Microsoft Word - Contents.doc

HUNGARIAN JOURNAL
OF INDUSTRIAL CHEMISTRY

VESZPRÉM
Vol. 34. pp. 15-20 (2006)

QUASI-CONTINUOUS ELUTION CHROMATOGRAPHIC PURIFICATION
OF A STEROID ACTIVE COMPOUND

K. TEMESVÁRI1, , A. ARANYI1, Z. HORVÁTH1, M. NAGY2, T. SZÁNYA2 and L. HANÁK2

1Gedeon Richter Ltd., H-1475 Budapest 10. PO Box 27. HUNGARY;
E-Mail: k.temesvari@richter.hu

2University of Pannon, Department of Chemical Engineering POB 158, Veszprém, H-8201, HUNGARY

A pilot-scale preparative batch elution chromatographic separation of a steroid active compound from its impurities was
developed and is applied at our company. The process consists of the following steps: (1.) injection of the crude sample
solved in the eluent onto the chromatographic column, (2.) elution of the less retained impurities with the eluent, (3.)
elution of the pure fraction of the active compound (main fraction) with the eluent, (4.) back-flushing the column with the
more polar solvent component of the eluent to remove the strongly retained compounds, (5.) conditioning the column
with the eluent to prepare for the next injection. The above mentioned five-step chromatographic process can be realized
quasi-continuously by applying the Simulated Moving Bed (SMB) instrument. The subsequent steps can be carried out in
parallel in the zones of the SMB unit, which are independent from each other in this special case.

In our present paper the elaborated quasi-continuous elution chromatographic separation is shown using the following
instrument set up and experimental conditions: a KNAUER CSEP 9116 laboratory-scale SMB unit with five KNAUER
K-501 HPLC pumps and eight columns (column dimension: 250X16 mm I.D.). The columns were packed in our
laboratory by dry-packing method using vibration, the applied packing material was UETIKON C-490 15-35 μm Si gel.
The appropriate eluent was methylene chloride/ethyl acetate mixture and as back-flushing solvent, pure ethyl acetate was
used. After the elaboration of quasi-continuous elution chromatographic purification, the intensification of the process
was carried out. The specific capacity parameters of the best quasi-continuous experiments were compared with the
performance parameters of the current batch pilot-scale separation of the compound.

Keywords: quasi-continuous, pilot-scale, preparative chromatographic separation, steroid, Simulated Moving Bed (SMB)

Introduction

In pharmaceutical industry there is a strict demand on
the quality of active compounds. In most cases the
amount of all impurities must be under 0.5-1 %, the
maximum level of individual impurities had to be under
0.1 %. In many cases these demands can not be fulfil
with traditional techniques, such as distillation or
crystallization. Pilot- and process-scale preparative
chromatographic processes can solve these difficult
separation problems. [1]

The preparative chromatographic processes can be
divided into batch and continuous methods, depending
on the actual separation task. In some cases these
techniques can compete with each other. [2-5]

The application areas of continuous counter-current
separation methods are chiral, isomer, or other two-
component (or pseudo two-component) separation
problems. [6-11]

The Simulated Moving Bed (SMB) technology is
the up-to-date solution for the practical realization of
continuous, counter-current preparative chromatographic
processes. It was invented and patented in the late
1950’s and was mainly applied in the petrochemical and
sugar industries for large-scale separations. SMB has

been introduced in the pharmaceutical industry since the
1990’s. [12]

The principle of the technique has been described in
numerous publications [13-15]. The essence of the
process is that instead of the continuous transportation
of the particulate solid phase, which leads to a lot of
difficulties, (bad efficiency because of high HETP and
mechanic problems with transportation of the solid
phase) the moving of the solid phase in packed columns
is simulated. This simulated moving can be solved
either by switching the inlets and outlets co-current with
the liquid flow direction, or by the real rotation of the
packed columns, connected to special distributing valves.
(Fig. 1.)

In the SMB process the strong (S, more retained)
and weak (W, less retained) components of the feed (F)
of concentration cF,S and cF,W are separated into two
outlet flows. In the extract (E, cE,S, cE,W) the better
adsorbing strong, while in the raffinate (R, cR,S, cR,W) the
less adsorbing weak components are enriched,
respectively. (Fig. 2.)

In purification tasks there are usually one (or some
more) main compounds in the presence of numerous
impurities of small amount. In such cases SMB is
generally not applicable, mainly preparative batch elution
techniques are used. [16-19]

Fig. 1: Conventional four-zone SMB unit with eight

columns

The conventional SMB instrument set-up can be
utilized to solve a batch elution chromatographic
purification task quasi-continuously, if an astute
assembling mode of the available unit is applied.

In this present study the quasi-continuous
implementation of a current pilot-scale batch elution
purification of a steroid active compound is shown. The
process consists of the following steps: (1.) injection of
the crude sample solved in the eluent onto the
chromatographic column, (2.) elution of the less retained
impurities with the eluent, (3.) elution of the pure
fraction of the active compound (main fraction) with the
eluent, (4.) back-flushing the column with the more polar
solvent component of the eluent to remove the strongly
retained compounds, (5.) conditioning the column with
the eluent to prepare for the next injection.

The subsequent steps can be carried out in parallel in
the zones of the SMB unit, which are independent from
each other in this special case.

The Implementation of Quasi-Continuous Process
Starting from Usual SMB Instrument Set-Up

The conventional SMB unit has four zones as can be
seen in Fig.1. and Fig.2.

Fig. 2.: One of the applied practical solutions of the

Simulated Moving Bed

It is possible to carry out the above-mentioned actual
five-step batch elution chromatographic purification
procedure quasi-continuously in these zones with the
help of the rotating switching valve of the SMB
instrument. As in our special case the zones are
independent from each other, they can be decomposed
to six parts and the flow rate can be reversed in zone I,
as Fig. 3 shows.

Fig. 3.: Quasi-continuous operation with the help of

rotating switching valve of the SMB unit

The minimum number of columns required in the
process is eight and in every second position an empty
short capillary is inserted in the unit. One column and
one empty position move together to the next position
in every switching.

With the knowledge of the solvent volumes necessary
to the subsequent steps of the separation, the switching
time of the rotating SMB valve and flow rates of the
pumps in the zones can be calculated.

The main steps of the elaboration of the quasi-
continuous process are:

(A.) Preliminary elution experiment on one of the
eight 250X16 mm I.D. columns used in the SMB unit
for quasi-continuous elution separation, for the
determination of the solvent volumes necessary to the
steps of the separation. From these data the switching
time of the rotating SMB valve and the pump flow rates
can be calculated. (B.) Carrying out a quasi-continuous
experiment applying the calculated parameters. (C.)
Refining the process parameters in experimental way.
(D.) Improvement the productivity by decreasing the
switching time and proportionally increasing the flow
rates of the pumps at the same time.

Experimental

The problem under investigation was the solution of
quasi-continuous elution chromatographic purification
of a steroid active compound, produced by Gedeon
Richter Ltd. The starting point of the investigation was
the pilot-plant batch elution chromatographic separation
of the compound from its impurities, followed by a
crystallisation step (not published here). This procedure
has been applied successfully for years. The purity of
the crude material was 94 % for the active compound.
The goal was to produce the active compound with a
purity of >99 %, the maximum level of individual
impurities had to be under 0.3 %.

Chemicals

UETIKON C-490 15-35 μm Si gel was used as the
stationary phase, it was purchased from ZEOCHEM
AG (Switzerland).

The mobile phase was methylene chloride/ethyl
acetate mixture. The back-flushing solvent was pure
ethyl acetate. Solvents were purchased from Merck
KgaA (Darmstadt, Germany).

Instrumentation

Preliminary Elution Experiment

The applied instrument consisted of a KNAUER K-501
HPLC pump, the volumetric flow rate was set to
5 mL min-1, one of the eight 250X16 mm I.D. columns
used in the SMB unit for quasi-continuous elution
separation, packed with UETIKON C-490 15-35 μm Si
gel packing material in our laboratory by dry-packing
method using vibration, and a KNAUER UV/VIS filter
photometer. The wavelength of the measurement was
254 nm. For the purpose of recording the chromatogram
a Radelkis OH-850 potentiometric recorder was applied.
The temperature was set to 25 ˚C.

Quasi-Continuous Experiments

A laboratory-scale SMB unit (KNAUER CSEP 9116)
was used for the quasi-continuous elution experiments
with five KNAUER K-501 HPLC pumps and eight
columns. The columns were packed in our laboratory by
dry-packing method, using vibration, as we described
above. (Column dimension was 250X16 mm I.D.). The
uniformity of columns was tested by elution investigations,
injecting pure active compound on the columns and
eluting it with the eluent. In the zone of the pure main
fraction a KNAUER UV/VIS filter photometer was
connected. The wavelength was set to 254 nm. For the
purpose of recording the successive chromatograms a
Radelkis OH-850 potentiometric recorder was applied.
The temperature was set to 25 ˚C. (See Fig. 4.)

Fig. 4.: KNAUER CSEP 9116 SMB unit assembled for

quasi-continuous process

Analytical Method Applied for the Quality Control
of the Separation

The quality control of the purified active compound was
performed by a validated analytical chromatographic
method, which is applied for the time being in the
current manufacturing process of the product.

Preliminary Elution Experiment

Before beginning the experiment, the column was
equilibrated with the eluent. The injected sample
amount was 0.32 g of crude material, which was solved
in 5 ml methylene-chloride. After the injection step, the
elution of less retained impurities was started and seven
fractions were collected, for the purpose of determination
of the first cut point of the pure main fraction. The end
cut point of the peak was not so critical, because a
certain impurity of small amount, eluting in the tail of
the active compound peak can be removed by the
crystallization step, following the chromatography. That
was the reason, why there was no fraction collection
there.

Fig. 5.: Chromatogram of the preliminary elution

experiment

When the elution of the main compound was
finished, the back-flushing with ethyl acetate was
started and continued until a stable baseline was
achieved. At the end of the back-flushing, the column
conditioning was carried out with twice of the column
volume of the eluent. Based on the pump flow rate and
speed of the recorder, the solvent volumes, necessary to
the subsequent steps of the separation could be
calculated. (Fig.5.)

Quasi-Continuous Elution Experiments

The switching time of the rotating SMB valve was
calculated based on the critical data of the preliminary
elution experiment, such as the elution volumes of less
retained impurities and the pure active compound. (See
Table 1.) The pump flow rates were set in accordance
with the calculated switching time and the solvent
volumes, necessary to the steps of the separation.

Table 1: Results of the preliminary elution experiment

Steps of the separation Solvent volumes necessary to the subsequent steps of the separation (mL)
Elution of less retained impurities ≈ 105

Elution of the pure active compound ≈ 84
Back-flushing with ethyl acetate minimum 65

Column conditioning twice of the column volume, minimum 100

The crude sample was solved in methylene-chloride,
which was the weaker solvent component of the eluent.
The compounds were adsorbed on the surface of
packing material from this solution in a narrow band at
the beginning of the columns, therefore the injected
sample volume could be higher, in most cases exactly
10.5 mL. In this case the sample pump (P1) flow rate
was set to 0.5 mL min-1, which was a reasonable value.

In the quasi-continuous experiments the sample load was
the same as in the preliminary elution experiment,
namely 0.32 g of crude material. The exception was
only the second experiment, in which the sample load
was lower, exactly 0.29 g of crude material. The process
parameters of the experiments are summarized in Table
2.

Table 2: Process parameters of quasi-continuous elution experiments

Flow rates of the pumps (mL min-1)

Experiment
Switching

time
(min)

Sample
injection

Elution of less
retained

impurities P5

Elution of the
pure active

compound P4

Back-
flushing

Column
conditioning P3

1. 21 0.5 5 4 4 3
2. 19 0.5 5 4 4 3
3. 21 0.5 4.7 4 4 3
4. 21 0.5 4.5 2.5 4 3
5. 21 0.5 4.5 3.2 4 3
6. 21 0.5 4.5 3.5 4 3
7. 21 0.5 4.5 3.5 4 3
8. 14 0.75 6.75 5.2 6 4.5

Results and Discussion

Results of the Preliminary Elution Experiment

The aim of this experiment was to determine the solvent
volumes necessary to the steps of the separation. The
received data can be seen in Table 1.

Elaboration and Intensification of the Quasi-
Continuous Process

As we mentioned before, based on the critical data of
the preliminary elution experiment, such as the elution
volumes of less retained impurities and the pure active
compound, (Table 1.) the applied switching time was 21
min in the first experiment. In this first run the purity
was very good, but the recovery was poor, because a
significant part of the main fraction eluted in the zone of
the less retained impurities. The reason of the deviation
between the preliminary elution experiment and the first
quasi-continuous run was the difference of the extra
column volumes of the two systems in which the
experiments were carried out. To improve the result of
the first quasi-continuous run, two possible ways were
tried:

The first was the reduction of the switching time, but in
this case the sample load and volumes of solvents in
each zone were also decreased, therefore finding the
optimal set of process parameters was difficult. (2.
experiment in Table 2.)
• The other (easier) way was decreasing the flow rate

in the zone of the less retained impurities. (3.
experiment)

Fig. 6.: Chromatograms of the pure main fractions

eluted from successive columns

In the 4., 5. and 6. experiments the cut point at the
end tail of the main fraction was optimized with varying
the flow rate in the zone of the active compound for the
purpose of keeping the recovery as high as possible and
the level of the above-mentioned impurity under a
certain limit, with respect to the cleaning effect of the
crystallisation step.

With the best process parameters a repetition was
carried out. (7. experiment)

In the final step (8. experiment) the productivity was
increased with decreasing the switching time and
proportionally increasing the flow rates at the same

time. The last two runs, switching from parameter set of
7. experiment to the ones of 8. experiment, with two
transient peaks can be seen in Fig. 6.

The reason of transients is that, when the
experimental parameters were changed, the peak of the
main fraction from the actually injected sample eluted
after two switching. That is why the new valuable status
was possible to see on the recorder after two transient
peaks.

The specific capacity parameters of 7. and 8. runs, in
comparison with the pilot-plant batch elution separation
are shown in Table 3.

Table 3: The specific capacity parameters of 7. and 8. runs in comparison with the pilot-plant batch elution separation

Specific capacity parameters

Experiment
Specific solvent

consumption
Methylene chloride

(mL g-1)

Specific solvent
consumption Ethyl

acetate
(mL g-1)

Productivity
[g (kg packing hour)-1]

Recovery
(m m-1 %)

Batch elution
pilot-scale 499.51 214.34 10.15 84.26

7. 836.02 409.97 4.11 86.85
8. 761.58 374.30 6.75 95.06

In the 8. experiment a higher recovery resulted a little
bit higher level of impurities in the product, but the
quality still met the requirements.

Conclusion

The quasi-continuous elution chromatographic
purification of a steroid active compound was possible
to carry out successfully with the help of the rotating
switching valve of a laboratory-scale Simulated Moving
Bed (SMB) instrument. The quality of the product,
gained at optimum operating conditions met the purity
requirements (purity > 99 %, maximum level of
individual impurities < 0.3 %).

The main advantage of applying this quasi-
continuous separation method is that the staff need for
processing is less, than in case of batch elution
chromatographic separation. Fraction collector is not
needed, because cutting is done automatically by
switching the columns and the main fractions coming
from successive columns can be pooled in one vessel.

Comparing the specific capacity parameters of the
quasi-continuous process with the batch pilot-scale
separation, the specific solvent consumption is higher
and the productivity is lower. But mention must be
made, that the number of columns applied in the SMB
instrument were eight instead of the necessary five
because of the construction of the SMB unit. The
columns applied in the SMB instrument were shorter
than the pilot-scale column, so the theoretical plate
number of these columns was considerably lower,
therefore the specific sample load also had to be lower.
(In case of pilot-scale column the specific sample load
was 0.02 [g sample g-1 packing material], while in

quasi-continuous process it was only 0.0143 [g sample
g-1 packing material].)

The robustness of the process demands the
uniformity of columns inserted in the SMB instrument.

The change of temperature has strong impact on the
adsorption behaviour of compounds, therefore transients
in temperature can detrimentally influence the
separation.

SYMBOLS

I.D. Internal diameter
F Feed
c Concentration of the compounds
R Raffinate
E Extract
L Liquid recycle
D Fresh solvent
P “Recycle of the packing”, column switching

time

INDICES

W Less retained compound
S More retained compound
F Feed
R Raffinate
E Extract

REFERENCES

1. MANN G. (1998) Analusis 26 (7): 76-82
2. COLIN H. (1998) Analusis 26 (7): 15-17
3. KATTI A. M., JAGLAND P. (1998) Analusis 26 (7):

38-46
4. A. SEIDEL-MORGENSTERN (1998) Analusis 26 (7):

46-55
5. GRILL C. M., MILLER L., YAN T. Q. (2004) J

Chromatography A 1026: 101-108
6. NICOUD R.-M. (1999) Pharmaceutical Technology

Europe 11: 36-44
7. CHARTON F., NICOUD R.-M. (1995) J

Chromatography A 702: 97-112
8. KHATTABI S., CHERRAK D. E., MIHLBACHLER K.,

GUIOCHON G. (2000) J Chromatography A 893:
307-319

9. PEDEFERRI M., ZENONI G., MAZZOTTI M.,
MORBIDELLI M. (1999) Chem Eng Sci 54: 3735-
3748

10. LEE H.-J., XIE Y., KOO Y.-M., WANG N.-H. L.
(2004) Biotechnol Prog 20: 179-192

11. SANTOS M. A. G., VEREDAS V., SILVA JR. I. J.,
CORREIA C. R. D., FURLAN L. T., SANTANA C. C.
(2004) Brazilian Journal of Chemical Engineering
21 (1): 127-136

12. TEMESVÁRI K., ARANYI A., CSUKÁS B., BALOGH S.
(2004) Chromatographia Suppl. 60: 189-199

13. NICOUD R.-M., BAILLY M. (1992) Proceedings of
“PREP 92”, NANCY (France), 6-8 April, ISBN 2-
905267-18-6: 205-220

14. BLEHAUT J., NICOUD R.-M. (1998) Analusis 26 (7):
60-70

15. GUIOCHON G. (2002) J Chromatography A 965:
129-161

16. HIMBERT F., PENNANEC R., GUILLAUMET G.,
LAFOSSE M. (2004) Chromatographia 60: 269-274

17. YANG X., LIU K., XIE M. (1998) J Chromatography
A 813: 201-204

18. ZHANG X., GEOFFROY P., MIESCH M., JULIEN-DAVID
D., RAUL F., AOUDÉ-WERNER D., MARCHIONI E.
(2005) Steroids ‘in press’

19. DOI: 10.1016/j.steroids.2005.06.003
20. HEITMANN D., ZIEHR H., MÜTHING J. (1998) J

Chromatography B 710: 1-8