Preparative enantiomer separation of dichlorprop with a cinchona-derived chiral selector employing centrifugal partition chromatography and high-performance liquid chromatography: A comparative study

Article abstract of DOI:10.1021/ac040102y

A countercurrent chromatography protocol for supportfree preparative enantiomer separation of the herbicidal agent 2-(2,4-dichlorphenoxy)propionic acid (dichlorprop) was developed utilizing a purposefully designed, highly enantioselective chiral stationary-phase additive (CSPA) derived from bis-1,4-(dihydroquinidinyl)phthalazine. Guided by liquid-liquid extraction experiments, a solvent system consisting of 10 mM CSPA in methyl tert-butyl ether and 100 mM sodium phosphate buffer (pH 8.0) was identified as a suitable stationary/mobile-phase combination. This solvent system provided an ideal compromise among stationary-phase retention, enantioselectivity, and well-balanced analyte distribution behavior. Using a commercial centrifugal partition chromatography instrument, complete enantiomer separations of up to 366 mg of racemic dichlorprop could be achieved, corresponding to a sample load being equivalent to the molar amount of CSPA employed. Comparison of the preparative performance characteristics of the CPC protocol with that of a HPLC separation using a silica-supported bis-1,4-(dihydroquinidinyl)phthalazine chiral stationary phase CSP revealed comparable loading capacities for both techniques but a significantly lower solvent consumption for CPC. With respect to productivity, HPLC was found to be superior, mainly due to inherent flow rate restrictions of title CPC instrument. Given that further progress in instrumental design and engineering of dedicated, highly enantioselective CSPAs can be achieved, CPC may offer a viable alternative to CSP-based HPLC for preparative-scale enantiomer separation.

Full text of DOI:10.1021/ac040102y

Anal. Chem. 2004, 76, 5837-5848
Preparative Enantiomer Separation of Dichlorprop
with a Cinchona-Derived Chiral Selector Employing
Centrifugal Partition Chromatography and
High-Performance Liquid Chromatography:
A Comparative Study
Elena Gavioli,† Norbert M. Maier,*,† Cristina Minguillo´n,‡ and Wolfgang Lindner†
Institute of Analytical Chemistry, University of Vienna, Wa¨hringerstrasse 38, A-1090 Vienna, Austria, and
Institut de Recerca Biome`dica, Parc Cient´ıfic de Barcelona (IRBB-PCB), Universitat de Barcelona, Josep Samitier, 1-5,
E-08028 Barcelona, Spain
maturation from a purely analytical to a widely accepted industrial
production tool was facilitated by major progress achieved in
chromatography process engineering2 and, in particular, enantio-
selective adsorbent development.3 The systematic evolution of
enantioselective receptors provided a rich toolbox of chiral
selectors (CSs), capable of resolving virtually any racemic mixture
of interest. Developing dedicated immobilization strategies to
confine these CSs onto the surface of appropriate support materials
was a key for the generation of robust bonded CSPs, demonstrat-
ing long-term stability under continuous operation conditions.
Nevertheless, there are also inherent limitations associated
with preparative applications of bonded CSPs. CSP preparation
requires expensive support materials fulfilling strict criteria in
terms of chemical inertness, particle size and shape, controlled
porosity and surface properties. Also, CS immobilization often
involves sophisticated and cost-intensive chemistries. Additional
efforts have to be invested into column packing, a nontrivial task
at technical scale.4 Drawbacks may arise from the chemical
microenvironment in which the CS units are embedded in bonded
CSPs. The physical constraints imposed by the attachment may
compromise CS accessibility and induced fit-type chiral recognition
phenomena.5 Further, the inevitable introduction of additional
chemical entities (residual surface functionalities, spacer and linker
groups) may give rise to competing nonspecific interactions, which
interfere with stereoselective association processes.6-8 A serious
restriction of bonded CSPs are the notoriously low CS densities
that can be generated on the limited surface areas (e.g., for silica
gels <350 m2/ g) of conventional support matrixes, resulting in
relatively low preparative capacities.1 Depending on the size and
shape of the CSs, and the immobilization chemistry used for
grafting, the supports permit surface loading levels in
A countercurrent chromatography protocol for support-
free preparative enantiomer separation of the herbicidal
agent 2-(2,4-dichlorphenoxy)propionic acid (dichlorprop)
was developed utilizing a purposefully designed, highly
enantioselective chiral stationary-phase additive (CSP A)
derived from bis-1 ,4 -(dihydroquinidinyl)phthalazine.
Guided by liquid-liquid extraction experiments, a solvent
system consisting of 1 0 mM CSP A in methyl ter t-butyl
ether and 100 mM sodium phosphate buffer (pH 8.0) was
identified as a suitable stationary/ mobile-phase combina-
tion. This solvent system provided an ideal compromise
among stationary-phase retention, enantioselectivity, and
well-balanced analyte distribution behavior. Using a com-
mercial centrifugal partition chromatography instrument,
complete enantiomer separations of up to 3 6 6 mg of
racemic dichlorprop could be achieved, corresponding to
a sample load being equivalent to the molar amount of
CSP A employed. Comparison of the preparative perfor-
mance characteristics of the CP C protocol with that of a
HP LC separation using a silica-supported bis-1 ,4 -(dihy-
droquinidinyl)phthalazine chiral stationary phase CSP
revealed comparable loading capacities for both tech-
niques but a significantly lower solvent consumption for
CP C. With respect to productivity, HP LC was found to
be superior, mainly due to inherent flow rate restrictions
of the CP C instrument. Given that further progress in
instrumental design and engineering of dedicated, highly
enantioselective CSP As can be achieved, CP C may offer
a viable alternative to CSP -based HP LC for preparative-
scale enantiomer separation.
Chromatographic separation of racemic mixtures on chiral
stationary phases (CSPs) is recognized as an efficient approach
to access scaleable amounts of pure enantiomers.1 Its rapid
(2) Schulte, M.; Strube, J. J. Chromatogr., A 2 0 0 1 , 906, 399-416.
(3) L¨ammerhofer, M.; Lindner, W. In Handbook of Analytical Separations; Valko,
K., Ed.; Elsevier Science B. V.: Amsterdam, 2000; Vol. 1, pp 337-437.
(4) Moscariello, J.; Purdom, G.; Coffman, J.; Root, T. W.; Lightfoot, E. N. J.
Chromatogr., A 2 0 0 1 , 908, 131-141.
* Corresponding author: (e-mail) norbert.maier@univie.ac.at; (fax) ++43-1-
4277-9523; (phone) ++43-1-4277-52373.
(5) Franco, P.; Senso, A.; Oliveros, L.; Minguillon, C. J. Chromatogr., A 2 0 0 1 ,
906, 155-170.
† University of Vienna.
(6) Pirkle, W. H.; Welch, C. J. J. Chromatogr. 1 9 9 2 , 589, 45-51.
(7) Fornstedt, T.; Sajonz, P.; Guiochon, G. Chirality 1 9 9 8 , 10, 375-381.
(8) Gotmar, G.; Fornstedt, T.; Guiochon, G. Chirality 2 0 0 0 , 12, 558-564.
‡ Universitat de Barcelona.
(1) Francotte, E. R. J. Chromatogr., A 2 0 0 1 , 906, 379-397.
10.1021/ac040102y CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/25/2004
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004 5837

the modest range of 0.3-1.0 µmol of CS/ m2. Higher CS loading
may be achieved with polymer coating technology,5 but at the
risk of compromised CS accessibility and function due to steric
overcrowding. Another problematic issue in context with the use
of bonded CSPs at preparative scale concerns irreversible adsorp-
tion of contaminants.2,9 This may alter the overall adsorption
characteristics of bonded CSPs, enforcing reoptimization of
operation parameters, elaborate washing protocols, or ultimately
a complete exchange of the affected CSP material.
Many of the problems inherent to bonded CSPs may be
resolved by resorting to alternative preparative chromatographic
methodologies that completely circumvent CS immobilization.
Specifically, support-free liquid-liquid partition chromatographic
technologies, e.g., countercurrent chromatography (CCC) and the
conceptually closely related centrifugal partition chromatography
(CPC), may provide such alternatives.10,11 These techniques use
immiscible solvents (or solvent mixtures) as stationary and mobile
phases. During the chromatographic process, the liquid stationary
phase is ”immobilized” in the column compartment by a strong
gravitational field, generated by centrifugation, while the mobile
phase is forced to percolate the former by pumping. To create an
enantioselective stationary phase, a suitable CS is dissolved in an
appropriate solvent, acting as a chiral stationary-phase additive
(CSPA). The separation of the enantiomers introduced with the
mobile phase is then effected by selective partition due to the
differential stabilities of the diastereomeric CSPA-analyte com-
plexes formed in the stationary phase.
ployed for enantiomer separation of amino acid derivatives, drugs,
and metabolites. Preparative runs performed in the course of these
studies gave promising results,16,19-21 encouraging further research
in this field.
However, a problem currently limiting routine application of
CPC/ CCC for enantiomer separation is the lack of variety of
efficient CSPAs. CSs developed for bonded CSPs frequently fail
in CPC/ CCC application due to insufficient enantioselectivity,
unfavorable solubility and phase distribution characteristics, and
incompatibility of the molecular recognition mechanisms with
mobile- or stationary-phase solvents.11,20 Specific chemical modi-
fication of existing highly enantioselective CS systems, however,
may allow generation of dedicated CSPAs fulfilling the multifaceted
criteria in terms of enantioselectivity, solubility, and phase-transfer
properties.21
Addressing these issues, we report here on the development
of preparative CPC and HPLC enantiomer separation protocols
for dichlorprop utilizing a dedicated CSPA and a bonded CSP,
respectively, both derived from bis-1,4-(dihydroquinidinyl)phthala-
zine (Figure 1). We outline important aspects concerning design
and synthesis of the CSPA and discuss the systematic optimization
of the operation conditions of the chromatographic separation
protocol. The preparative performance characteristics of these
complementary enantiomer separation methodologies are as-
sessed and critically evaluated on the basis of productivity-related
criteria and environmental considerations.
The use of stationary phases with physically unconfined CS
units may offer various practical advantages over the immobilized
CS regime in bonded CSPs. Most appealing, the process of
preparing an enantioselective stationary phase is simplified to
dissolving the CSPA in an appropriate solvent, suspending any
need for expensive solid supports and sophisticated immobilization
chemistries. Column packing can be achieved with ease by filling
the column compartment with the CSPA solution by pumping.
Using CSs as free solution species also obviates any support-
induced conformational restrictions and nonspecific interaction
interfering with selective CS-analyte association. The CS loading
level (and thus the preparative capacity) of the stationary phase
may conveniently be adjusted over a broad concentration range,
with limits being theoretically dictated only by CSPA solubility.
Attracted by these potential benefits, several research groups
have studied the utility of CCC and CPC for enantiomer separa-
tions. The results of these investigations have been discussed in
detail in recent review articles.11 Various CS systems, including
bovine serum albumin,12-14 π-acidic amino acid deriva-
tives,15-18 sulfated â-cyclodextrins,19 the antibiotic vancomycin,20
and cinchona alkaloid derivatives,21 have been successfully em-
EXPERIMENTAL SECTION
General Information. Unless stated otherwise, all reactions
were carried out under strictly anhydrous conditions and under
nitrogen atmosphere. All solvents were dried according to
standard procedures and distilled prior to use. The 1H NMR
spectra were acquired on a Bruker DRX 400-MHz spectrometer.
The chemical shifts (δ) are given in parts per million (ppm)
relative to TMS as internal standard. IR spectra were recorded
with a Perkin-Elmer Spectrum 2000 spectrometer. Mass spectra
were acquired on a PESciex API 365 triple quadrupole instrument
using electrospray ionization. Sample solutions in appropriate
solvents (chloroform/ methanol) were infused at concentrations
of ∼0.1 mg/ mL via a syringe pump at a flow rate of 5 µL/ min.
The electrospray voltage was typically set to 5250 V. Optical
rotation values were measured on a Perkin-Elmer 341 polarimeter
at 25 °C. Melting points were determined with a Kofler apparatus,
equipped with a Leica Galen III microscope. Thin-layer chroma-
tography was carried out with Silica gel 60 F254 aluminum sheets
provided by Merck (Darmstadt, Germany). Flash chromatography
was performed on Silica 60 (0.040-0.063-mm particle size
(Merck).
Materials. 2-(2,4-Dichlorophenoxy)propionic acid (dichlor-
prop), butyllithium, 1,4-dichlorophthalazine, tetrahydrofuran (THF),
potassium carbonate, potassium hydroxide, and octadecylmer-
(9) Miller, L.; Grill, C.; Yan, T.; Dapremont, O.; Huthmann, E.; Juza, M. J.
Chromatogr., A 2 0 0 3 , 1006, 267-280.
(10) Foucault, A. P.; Chevolot, L. J. Chromatogr., A 1 9 9 8 , 808, 3-22.
(11) (a) Foucault, A. P. J. Chromatogr., A 2 0 0 1 , 906, 365-378. (b) Oliveros, L.;
Minguillo´n, C.; Franco P.; Foucault, A. P. Enantioseparations in Counter-
current Chromatography (CCC) and Centrifugal Partition Chromatography
(CPC). In Countercurrent Chromatography, the support-free liquid stationary
phase; Berthod, A., Ed.; Comprehensive Analytical Chemistry Series;
Elsevier: New York, 2002; Chapter 11.
(15) Oliveros, L.; Franco Puertolas, P.; Minguillo´n, C.; Camacho-Frias, E.;
Foucault, A.; Goffic, F. L. J. Liq. Chromatogr. 1 9 9 4 , 2301-2318.
(16) Ma, Y.; Ito, Y.; Foucault, A. J. Chromatogr., A 1 9 9 5 , 704, 75-81.
(17) Ma, Y.; Ito, Y. Anal. Chem. 1 9 9 5 , 67, 3069-3074.
(18) Ma, Y.; Ito, Y. Anal. Chem. 1 9 9 6 , 68, 1207-1211.
(12) Shinomiya, K.; Kabasawa, K.; Ito, Y. J. Liq. Chromatogr. 1 9 9 8 , 21, 135-
141.
(13) Arai, T.; Kuroda, H. Chromatographia 1 9 9 1 , 32, 56-60.
(14) Ekberg, B.; Sellergren, B.; Albertsson, P. A. J. Chromatogr. 1 9 8 5 , 333, 211-
214.
(19) Breinholt, J.; Lehmann, S. V.; Varming, A. R. Chirality 1 9 9 9 , 11, 768-771.
(20) Duret, P.; Foucault, A.; Margraff, R. J. Liq. Chromatogr., Relat. Technol.
2 0 0 0 , 23, 295-312.
(21) Franco, P.; Blanc, J.; Oberleitner, W. R.; Maier, N. M.; Lindner, W.;
Minguillon, C. Anal. Chem. 2 0 0 2 , 74, 4175-4183.
5838 Analytical Chemistry, Vol. 76, No. 19, October 1, 2004

Figure 1. Structures of the chiral herbicidal agent dichlorprop and the co-generic CSP and CSPA derived from (DHQD)2PHAL.
captan were purchased from Aldrich (Vienna, Austria). Dihydro-
quinidine (DHQD) was provided by Boehringer Mannheim
(Mannheim, Germany) and quinidine (QD) was bought from
Buchler (Braunschweig, Germany). Ammonium acetate, acetic
acid, and methyl tert-butyl ether (MTBE) were purchased from
Fluka (Buchs, Switzerland). R,R’-Azoisobutyronitrile was from
Merck. Magnesium sulfate was purchased from Riedel de Haen
(Vienna, Austria). Spherical silica gel (ProntoSIL 120-5-Si, 5 µm,
320 m2/ g) was from Bischoff Chromatography GmbH (Leonberg,
Germany) and was modified with mercaptopropyl groups following
a literature procedure.23 Aqueous ammonia solution (25%), di-
sodium hydrogen phosphate, sodium dihydrogen phosphate,
orthophosphoric acid, sodium hydroxide, and methyl isobutyl
ketone (MIBK) were purchased from Panreac Qu´ımica (Barce-
lona, Spain). Ultrapure water used for mobile-phase preparation
was obtained from a MilliQ Academic A10 system. All the buffered
mobile phases were filtered under reduced pressure through a
0.45-µm membrane filter.
Instruments. The CPC runs were performed on a high-
performance CPC apparatus model LLB-M (EverSeiko, Tokyo,
Japan), equipped with a stacked circular partition disk rotor (2136
channels, 220 mL of internal volume). The latter was connected
to a Hewlett-Packard 1100 chromatography system. A Rheodyne
injector valve with a 2.4-mL loop was employed for manual sample
loading. Semipreparative HPLC separation and enantiomeric
excess determination were performed on a Merck-Hitachi HPLC
system, consisting of a semipreparative pumping system series
L-7150, an autosampler series L-7250 equipped with a 100-µL
sample loop, a diode array detector series L-7455, and an interface
series D-7000. Data processing was performed with D-HSM 7000
software. Determination of the enantiomeric excess of dichlorprop
was performed with a Chiral AX-QN1 CSP (125 × 4 mm i.d., CS
loading 190 µmol/ g) from Bischoff Chromatography.
Synthesis of the ((DHQD)2 P HAL-Type CSP and CSP A
(Figure 2 ). 1 -Chloro-4 -(9 -O-dihydroquinidinyl)phthalazine
(1 ). Dihydroquinidine (10.80 g, 33.0 mmol) was dissolved in dry THF
(200 mL), and the solution was cooled in an ice-water bath. Butyl-
lithium (2 M solution in pentane, 17 mL, 33.0 mmol) was added via
syringe. Solid 1,4-dichlorophthalazine (8.00 g, 39.6 mmol) was added
in a single portion, and the mixture was stirred at ambient temperature.
After 5 h, TLC analysis (mobile phase: CHCl3/ MeOH, 10:1) showed
complete consumption of dihydroquinidine. The solvent was evaporated
under reduced pressure at 40°C. The solid residue was dissolved in
CHCl3 (200 mL) and washed with water (3 × 200 mL). The organic
phase was dried (MgSO4) and concentrated to a volume of 100 mL
under reduced pressure. The concentrated solution was purified by
flash chromatography on silica gel (300 g), eluting with CHCl3/ MeOH,
10:1, to give 12.11 g (24.8 mmol, 75%) of a white solid: mp 194-196
°C; IR (KBr) 2932, 1620, 1538, 1387 cm-1; MS (ESI) 489.3 [M + H]+,
977.5 [2M + H]+, 1467.8 [3M + H]+; 1H NMR (CDCl3) δ 8.67 (d,
1H), 8.38 (m, 1H), 8.18 (m, 1H), 7.99 (m, 3H), 7.62 (d, 1H), 7.47 (d,
1H), 7.34 (dd, 1H), 7.28 (d, 1H), 3.98 (s, 3H), 3.52 (m, 1H), 2.91 (m,
1H), 2.84 (m, 2H), 2.74 (m, 1H), 2.04 (m, 1H), 1.79 (m, 1H), 1.70-
1.42 (overlapped m, 6H), and 0.92 (t, 3H); optical rotation [R]589
-184.9°, [R]436 -533.7°, [R]546 -232.9°, (c 1.0, CHCl3).
1 -(9 -O-Qu inidinyl)-4 -(9 -O-dihydroqu inidinyl)phthala-
zine (2). Quinidine (10.41 g, 32.2 mmol) was dissolved in dry toluuene
(300 mL). From this solution, a volume of ∼80 mL was distilled to
remove traces of water. The remaining solution was added to added
to a suspension of 1 (12.11 g, 24.8 mmol) in toluene (80 mL). After
addition of finely powdered K2CO3 (4.80 g, 35.5 mmol) and KOH (2.40
(22) Han, H.; Janda, K. D. Tetrahedron Lett. 1 9 9 7 , 38, 1527-1530.
(23) Maier, N. M.; Nicoletti, L.; L¨ammerhofer, M.; Lindner, W. Chirality 1 9 9 9 ,
11, 522-528.
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004 5839

1 -(9 -O-Dihydroquinidinyl)-4 -(1 1 -octadecylthia-9 -O-dihy-
droquinidinyl)phthalazine ((DHQD)2P HAL-Type CSP A). Pre-
cursor 2 (13.60 g, 17.5 mmol) was dissolved in dry chloroform
(125 mL). Octadecylmercaptan (25.10 g, 87.5 mmol) and azoisobu-
tyronitrile (150 mg, 0.9 mmol) were added. The mixture was
refluxed with stirring for 8 h. The reaction mixture was loaded
onto a flash chromatography column (200 g silica) preconditioned
with chloroform. After elution of excess octadecylmercaptan with
chloroform, the product was isolated by elution with chloroform/
methanol, 20:1. Yield, 11.7 g (11.0 mmol, 63%) of a white foam:
mp 59-62 °C; IR (KBr) 2928, 2853, 1621, 1592, 1552, 1508 cm-1
;
MS (ESI) 532.6 [M + 2H]2+, 1063.8 [M + H]+, 2128.0 [2M +
1
H]+; H NMR (CDCl3) δ 8.65 (m, 2H), 8.35 (m, 2H), 7.57 (m,
4H), 7.53 (m, 2H), 7.43 (m, 2H), 7.38 (m, 2H), 7.02 (m, 2H), 3.91
(s, 6H), 3.42 (m, 2H), 2.86-2.62 (m, 8H), 2.42 (m, 4H), 2.08-
1.92 (m, 2H), 1.78-1.18 (m, 46H), 0.88 (m, 3H), and 0.77 (m, 3H);
optical rotation [R]589 -147.1°; [R]436 -403.9°; [R]546 -184.6° (c
1.0, CHCl3).
Screening by Liquid-Liquid Extraction. Stock Solutions.
The following stock solutions and buffers were prepared for
liquid-liquid extraction experiments: 100 mM CSPA (1.063 g of
CSPA in 10 mL of chloroform); 100 mM dichlorprop (235 mg of
racemic dichlorprop in 10 mL of methanol); 100 mM sodium
phosphate and 100 mM ammonium acetate buffers, pH 6.8, 7.0,
and 8.0. Prior to use in liquid-liquid extraction experiments, these
buffers were equilibrated with MIBK and MTBE, respectively,
by shaking in a separation funnel.
Figure 2. Synthetic scheme for the preparation of the (DHQD)2PHAL-
type CSP and CSPA.
Liquid-Liquid Extraction Experiments. A set of 24 glass
centrifuge tubes was charged with 200 µL of CSPA and 200 µL of
dichlorprop stock solution and dried under a stream of nitrogen.
The residues were reconstituted in 2 mL of the respective organic
solvent (MIBK or MTBE) and mixed with 2 mL of the respective
aqueous buffers. The tubes were sealed with Teflon-lined screw
caps and allowed to equilibrate for 6 h at 25 °C with shaking.
After phase separation, 250-µL aliquots were withdrawn from the
aqueous (lower) phase, transferred into HPLC vials, and diluted
with methanol (1.25 mL). The samples were analyzed on a Chiral
AX QN-1 CSP (125 × 4 mm i.d.) using methanol/ 0.1 M am-
monium acetate (80:20), pHa 6.0 as mobile phase. A flow rate of
1 mL/ min, a detection wavelength of 230 nm, and a column
temperature of 25 °C were used. The retention factor of the first
eluted R-enantiomer was 5.47 and that of the more retained
S-enantiomer was 6.61. The dichlorprop concentrations in the
individual phases were calculated by comparison with calibration
standards. The resultant data were used to calculate CSPA-
mediated distribution ratios for the individual enantiomers (DR;
DS) and the corresponding enantioselective distribution factor Rex
(DS/ DR). The non-CSPA-mediated distribution ratio (D0) of dichlor-
prop was performed in an analogous fashion, omitting the CSPA
in the organic phase.
g, 43.6 mmol), the mixture was refluxed with stirring for 10 h. TLC
analysis of the resulting mixture (CHCl3/ MeOH 10:1) showed complete
consumption of the starting material. The reaction mixture was allowed
to cool, diluted with water (300 mL), and extracted with ethyl acetate
(300 mL). The organic phase was washed with brine (2 × 250 mL),
dried (MgSO4), and concentrated under reduced pressure to give a
yellowish oil. Purification by flash chromatography (280 g of silica,
gradient elution with CHCl3/ MeOH 100/ 0-90/ 10) gave 17.90 g (23.0
mmol, 93%) of a white foam: mp 119-122 °C; IR (KBr) 2934, 1621,
1
1551, 1509 cm-1; MS (ESI) 777.8 [M + H]+, 1554.1 [2M + H]+; H
NMR (CDCl3) δ 8.65 (m, 2H), 8.34 (m, 2H), 8.01 (m, 2H), 7.94 (m,
2H), 7.58 (m, 2H), 7.43 (m, 2H), 7.37 (m, 2H), 7.05 (d, 1H), 6.99 (d,
1H), 5.95 (m, 1H), 4.98 (m, 2H), 3.93 (s, 6H), 3.41 (m, 2H), 2.94 (m,
1H), 2.85-2.61 (m, 5H), 2.24 (m, 1H), 2.09 (m, 1H), 1.96 (m, 1H),
1.82 (s, 1H), 1.74-1.36 (m, 12H), and 0.78 (m, 3H); optical rotation
[R]589 ) -170.1°; [R]436 ) -477.1°; [R]546 ) -214.3°, (c 1.0, CHCl3).
Immobilization of the Chiral Selector onto Mercaptopro-
pyl-Silica ((DHQD)2 P HAL-Type CSP ). A 3.0-g amount of
mercaptopropyl-modified silica gel23 was suspended in methanol
(40 mL). Precursor 2 (256.0 mg, 330 µmol) and azoisobutyronitrile
(5.0 mg, 30 mmol) were added. The mixture was refluxed with
gentle mechanical stirring for 5 h. The modified silica gel was
isolated by filtration through a sintered glass funnel (porosity 4).
The silica gel was washed with CHCl3 (50 mL), CHCl3/ MeOH
1:1 (2 × 50 mL), and hot MeOH (4 × 50 mL) and dried in high
vacuum at 60 °C. Elemental analysis gave 8.59% C, 1.54% H, 0.69%
N, and 2.64% S, corresponding to a CS loading of 82 µmol/ g, 0.26
µmol/ m2 (calculated based on N content).
CP C Experimental Conditions. Stationary phases for CPC
experiments were prepared by dissolving 4.26 g (4.0 mmol) of
CSPA in 400 mL of MTBE or MIBK, respectively. The corre-
sponding stationary phase (in all cases the one with lower density)
was transferred into the CPC rotor compartment by descending
5840 Analytical Chemistry, Vol. 76, No. 19, October 1, 2004

mode pumping, replacing the previousely charged aqueous mobile
phase. As soon as the stationary phase appeared at the outlet,
the rotor valve was closed and the rotation (1100-1200 rpm) was
started. After 30 min of centrifugation, the respective aqueous
mobile phase was pumped into the rotor compartment in descend-
ing mode. After initially occurring displacement of stationary
phase, clear mobile phase appeared at the outlet. Stabilization of
the back pressure indicated that the two-phase system in the rotor
had reached a steady-state equilibrium. The volume of stationary
phase retained in the equilibrated CPC system was calculated as
the difference of the volumes originally loaded and the fraction
displaced on introduction of the mobile phase. The amount of
racemic dichlorprop to be injected was calculated from the total
CSPA amount available in the equilibrated CPC system. Racemic
dichlorprop samples corresponding to 0.25, 0.5, 1, and 2 times
the molar amount of the total CSPA amount were dissolved in
the corresponding organic solvent (2.4 mL) and injected. For all
experiments, the flow rate of the mobile phase was set at 3 mL/
min. The eluate was collected in 3-mL fractions until the return
of the online UV signal indicated complete dichlorprop elution.
Due to signal saturation, well-resolved elution profiles could not
be obtained with the employed analytical UV detector. Therefore,
the corresponding CPC elution profiles were constructed by
analyzing the individual fractions by HPLC and plotting the
corresponding concentrations of R- and S-enantiomers versus time.
For this purpose, a 500-µL aliquot of each fraction was withdrawn,
diluted with 500 µL of methanol, and analyzed by the above-
described chromatographic procedure.
Recovery of Dichlorprop Enantiomers and CSP A. The
individual dichlorprop enantiomers were recovered from the
fractions of the separation performed in MTBE/ phosphate buffer
at pH 8.0 with equimolar carrier/ analyte concentrations as follows.
The fractions containing the individual enantiomers were pooled,
acidified with concentrated hydrochloric acid, and extracted with
chloroform (3 × 100 mL). The combined organic phases were
dried over anhydrous sodium sulfate, and the solvent was
evaporated under reduced pressure. Gravimetric quantification
gave recoveries of 165 (90%) and 174 mg (95%) for the R-
enantiomer and S-enantiomer, respectively. For recovery of the
CSPA, water was pumped in ascending mode through the system
without any rotation. The resulting effluent was saturated with
sodium chloride to break the emulsion. Separation of the organic
phase, drying over sodium sulfate, and evaporation of the solvents
allowed >90% recovery of the CSPA.
of more than 23 000-30 000 tons in the United States.24 Dichlorprop
exhibits pronounced enantioselective biological activity with
herbicidal properties being restricted to the R-enantiomer, whereas
the S-enantiomer shows even some anti-auxin effects.25 The
intense and worldwide application of racemic dichlorprop con-
tributes considerable to global pollution, the level of which could
be cut down by half using the active enantiomer only. Therefore,
the development of efficient enantiomer separation methodologies
for this target compound can be considered a highly relevant
challenge.
Our efforts to identify enantioselective receptors for environ-
mentally relevant chiral compounds led to the discovery that bis-
1,4-(dihydroquinidinyl)phthalazine, a cinchona ligand widely used
for asymmetric synthesis protocols,26 also possesses excellent
chiral recognition properties for dichlorprop (Figure 1). Thus, a
silica-supported CSP, derived from (DHQD)2PHAL, produced an
RHPLC of 15.3 for dichlorprop under optimized chromatographic
conditions. This unprecedented enantioselectivity and the ready
accessibility of (DHQD)2PHAL recommend this ligand as a
promising CS candidate for the development of bonded CSPs for
preparative chromatographic enantiomer separation of dichlor-
prop. Unfortunately, our attempts to generate high-capacity
(DHQD)2PHAL-type CSPs resulted in materials with disappoint-
ingly low CS coverages, the best results being in the range of
<150 µmol/ g. Evidently, the high molecular weight (∼760) and
the spherical molecular shape of (DHQD)2PHAL prohibit more
densely grafted surfaces due to steric congestion.
Faced with this limitation, we decided to explore the utility of
(DHQD)2PHAL for support-free CPC enantiomer separation,
avoiding surface immobilization and allowing for more flexibility
in terms of CS loading. Considering environmental issues, we
projected a CPC protocol operating with a CSPA-charged organic
stationary phase and a purely aqueous mobile phase. Successful
realization of this concept called for a (DHQD)2PHAL species
showing excellent solubility in the organic stationary phase while
being practically insoluble in the aqueous mobile phase. This
crucial requirement was addressed by enhancing the intrinsic
hydrophobicity of (DHQD)2PHAL by covalent attachment of a
lipophilizing moiety (Figure 1). To preserve the functional integrity
of (DHQD)2PHAL, an octadecylthia group was attached at the
C
11 position of the common precursor, reproducing the molecular
microenvironment of the silica-supported CS. The corresponding
synthetic aspects are discussed in detail in the next section.
Synthesis. The synthetic routes employed for the preparation
of the silica-supported (DHQD)2PHAL-type CSP and the corre-
sponding CSPA from a common precursor are outlined in Figure
2.
Using a modified literature procedure,22 monochlorophthala-
zine 1 was prepared in 75% yield from DHQD by deprotonation
with butyllithium and subsequent alkylation of the resultant
alkoxide with 1,4-dichlorophthalazine. Condensation of 1 with QD
under basic conditions provided the common precursor 2 in 93%
yield,22 comprising a single vinyl group as a handle for further
functionalization. Reaction of 2 with mercaptopropyl-modified silica
gel23 in the presence of AIBN as a free radical initiator gave the
HP LC Semipreparative Enantioseparation of Dichlorprop.
Semipreparative separation of 4.9, 9.7, and 19.5 mg of racemic
dichlorprop was carried out on a 150 × 4 mm i.d. column packed
with 1.10 g of (DHQD)2PHAL-type CSP with a loading level of 82
µmol of CS/ g. The mobile phase consisted of methanol/ acetic
acid/ ammonium acetate (92:2:0.5, v/ v/ w). A flow rate of 1 mL/
min and a column temperature of 25 °C (controlled with a Haake
C40 water thermostat) were used. Peak detection was performed
at 300 nm. Samples were dissolved in methanol at a concentration
of 100 mg/ mL. The enantiomeric excess of each fraction collected
was determined as described above.
(24) Archibald, S. O.; Winter, C. K. Chemicals in the Human Food Chain; Van
Nostrand-Reinhold: New York, 1990.
(25) Loos, M. A. In Phenoxyalkanoic acids; Kearny, P. C., Kaufmann, D. D., Eds.;
Marcel Dekker: New York, 1975; Vol. 1, pp 1-101.
(26) Kacprzak, K.; Gawronski, J. Synthesis 2 0 0 1 , 7, 961-998.
RESULTS AND DISCUSSION
HP LC versus CP C: Selection of the Model System.
Dichlorprop represents one of the most frequently employed chiral
herbicidal agents in crop protection, with an estimated annual use
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004 5841

enantioselective (DHQD)2PHAL-dichlorprop association mech-
anism can be drawn. The observed pH-retention profile, reflecting
most efficient analyte binding at an intermediate pH range, clearly
indicates that dichlorprop-(DHQD)2PHAL complex formation is
primarily driven by electrostatic interactions. The contribution of
these electrostatic interactions to complex stabilization, however,
appears to be largely nonenantioselective in nature, as their partial
disruption at the high pH values improves rather than impairs
enantioselectivity.
The fact that pH variation affects primarily binding affinity but
to a much lesser extent enantioselectivity has important practical
implication for selection of the CPC mobile-phase conditions. From
the pH-retention profile in Figure 3 it can be concluded that
adjusting mobile-phase pH to values of >7 offers a convenient
means to tune (DHQD)2PHAL-dichlorprop binding affinity with-
out compromising overall enantioselectivity. Consequently, aque-
ous phosphate and acetate buffers covering a pH range from 7.0
to 8.0 were selected as mobile phases for further optimization
studies.
Figure 3. Mobile-phase-pH dependency of the retention factors
(kS, kR) and the enantioselectivity factor (RHPLC) of dichlorprop on the
(DHQD)2PHAL-type CSP. Conditions: column 150 × 4 mm i.d.;
mobile phase methanol/20 mM sodium phosphate (80:20, v/v), pHa
adjusted with aqueous orthophosphoric acid (50%) and 6 M sodium
hydroxide solutions; flow rate 1 mL/min; UV detection 254 nm; column
temperature 25 °C.
Liquid-Liquid Extraction Screening. The prudent choice
of appropriate mobile/ stationary-phase systems is the most crucial
task in the development of CPC separation protocols.27 To
guarantee adequate performance in CPC applications, a solvent
combination has to fulfill several crucial requirements. Considering
the dynamic nature of liquid-liquid partition chromatography, the
phase separation kinetics of the stationary/ mobile-phase system
must be sufficiently fast. In CPC applications, slow phase separa-
tion may give rise to the formation of emulsions, leading to an
instable separation process or complete displacement of the
stationary phase.27 For enantiomer separation, the solvent system
must be fully compatible with the chiral recognition mechanism
of the chosen CSPA. The stationary phase must show excellent
dissolving capacity for the CSPA and the corresponding CSPA-
analyte complexes to ensure high preparative capacity and must
completely retain the CSPA under chromatographic operation
conditions. Leaching effects would inevitably lead to a continuous
loss of CSPA from the stationary phase and, consequently, to
system instability and severe contamination of the product
fractions. Concerning the analytes, the solvent combination must
show a well-balanced partition behavior, with optimal stationary/
mobile-phase distribution ratios in the range of 0.2-3.0.11 Lower
distribution ratios correspond to a preferential partition of the
analytes into the mobile phase and, thus, poor interaction with
the stereodiscriminating CSPA. On the contrary, exceedingly high
distribution ratios indicate very strong CSPA-analyte interactions,
resulting in loss of peak resolution and impracticably large elution
volumes and long run times.
corresponding CSP with a CS coverage of 82 µmol/ g. For the
preparation of the (DHQD)2PHAL-type CSPA, precursor 2 was
reacted with an excess of octadecylmercaptan in the presence of
AIBN, affording the corresponding highly lipophilic thioether in
61% yield. This compound showed excellent solubility in a wide
range of organic solvents (>300 mg/ mL for MTBE, MIBK, and
CHCl3) but was essentially insoluble in water over a pH range of
5.0-9.0.
Development of the CP C Enantiomer Separation P rotocol.
Influence of pH on Enantioselectivity/ Binding Affinity. To
facilitate the development of an efficient CPC enantiomer separa-
tion protocol, the impact of operational parameters on enantio-
selective binding of dichlorprop to (DHQD)2PHAL was explored.
Considering the acid-base character of the receptor-analyte
system under investigation, it was of particular interest to establish
to what extent enantioselectivity or binding affinity could be
controlled by pH variations of the mobile phase. Therefore, the
retention of the dichlorprop enantiomers on the silica-supported
(DHQD)2PHAL-type CSP was studied in a pH range of 4.0-7.6
with a hydroorganic mobile phase consisting of methanol/ 20 mM
sodium phosphate buffer (80:20, v/ v). The corresponding data
are depicted in Figure 3.
On increasing the mobile-phase pH, the retention factor of the
more strongly bound S-enantiomer, kS, was significantly enhanced,
reaching a distinct maximum at pH 6.7. Changing the mobile-
phase pH from 4.0 to 6.7 increased kS from 6.5 to 37.0, while a
further increase to pH 7.5 induced a dramatic loss in retention to
kS ) 9.5. An analogous tendency, albeit far less pronounced, was
observed for the less strongly bound R-enantiomer. For enantio-
selectivity, however, a different pH profile was observed. The
enantioseparation factor (RHPLC) improved more or less linearly
with increasing mobile-phase pH through the entire range studied,
showing an enhancement from R ) 5.1 at pH 4.0 to R ) 11.1 at
pH 7.5.
Guided by these criteria, we explored the performance
characteristics of different solvent combinations for enantioselec-
tive partitioning of dichlorprop in the presence of the (DHQD)2-
PHAL-type CSPA by liquid-liquid extraction experiments. To
ensure operational simplicity, and to comply with environmental
and safety issues, we restricted our choice of organic mobile
phases to inexpensive, low-toxicity single solvents. For these
reasons, MIBK and MTBE were selected as stationary-phase
solvents. These solvents dissolved up to 300 mg/ mL CSPA,
dichlorprop, and the corresponding diastereomeric complexes at
Although the underlying chiral recognition mechanism is still
under investigation, two preliminary conclusions concerning the
(27) Foucault, A. P. Centrifugal Partition Chromatography; Chromatographic
Science Series 68; Marcel Dekker: New York, 1999.
5842 Analytical Chemistry, Vol. 76, No. 19, October 1, 2004

organic solvents seems to enhance their intrinsic polarity to an
extent that allows solvation of the dichlorprop salts and thus their
transfer into the organic phase. The extent of non-CSPA-mediated
partitioning into the organic phase was influenced by the nature
of the buffer salts and the pH of the aqueous phase. Evidently,
the presence of ammonium acetate, as compared to sodium
phosphate, greatly facilitated the transfer of dichlorprop into the
organic phase. Thus, at pH 6.8, with MTBE a D0 ) 1.32 was
observed with ammonium acetate, whereas with sodium phos-
phate extraction into the organic phase was less pronounced (D0
) 0.21). Preferential hydrogen-bonding interactions of the am-
monium cation with the acceptor-type organic solvents (and
dissolved water) may play a role in this partition phenomenon.
At pH 8.0, partition of dichlorprop into the organic phases was
largely suppressed for both buffer systems. This may originate
from a shift of the dichlorprop dissociation equilibrium in favor
of the deprotonated species. The nature of the organic solvent
had only a minor influence on the partition behavior, indicating
that these media exhibit rather similar solvation capacities for
dichlorprop.
Performing the extraction experiments in the presence of the
(DHQD)2PHAL-type CSPA induced dramatic shifts in the partition
equilibrium of dichlorprop in favor of the organic phase. Generally,
the addition of CSPA to the organic phase enhanced the dichlo-
rprop extraction yields by a factor of up to >40. In all cases, (S)-
dichlorprop was selectively extracted into the CSPA-comprising
organic phase, providing impressive enantiomer enrichments up
to ee > 50%. The observed preference in enantioselective binding
is consistent with the retention characteristics of the bonded
(DHQD)2PHAL-type CSP, providing evidence that the extraction
and chromatographic processes are governed by closely related
chiral recognition phenomena.
The efficiency of the CSPA-mediated partition process was
primarily sensitive to the pH value and the nature of buffer salts.
For all studied solvent/ buffer combinations, extractions performed
at pH 6.8 produced the highest levels of dichlorprop enrichment
in the organic phase. However, already a rather incremental shift
in pH to 7.0 led to a significant drop in distribution ratios, with a
most dramatic decrease at pH 8.0. This parallels the pH-retention
profile observed for the silica-supported CSP and is indicative of
the crucial role of ion-pairing interactions to complex stabilization.
In terms of extraction efficiency, MTBE and MIBK showed
similar performance characteristics, with MIBK offering somewhat
improved yields at lower pH values. This may reflect more efficient
ion solvation properties of MIBK due to its higher dipole moment
and dielectric constant. Concerning the buffer salts, the trends
were similar to those observed in the absence of the CSPA.
Ammonium acetate facilitated dichlorprop transfer into the organic
phase relative to sodium phosphate. Again, superior solvation
capacity of the employed organic solvents for the ammonium
cation may account for this behavior.
The levels of extraction enantioselectivity observed in these
liquid-liquid partition experiments, expressed as Rex ) DS/ DR,
were located in a rather narrow range from 3.9 to 7.6. The nature
of the organic solvent had some influence on chiral recognition,
with MTBE showing generally improved enantioselective parti-
tioning properties relative to MIBK. In contrast to the extraction
yields, Rex was found to be relatively insensitive to pH variation
of the aqueous buffer phase.
Table 1. Influence of Buffer Salts, pH, and Organic
Solvents on the Passive (Nonenantioselective) and
CSPA-Mediated (Enantioselective) Liquid-Liquid
Partition Behavior of Dichlorpropa
aqueous
buffer
organic
solvent
pH
eeorg
7.64
15.61 0.13
39.82 0.03
7.55
7.22
46.02 0.63
12.94 0.21
29.50 0.28
44.65 0.1
12.92 1.32
12.19 1.36
51.61 0.84
D0
DS
DR
Rex
phosphate
phosphate
phosphate
acetate
acetate
acetate
phosphate
phosphate
phosphate
acetate
acetate
acetate
6.8
7.0
8.0
6.8
7.0
8.0
6.8
7.0
8.0
6.8
7.0
8.0
MIBK
MIBK
MIBK
MIBK
MIBK
MIBK
MTBE
MTBE
MTBE
MTBE
MTBE
MTBE
0.31
16.7
7.31
1.58
24.57 4.75 5.17
15.34 4.33 3.54
1.22
15.02 2.61 5.76
5.88
1.91
18.25 2.72 6.72
15.51 2.78 5.58
2.08
4.25 3.93
1.79 4.08
0.36 4.42
1.04
1.07
0.26 4.79
0.87 6.76
0.33 5.69
0.27 7.56
a
A 2-mL aliquot of 100 mM aqueous buffer and 2 mL of 10 mM
CSPA-racemic dichlroporp were equilibrated at 25 °C. Phosphate,
sodium phosphate; acetate, ammonium acetate; MIBK, methyl isobutyl
ketone; MTBE, methyl tert-butyl ether; eeorg, enantiomeric excess (%)
of (S)-dichlorprop in the organic phase after extraction; D0, organic/
aqueous distribution ratio of dichlorporp in absence of the CSPA; DS,
DR, organic/ aqueous distribution ratios of (S)- and (R)-dichlorprop in
the presence of 10 mM CSPA; Rex, enantioselectivity of extraction
(DS/ DR).
ambient temperature. For reasons discussed in the preceding
section, aqueous sodium phosphate and ammonium acetate
buffers, 100 mM, at pH 6.8, 7.0, and 8.0 were tested as mobile
phases.
Concerning phase separation kinetics, different behaviors were
observed depending on the nature of the buffer salts in the
aqueous phase. In the presence of phosphate buffers, all systems
showed relatively fast phase separation, being typically complete
in less than 30 s. With acetate buffers, phase separation was
significantly slower (∼1 min), with a tendency to form short-lived
emulsions.
Next we explored passive (in absence of CSPA) and CSPA-
mediated liquid-liquid partition of dichlorprop for all possible
organic solvent and buffer combinations. For these experiments,
equal volumes of organic solvents and aqueous buffers, 10 mM
in (DHQD)2PHAL-type CSPA and racemic dichlorprop, respec-
tively, were equilibrated at 25 °C. The efficiency of the enantio-
selective analyte transfer to the organic phase was determined
by quantifying the relative amount and the enantiomeric ratio of
dichlorprop remaining in the aqueous phase. From these data,
the organic/ aqueous distribution ratios for the individual enan-
tiomers (DR, DS) and the corresponding enantioselectivity (DS/
DR) were calculated. In addition, control experiments in the
absence of the CSPA were performed to establish the nonselective
distribution ratio (D0) of dichlorprop. The corresponding results
are summarized in Table 1.
Inspection of the D0 values listed in Table 1 reveals that, even
in absence of the CSPA, significant amounts of dichlorprop could
be extracted into MTBE and MIBK at pH 6.8-7.0. Considering
the fact that in this pH range dichlorprop (pKa ) 2.93) exists
preferentially as a deprotonated and well water-solvated species,
this finding is rather unexpected. However, under the experimen-
tal conditions employed, both MTBE and MIBK are saturated with
the aqueous phase and may contain up to 1.5 and 1.9% (w/ w)
water, respectively. The presence of water molecules in the
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004 5843

Figure 4. Schematic representation of the chromatographic equipment and (inset) the descending mode mobile-phase flow regime used for
the CPC separation of the dichlorprop enantiomers.
As outlined above, solvent systems suitable for CPC enanti-
omer separation must combine the attributes of high levels of
enantioselectivity and well-balanced distribution behavior for the
analyte of interest, ideally with stationary/ mobile-phase distribu-
tion ratios in the range of 0.2-3.0. Inspection of the data
summarized in Table 1 indicates that all studied solvent combina-
tions fulfill these criteria in terms of enantioselectivity. Combina-
tions comprising pH 7.0 and 7.5 buffers, however, suffer from
exceedingly high distribution ratios and are therefore unsuitable
for economical CPC applications. Solvent systems with pH 8
buffers met both requirements in terms of distribution charac-
teristics (0.26 < DR, DS < 2.08) and enantioselectivity (Rex ) 4.42-
7.56) and were therefore evaluated as mobile phases for CPC
separations.
CP C Enantiomer Separation of Dichlorprop. A schematic
representation of the CPC apparatus and the operational principle
of the employed descendent chromatographic mode are given in
Figure 4. The CPC “column” is represented by a rotor assembled
from a stack of disks, having an engraved, interconnected system
of 2136 partition cells. An individual partition cell consists of a
channel and a duct compartment. The channels are the locations
where analyte partition through intense mobile/ stationary-phase
mixing occurs during the chromatographic process, while the
ducts provide the transfer of the mobile phase between the
channels.
retained under operational conditions in the CPC column defines
the total amount of CS available for analyte interactions and thus
the preparative capacity of the chromatographic system.
For preparative CPC enantiomer separation of dichlorprop, the
solvent systems comprising pH 8.0 buffers and displaying DR, DS
values in the most economic range of 0.2-2.2, were employed.
To ensure good reproducibility of the enantioselective partition
efficiencies observed in exploratory liquid-liquid extractions
study, all CPC runs were performed with stationary phases
comprising 10 mM (DHQD)2PHAL-type CSPA. For the selection
of operational parameters, we drew from our experience gained
in previous work.21 A rotational speed of 1100-1200 rpm and a
flow rate of 3 mL/ min were considered to provide a good
compromise among system stability, throughput capacity, and
upper pressure limit of the instrument (80 bar).
Under these conditions, the amount of stationary phase (and
thus the total amount of CSPA) retained in the CPC column was
found to depend on the nature of the organic stationary-phase
solvent. With MTBE-comprising solvent systems, stationary-phase
retention was 155 mL (70% of the total column volume), while
with the MIBK systems, only 130 mL (59% of the total column
volume) was retained. These different levels of CSPA loading were
considered when the sample amounts to be processed in the
preparative CPC runs were calculated.
To identify the most suitable solvent systems for the prepara-
tive enantiomer separation, a series of CPC chromatographic
experiments was performed with the set of selected solvent
systems and increasing amounts of racemic dichlorprop. Com-
parative runs conducted with the ammonium acetate/ MIBK and
phosphate/ MIBK systems and a sample load corresponding to a
molar dichlorprop/ CSPA ratio (r) of 0.25 both produced clean
baseline separation (data not shown). However, the CPC run
performed with acetate buffer as mobile phase showed signifi-
cantly increased analyte retention (acetate buffer kR ) 1.27, kS )
In contrast to HPLC with bonded CSPs, column packing in
CPC is generally performed under dynamic conditions. For this
purpose, the CPC column is first charged with the stationary
phase, followed by the setup of the appropriate operational rotation
speed and mobile-phase flow rate. During the initial equilibration
stage, the mobile phase displaces a certain fraction of the
“gravitationally immobilized” stationary phase until a steady state
is reached, depending on the nature of the solvent systems and
the operational conditions used. The volume of stationary phase
5844 Analytical Chemistry, Vol. 76, No. 19, October 1, 2004

Figure 5. Preparative CPC enantiomer separation of dichlorprop performed with increasing sample loads under different stationary-phase
conditions. All experiments were carried out at a flow rate of 3 mL/min with aqueous sodium phosphate buffer (100 mM, pH 8.0) as mobile
phase at a rotor speed of 1100 (MTBE) and 1200 rpm (MIBK), and a temperature of 25 °C. Series A: 10 mM (DHQD)2PHAL-type CSPA in
MIBK as stationary-phase solvent. Series B: 10 mM (DHQD)2PHAL-type CSPA in MTBE as stationary-phase solvent. The amounts of racemic
dichlorprop injected are indicated in the framed boxes. r refers to the molar ratio of the loaded amount of racemic dichlorprop and the total
amount of CSPA present in the CPC rotor.
6.28; phosphate buffer kR ) 0.36, kS ) 1.27), indicating less efficient
mass transport properties. In context with preparative applications,
extensive retention compromises the overall productivity of the
process by limiting throughput and increasing solvent consump-
tion. As a consequence, solvent combinations with acetate buffers
were excluded from further studies.
With the phosphate-buffered mobile phase, both the MIBK and
the MTBE stationary-phase system showed favorably fast elution
behavior. Thus, clean baseline enantiomer separation could be
achieved with sample loads corresponding to a molar analyte/ CS
ratio r ) 0.5 (see Figure 5, A1 and B1), indicating excellent
preparative capacities for both systems.
combination, minor peak overlap was observed, resulting in a few
mixed fractions containing 14% of the injected sample.
Finally, the sample loading was further increased to r ) 2.0
(Figure 5, A3 and B3). Achieving baseline separation under these
loading conditions would realize the economically highly desirable
scenario of “total CS exploitation”, a situation in which every single
CSPA molecule physically effects the separation of a pair of
enantiomer molecules. Enantiomer separations were incomplete
with both solvent systems under these conditions; however, peak
overlap was moderate (38 and 21%, respectively), permitting
isolation of major fractions of the injected racemic sample as pure
enantiomers.
The excellent preparative capacity motivated us to probe the
CPC performance under equimolar sample loading conditions (r
) 1.0, Figure 5, A2 and B2). Much to our delight, the phosphate/
MTBE system provided, even under these challenging conditions,
complete enantiomer separation. With the phosphate/ MIBK
Enantiomer Separation of Dichlorprop on CSP 1 . The
exceptional preparative performance observed with the (DHQD)2-
PHAL-type CSPA under CPC conditions motivated us to carry out
a comparative HPLC study, employing the corresponding CSP
with the parent cinchona-type CS covalently linked onto the
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004 5845

Figure 6. Analytical HPLC enantiomer separation of dichlorprop on the (DHQD)2PHAL-type CSP under optimized mobile-phase conditions:
column (150 × 4 mm i.d.); mobile phase methanol/acetic acid/ammonium acetate (92:2:0.5, v/v/w); flow rate 1 mL/min; UV detection 254 nm;
column temperature 25°C; injected sample amount 20 µg.
surface of 5-µm spherical silica particles. The (DHQD)2PHAL-type
CSP was available as an analytical column (150 × 4 mm i.d.) only,
containing 1.10 g of CSP with a CS loading level of 82 µmol/ g.
The mobile-phase conditions employed in CPC were found to be
incompatible with HPLC separation due to the extreme analyte
retention (kR > 10) and prohibitively high back pressure (>150
bar at a flow rate of 1 mL/ min) observed. Optimization of the
mobile-phase conditions identified a mixture of ammonium acetate
and acetic acid in methanol as a favorable alternative. This mobile
phase provided the benefits of excellent enantioselectivity (RHPLC
> 15; see Figure 6), along with acceptably short retention times
and workable back pressure (60 bar).
Employing these polar organic mode conditions, the (DHQD)2-
PHAL-type CSP was challenged with increasing amounts of
racemic dichlorprop corresponding to r values of 0.25, 0.50, and
1.0, respectively. The collected product fractions were analyzed
with respect to enantiomer purity. The chromatograms of these
preparative runs, and the corresponding quality control runs, are
depicted in Figure 7. Inspection of the elution profiles of the
preparative runs reveals a particular behavior in terms of peak
shapes: While sample overloading induced strong peak distortion
for the more retained S-enantiomer, the peak quality and sharp-
ness of the first-eluted enantiomer was almost unaffected. Baseline
and near-baseline enantiomer separations could easily be achieved
for 4.9 mg (r ) 0.25). and 9.7 mg (r ) 0.50) racemic dichlorprop.
For the highest sample load tested, i.e., 19.5 mg of racemic
dichlorprop corresponding to r ) 1.0, the peaks merged com-
pletely into a single unstructured band, suspending the possibility
for UV signal-guided fractionation. However, due to the favorable
“peak compression” of the first-eluted R-enantiomer, time-
controlled peak fractionation still allowed isolation of enantiomeri-
cally highly enriched product fractions. The analytical assessment
of the enantiomeric excess values of the collected fractions showed
ee > 98% for the first-eluted R-enantiomer and >96% for the more
retained S-enantiomer. In the case of the highest sample loading,
however, poor peak resolution resulted in a somewhat reduced
enantiomer purity for the less retained enantiomer (ee > 92%).
P reparative Capacity of CP C versus HP LC. The application
of cogeneric (DHQD)2PHAL-type CS systems for HPLCsas well
as CPC-based enantiomer separationsoffers the opportunity to
assess merits and limitations of these complementary techniques.
For this purpose, an overview of the corresponding operational
parameters, specific productivities, and mobile-phase consump-
tions for CPC and HPLC runs performed with a r ) 1.0 is given
in Table 2. The data given for CPC refer to run B2 in Figure 5,
and for HPLC, to run C in Figure 7.
The most significant finding emerging from this comparative
study is that CSPA-mediated CPC shows a level of CS utilization
comparable to that of CSP-based HPLC. Evidently, both chro-
matographic technologies allow exceptionally high levels of
loadability, with complete dichlorprop enantiomer separations still
achievable with racemic sample amounts equimolar to the
incorporated CS (r ) 1.0).
Significant differences in the performance characteristics of
CPC and HPLC, however, become obvious when comparing
productivity-related figures. Considering the different molar
amounts of (DHQD)2PHAL employed in CPC and HPLC experi-
ments, the productivity data and mobile-phase requirements
observed on the analytical column were appropriately normalized
with respect to the effective CS concentration, according to well-
established linear scale-up rules.28 Thus, relating the mass of
resolved racemate to the molar amount of applied CS and time,
HPLC should be capable of separating roughly 7.8 kg of racemic
dichlorprop per mol of CS per day, while CSPA-mediated CPC
can only separate 2.1 kg per mol of CS per day.
Inspection of the operational parameters listed in Table 2
identifies the mobile-phase flow rate as the major productivity-
limiting factor. The rather slow flow rate (3 mL/ min) employed
in CPC leads to extended run times (160 min) compared to HPLC
(40 min) and thus to decreased throughput capacity. The pos-
sibility to enhance CPC productivity by increasing flow rate seems
to be limited as this may reduce the volume of stationary phase
retained within the CPC column and thus compromise preparative
enantiomer separation capacity. In addition, higher flow rates may
increase the back pressure beyond the instrumental limit (80 bar).
(28) Heuer, C.; Hugo, P.; Mann, G.; Seidel-Morgenstern, A., J. Chromatogr., A
1 9 9 6 , 752, 19-29.
5846 Analytical Chemistry, Vol. 76, No. 19, October 1, 2004

Figure 7. Preparative HPLC enantiomer separation runs performed with the (DHQD)2PHAL-type CSP with increasing sample loads of racemic
dichlorprop. The fractionation times are indicated by the broken lines. The chromatograms depicted in the right column give the enantiomeric
ratios of the collected fractions, obtained with a different type of CSP (see Experimental Section). The amounts of racemic dichlorprop injected
are indicated in the framed boxes. r refers to the molar ratio of the loaded amount of racemic dichlorprop and the total amount of immobilized
(DHQD)2PHAL-type CS present in the column. For experimental conditions see information given for Figure 6.
With respect to solvent consumption, the developed CPC
separation protocol is much more economical than HPLC. The
mobile-phase consumption of CPC amounts to 1.30 L/ g of resolved
racemate, while under HPLC conditions, 2.10 L/ g of resolved
racemate is required. Apart from its lower solvent consumption,
the CPC separation protocol also appears to be more attractive
from an environmental viewpoint. In contrast to HPLC, which
operates with a mobile phase largely composed of methanol, CPC
enantiomer separation can be achieved with an aqueous phosphate
buffer mobile phase. Thus, with CPC, the utilization of organic
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004 5847

Table 2. Comparison of the Preparative Performance Characteristics of Developed CPC and HPLC-Based
Enantiomer Separation Protocols for Dichlorpropa
CPC
HPLC
stationary phase
total amount of CS (mmol)
mobile phase
10 mM CSPA in MTBE (155 mL)
1.55
100 mM sodium phosphate, pH 8.0
CSP grafted with 82 µmol of CS g-1 (1.10 g)
0.092
methanol/ acetic acid/ ammonium acetate
92/ 2/ 0.5 (v/ v/ w)
racemate resolved per run (mg/ mmol)
r
366/ 1.55
1.0
19.5/ 0.083
1.0
ee of isolated S/ R-enantiomer (%)
recovery (%)
98/ 98
>95
480
160
2.1
98/ 92
>90
mobile-phase consumption per run (mL)
run time (min)
40
40
productivity (kg of racemate d-1 mol-1 CS)
solvent consumption (L g-1 racemate)
7.8b
1.30
2.10b
a
r, molar ratio of the racemic dichlorprop loaded and the chiral selector incorporated in the separation system. b These figures were calculated
for a column comprising an equivalent amount of (DHQD)2PHAL-type CSP, according to well-established linear HPLC scale-up rules.28
solvents is reduced to a minimum (i.e., the amount of MTBE used
as stationary-phase solvent), allowing for significant savings in
chemicals and waste stream management.
suitable CSPAs and the technically still immature state of large-
scale CPC instrumentation. Efficient CSPAs for preparative CPC
application must exhibit high levels of target-specific enantio-
selectivity, excellent solubility, and complete retention in the
stationary-phase media and readily tunable binding affinity to
control analyte distribution behavior. Concerning instrumentation,
CPC equipment is less readily available and significantly more
expensive as compared to HPLC systems. Adaptation of CPC
enantiomer separation to industrial-scale application appears to
be more challenging than for HPLC, particularly in view of
engineering continuously operating process configurations.
Nevertheless, the impressive preparative performance of the
model CPC enantiomer separation protocol presented in this paper
justifies further research in these directions. The possibility to
achieve efficient enantiomer resolution with support-free CSs may
help master preparative separations under conditions incompatible
with conventional silica-supported CSPs, for example, with salt-
rich aqueous mobile phases at extreme pH values. The excellent
compatibility of CPC with aqueous mobile phases may also provide
unique opportunities for the implementation of low-cost and
environmentally friendly enantiomer separation schemes.
We wish to point out that the above figures were calculated
under ideal “linear upscaling conditions”. In practice, upscaling
efforts may be associated with severe limitations. These may
include, for example, the need of handling excessively large
stationary- and mobile-phase volumes, the limited availability of
large-scale CPC instruments, and the intrinsic loss of HPLC
efficiency on changing to CSP materials of larger particle size.
Thus, the productivity data outlined in Table 2 reflect “best case
scenarios”, rather than real world situations.
CONCLUSIONS
A support-free CPC enantiomer separation protocol for dichlo-
rprop was developed utilizing a purposefully designed CSPA
derived from (DHQD)2PHAL. The selection of solvent systems
suitable as stationary/ mobile phases for CPC application was
guided by enantioselective liquid-liquid extraction experiments
in the presence of the (DHQD)2PHAL-type CSPA. A system
consisting of 10 mM CSPA in MTBE/ sodium phosphate buffer
(pH 8.0) was identified to provide the benefits of high enantio-
selectivity, well-balanced distribution behavior, and satisfactory
system stability. Preparative CPC runs performed with this solvent
combination showed excellent loadability, producing clean base-
line separations with sample loads up to amounts equimolar to
the CSPA present in the stationary phase. Comparison with the
HPLC performance characteristics of a silica-supported version
of (DHQD)2PHAL revealed that CPC offers a comparably high
preparative loadability at significantly reduced solvent consump-
tion. CPC, however, gave lower specific productivity, mainly
imposed by instrument-inherent flow rate restrictions. These
limitations in productivity may be overcome by (i) enhancing
system loadability by increasing the CSPA concentration in the
stationary phase and (ii) improving peak capacity through pH zone
refining.11,16
ACKNOWLEDGMENT
This project was supported by grants from the Austrian
Research Fund (FWF) Projects P14179-CHE and P16300-B1; the
European Community under the Industrial & Materials Technolo-
gies Program (Brite EuRam III), Project BE 96-3159; the Austrian
Academic Exchange Service (OEAD); and the Ministerio de
Ciencia y Tecnologia (Spain), Integrated Actions Austria-Spain,
Projects 13/ 2003 and HU2002-0025. Our thanks to Dr. Pilar Franco
for critical discussions and Dr. Kevin Schug for providing
competent linguistic advice. Dedicated to Prof. Domenico Misiti
on the occasion of his 70th birthday.
Received for review May 20, 2004. Accepted July 18,
2004.
A serious drawback currently prohibiting the broad use of CPC
technology for preparative enantiomer separation is the lack of
AC040102Y
5848 Analytical Chemistry, Vol. 76, No. 19, October 1, 2004