Effect of achiral support on the resolution of tetramisole by supercritical fluid extraction

Article abstract of DOI:10.1016/S0957-4166(02)00337-3

The enantiomers of tetramisole were produced by partial diastereomeric salt formation with O,O′-dibenzoyl-(2R,3R)-tartaric acid monohydrate and subsequent supercritical fluid extraction (SFE) of the unreacted enantiomers in the presence of an achiral support. The effect of the activated carbon and Perfil 100 on the efficiency of chiral separation were studied. The kinetics of the process was found to be an important factor affecting enantioselectivity. When the parameters were properly set, much better resolution efficiency (F) and higher enantiomeric purity were achieved than in the equilibrium. The presence of Perfil 100 and activated carbon caused an increase as high as 53 and 84% in F, respectively, compared to that achieved without using any supporting material. Thus during the development of a resolution procedure, beside the proper resolution agent, optimised molar ratio and other parameters, the selection of achiral addict(s) may be also an important point.

Full text of DOI:10.1016/S0957-4166(02)00337-3

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 13 (2002) 1429–1434
Effect of achiral support on the resolution of tetramisole by
supercritical fluid extraction
Edit Sze´kely,a,* Be´la Sima´ndi,a Krisztina La´szlo´,b Eleme´r Fogassy,c Gyo¨rgy Pokold and
Ildiko´ Kmecza
aDepartment of Chemical Engineering, Budapest University of Technology and Economics, H-1521 Budapest, Hungary
bDepartment of Physical Chemistry, Budapest University of Technology and Economics, H-1521 Budapest, Hungary
cDepartment of Organic Chemical Technology, Budapest University of Technology and Economics, H-1521 Budapest, Hungary
dInstitute of General and Analytical Chemistry, Budapest University of Technology and Economics, H-1521 Budapest, Hungary
Received 21 June 2002; accepted 27 June 2002
Abstract—The enantiomers of tetramisole were produced by partial diastereomeric salt formation with O,O%-dibenzoyl-(2R,3R)-
tartaric acid monohydrate and subsequent supercritical fluid extraction (SFE) of the unreacted enantiomers in the presence of an
achiral support. The effect of the activated carbon and Perfil 100™ on the efficiency of chiral separation were studied. The kinetics
of the process was found to be an important factor affecting enantioselectivity. When the parameters were properly set, much
better resolution efficiency (F) and higher enantiomeric purity were achieved than in the equilibrium. The presence of Perfil 100™
and activated carbon caused an increase as high as 53 and 84% in F, respectively, compared to that achieved without using any
support. Thus during the development of a resolution procedure, beside the proper resolution agent, optimised molar ratio and
other parameters, the selection of achiral additive(s) may be also an important point. © 2002 Elsevier Science Ltd. All rights
reserved.
1. Introduction
The development of a widely applicable method
requires a deep understanding of the influencing fac-
tors. The effect of the molar ratio, extraction pressure
and temperature on the resolution of different organic
acids3 and bases has been reported recently.7 Herein the
effect of the achiral support is examined.
According to the US Food and Drug Administration’s
policy statement, marketing of new chiral pharmaceuti-
cals has been allowed only as pure enantiomers from
the beginning of 1992, except when the therapeutic
properties of the racemic mixture are better.1 In many
cases these enantiopure drugs are produced by asym-
metric synthesis followed by chiral resolution.
2. Materials and methods
In recent years growing interest has been shown in new
and environmentally friendly resolution processes.2
Beside membrane separation, procedures using super-
critical fluids have also become promising on industrial
scale.3,4 Our group developed a novel technique based
on partial diastereomeric salt formation followed by
supercritical fluid extraction (SFE) of the free enan-
tiomers.3 The method was successfully applied to the
resolution of racemic organic acids and bases. Mean-
while, Bauza et al. also proved the feasibility of SFE of
enantiomers for analytical purposes.5,6
Racemic tetramisole (6-phenyl-2,3,5,6-tetrahydro-imi-
dazo-[2,1b]-thiazole) was prepared in our laboratory
and checked analytically by HPLC prior use. It was
found to be of 99.9% purity. The resolving agent (O,O%-
dibenzoyl-(2R,3R)-tartaric acid=DBTA) and all other
reagents and solvents were purchased from Reanal
Ltd., Budapest.
Perfil 100™ (expanded and milled perlite for use as a
filtering-aid) was obtained from Baumit Co.
(Budapest). The activated carbon applied was of medi-
cal grade, commercially available in pharmacies. The
specific surface area (Sa,BET) and the total pore volume
values are listed in Table 1.
* Corresponding author. Tel. +36-1-4632246; fax.: +36-1-4633197;
e-mail: sz-edit@vhsz.bme.hu
0957-4166/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(02)00337-3

1430
E. Sze´kely et al. / Tetrahedron: Asymmetry 13 (2002) 1429–1434
Table 1. Specification of the supports
Sa,BET (m2/g)
Total pore volumea (cm3/g)
Sa,aval (m2/g)
a
b
Support
Active carbon
Perfil 100™
1.72×103
2.89
1.02
82.9
2.89
5.02×10−3
a Determined from low temperature N2-adsorption isotherm (AUTOSORB, QUANTACHROME, USA).
b
,
It was supposed that pores below a diameter of 10 A are not available for the molecules studied. In the case of the highly porous activated
carbon sample, the Sa,aval value (available specific surface area) was derived from the t-plot of the nitrogen adsorption data.8
The CO2 used in this study was 99.5% (w/w) pure and
supplied by Messer Griesheim Hungaria Ltd.
(Budapest).
Optical rotation (h) measurements were carried out on
a Perkin–Elmer 241 polarimeter, and the enantiomeric
excess values were determined from a values according
to prior calibration.
The resolution process is shown in Scheme 1. A descrip-
tion of the extractor unit is given in a previous paper.9
In the sample preparation step the resolving agent and
racemic tetramisole, in 0.25:1 molar ratio, were dis-
solved in an appropriate solvent, in this case methanol,
and then the support was added to the clear solution.
After about 5 min the solvent was evaporated in vac-
uum at 40°C and the mixture was dried at 40°C for 24
h. The extractor vessel was filled with this dry material
and then the SFE was performed at 160 bar and 39°C.
Adsorption isotherms were derived from the initial and
equilibrium concentration of the liquid phase measured
by a UV–Vis spectrophotometer (Uvikon Kontron,
Zu¨rich) at u=224 nm.
TG and DSC measurements were carried out on TA
Instruments 2050 TGA V5.3C and TA Instruments
2920 MDSC V2.4F, respectively. Argon atmosphere
was used in a 10 L/h flow rate. The mass of samples
was 4–5 mg.
3. Results
3.1. Extraction curve
A typical extraction curve is shown in Fig. 1. The
extraction yields (Ye=mass of extract/mass of tetra-
misole in the sample) are plotted versus the specific
solvent mass (w=mass of CO2 passed through the
extractor/mass of tetramisole in the prepared sample).
The Ye=a(1−e−kw) regression function suggested by
Brunner10 fits very well to the measured data. The
quality of the support had no significant effect on the
estimated parameters (a, k).
3.2. Adsorption of tetramisole and DBTA on Perfil
100™ and activated carbon
The adsorption isotherms were derived as na=V(c0−
c)/m, where c0 and c are the initial and equilibrium
concentration, V is the volume of the solution and m
is the mass of the support. The na versus c function
was fitted according to the Langmuir model.11 The
deduced parameters, the monolayer capacity (nm) and
the adsorption constant (K) are listed in Table 2. The
K values are characteristic to the adsorption ‘affinity’.
The low K values mean a weak adsorption. The
strongest among them is the tetramisole/activated car-
bon interaction. Comparing the mrel,m and mrel,res val-
ues, it is obvious that in the experiments we were far
above the monolayer capacity of the support, as the
mrel,m is at least one order of magnitude smaller than
mrel,res
.
Scheme 1. Resolution by supercritical fluid extraction.

E. Sze´kely et al. / Tetrahedron: Asymmetry 13 (2002) 1429–1434
1431
various phases was established. The DSC curves of
tetramisole samples are presented in Fig. 2. The results
are summarised in Table 3. The presence of the sup-
ports does not affect the melting point of the crystalline
tetramisole, but the shape and the integral of the peaks
are influenced by both the Perfil 100™ and the acti-
vated carbon. The deviation of the reported Qm/Qcal
values from 1 suggests, that a part of the tetramisole is
present in non-crystalline form. The decrease of crys-
talline tetramisole in equilibrium samples was proved
by the X-ray diffraction measurements. [HCG-4 Carl
Zeiss (Jena)]. It could be concluded, that either the
presence of the support, or the long contact time due to
the slow evaporation, suppress crystal formation.
Figure 1. A typical extraction curve; R2=0.999.
A significant effect was also observed in the case of the
DBTA/support systems. No peak was detected in the
temperature range studied in the case of DBTA on
activated carbon. When Perfil 100™ was added, an
endothermic peak was detected in the range of the
decomposition temperature of the DBTA.
3.3. Thermal and X-ray diffraction analysis of the
equilibrium samples
Samples for thermal analysis were prepared by slow
evaporation of the solvent at 40°C. We supposed that
during the slow evaporation an equilibrium state of the
Table 2. Parameters fitted from the linear Langmuir plot11
a
b
Support
Adsorbate
nm (mmol/g)
K (dm3/mg)
R2
mrel,m (mg/m2)
mrel,res (mg/m2)
Activated carbon
Perfil 100™
Tetramisole
DBTA
Tetramisole
DBTA
0.0090
0.3129
0.0026
1.247
0.2742
0.0698
0.997
0.998
0.994
0.022
0.589
0.184
10–350
4.6–160
100–1100
46–760
Adsorption was below the detection limit
a mrel,m is the mass of tetramisole or DBTA/surface of the support at monolayer adsorption.
b mrel,res is the mass of tetramisole or DBTA/surface of the support at resolution experiments.
Figure 2. DSC curves of tetramisole samples; (1) pure tetramisole; (2) tetramisole on Perfil 100™; mrel=0.373, (3) tetramisole on
activated carbon, mrel=0.012.

1432
E. Sze´kely et al. / Tetrahedron: Asymmetry 13 (2002) 1429–1434
Table 3. DSC measurement results of tetramisole equilibrium samples
mrel (g/m2)
Qcal (J/g sample)
Qm (J/g sample)
Qm/Qcal
a
b
c
Support
None (pure)
Activated carbon
Perfil 100™
101.2
50.73
48.67
101.2
8.31
41.36
1
0.012
0.373
0.164
0.850
a Qcal=Qpure/(1+mrelSa,aval) is the calculated melting enthalpy from the amount of tetramisole. Qpure is the melting enthalpy of tetramisole.
b Qm is the measured melting enthalpy.
c By the ratio Qm/Qcal the crystalline ratio of tetramisole (mass of crystalline tetramisole versus mass of tetramisole in the sample) was estimated.
3.4. Effect of the support on chiral recognition
these results it is not likely, that the adsorption of
tetramisole results in a change of F/F0.
To evaluate the resolution efficiency, a merged F
parameter was used (F=ꢀYeeeeꢀ+ꢀYreerꢀ, where Yr is the
yield of tetramisole in raffinate, eee and eer are enan-
tiomeric excess values of extract and raffinate, respec-
tively). F may vary between 0 and 1. According to the
mass balance the terms ꢀYeeeeꢀ and ꢀYreerꢀ should be
equal. This new parameter provides a better estimation
of the resolution efficiency of the extract and raffinate,
and the degree of freedom becomes higher at statistical
evaluation.
Experimental results are shown in Figs. 3 and 4. The
F/F0 term shows the relative effect of the support on
resolution efficiency. F0 is the resolution efficiency
obtained without using any support (F0=0.341
0.0236).
Figure 3. Effect of the specific mass of tetramisole on perfil
(mrel) on resolution.
An increase in the mrel values leads to a continuous
increase of F/F0 up to the limits 1.53 and 1.84 using
activated carbon and Perfil 100™, respectively.
4. Discussion
After observing the interesting effects of the support,
our first aim was to establish whether these were caused
by the change in diastereomeric equilibrium, or resulted
from kinetic control of the resolution. To test this, a
sample was stirred with activated carbon (mrel=0.094
g/m2) for 1 h prior to evaporation of the solvent (to
reach the equilibrium). Much lower selectivity was
achieved (F/F0=0.998), than calculated from the
regression function (F/F0=1.755). Surprisingly the
measured F was almost the same as obtained without
using any support. These results suggest that the higher
resolution efficiencies at higher mrel values are caused
by non-equilibrium, kinetic process and the decrease of
the crystalline phase at quasi-equilibrium state (see part
3.3) suppress the chiral recognition as well.
Figure 4. Effect of the specific mass of tetramisole on acti-
vated carbon (mrel) on resolution.
Table 4. Effect of sample preparation on the resolution
efficiency. The solution of the given compound was stirred
1 h with the activated carbon then the solution of the
other compound was added only 5 min before the evapo-
ration
Almost the same results were achieved, when only the
solution of tetramisole or only the solution of DBTA
were stirred for an hour with activated carbon and the
solution of the other compound was added to the
mixture just before the evaporation step (see Table 4).
Although tetramisole showed higher adsorption
‘affinity’ than DBTA (see Section 3.2), according to
Stirred
compound
mrel (g/m2)
F/F0 measured F/F0
calculateda
Tetramisole
DBTA
0.095
0.094
0.902
0.909
1.777
1.771
a From the regression function in Fig. 4.

E. Sze´kely et al. / Tetrahedron: Asymmetry 13 (2002) 1429–1434
1433
According to the adsorption data (K values) none of
6. Experimental
the components studied show a strong adsorption inter-
action. Thus, among the components present in the
prepared sample, the adsorption of a diastereomeric
complex may be the strongest, due to its higher molecu-
lar mass and lower solubility. Since the solubility of
(S)-(−)-tetramisole–DBTA diastereomer in methanol is
much lower than that of (R)-(+)-tetramisole–DBTA
diastereomer, the (S)-(−)-tetramisole-containing dia-
stereomer has the highest probability to adsorb. All
these factors suggest, that the higher chiral separation
at higher mrel values is due to simple competition of the
components for adsorption sites. The adsorbed
molecule may induce its own crystallisation, and the
(S)-(−)-tetramisole-DBTA diastereomer is the most
likely to start this. When there is a huge amount of
bonding sites, all possible molecules in the solution
adsorb and start to crystallise, lowering the chiral dif-
ferentiation toward zero.
Racemic tetramisole (1.50 g, 7.40 mmol) and O,O%-
dibenzoyl-(2R,3R)-tartaric acid monohydrate (0.69 g,
1.85 mmol) were dissolved in methanol (20 mL). Active
carbon (0.05g) was added to the clear solution and the
solvent was evaporated at 40°C under reduced pressure
(water vacuum). The residue was extracted with super-
critical carbon dioxide (160 bar, 39°C, 900 g CO2). The
(R)-(+)-enantiomer (0.64 g) ([h]2D0=+72.6 (c 5, MeOH))
was recovered from the extract. The raffinate was sus-
pended in water (20 mL) and NH4OH was added (25
w/w%) to pH 9.5. The reaction mixture was stirred for
30 min at 10°C and the (S)-(−)-enantiomer. (0.66 g) was
isolated by filtration ([h]2D0=−78.8 (c 5, MeOH)).
A description of the enantiomeric enrichment by partial
salt formation with HCl and subsequent SFE was given
in a previous paper.7 Scheme 2 shows the whole resolu-
tion process with results of the given experimental. The
higher enantiomeric excess achieved in the first step,
with addition of activated carbon, results in higher total
yield values (Ytotal=mass of pure enantiomer/mass of
racemic tetramisole in the initial sample), see Table 5.
5. Conclusions
This resolution procedure by partial diastereomeric salt
formation followed by supercritical fluid extraction of
the unreacted enantiomers is a form of kinetic resolu-
tion. When process parameters are properly set, much
better resolution efficiency and higher enantiomeric
purity can be achieved than at the equilibrium.
Acknowledgements
In the development of a resolution procedure, besides
choosing an appropriate resolving agent, molar ratio
and other parameters, the selection of achiral addi-
tive(s) is also an important aspect.
This research work was supported by Ministry of Edu-
cation within a German-Hungarian Cooperation Pro-
ject (D-3/99).
Scheme 2. The whole resolution process of tetramisole using activated carbon at mrel=0.347 and 5 min sample preparation time.
Table 5. Comparison of yields of whole resolution procedures
Method
First step, mixture with
First step, mixture with
Second step Ytotal
(R)-(+) in excess
(S)-(−) in excess
Y
ee
Y
ee
(R)-(+)
(S)-(−)
Optimised SFE8
0.404
0.391
0.472
0.427
0.600
0.346
0.582
0.680
0.421
0.415
0.428
0.440
0.540
0.512
0.547
0.737
0.189
0.158
0.224
0.221
0.211
0.188
0.199
0.269
SFE, without support
SFE, perfil
SFE, activated carbona
a See Scheme 2.

1434
E. Sze´kely et al. / Tetrahedron: Asymmetry 13 (2002) 1429–1434
References
1. FDA Chirality 1992, 4, 338–340.
1999, 391, 253–256.
6. Bauza, R.; R´ıos, A.; Valca´rcel, M. Anal. Chim. Acta
2001, 450, 1–11.
7. Keszei, S.; Sima´ndi, B.; Sze´kely, E.; Fogassy, E.; Sawin-
sky, J.; Keme´ny, S. Tetrahedron: Asymmetry 1998, 10,
1275–1281.
8. La´szlo´, K.; Szucs, A. Carbon 2001, 39 (13), 1945–1953.
9. Sima´ndi, B.; Keszei, S.; Fogassy, E.; Keme´ny, S.; Sawin-
sky, J. J. Supcrit. Fluids 1998, 13, 331–336.
10. Brunner, G. Ber. Bunsenges. Phys. Chem. 1984, 88, 887–
891.
2. de Haan, B. A.; Sima´ndi, B. In Extraction Technology for
the Separation of Optical Isomers; Marcus, Y.; SenGupta,
A. K.; Marinsky, J. A., Eds. Ion exchange and solvent
extraction; Marcel Dekker: New York, 2002; Vol. 15, pp.
255–294.
3. Fogassy, E.; Acs, M.; Szili, T.; Sima´ndi, B.; Sawinsky, J.
Tetrahedron Lett. 1994, 35, 257–260.
4. Kordikowski, A.; York, P.; Latham, D. J. Pharm. Sci.
´
11. Kipling, J. J. Adsorption from Solutions of Non-elec-
trolytes; Academic Press: London, 1965.
1999, 88, 786–791.
5. Bauza, R.; R´ıos, A.; Valca´rcel, M. Anal. Chim. Acta