Commercial Scale Counter Current Chromatographic
Separation of a Pharmaceutical Intermediate:
Achieving purity when conventional chemical approaches
Eastern Analytical Symposium
November 16, 1999
T.G. Archibald, G.G. McParland, M. Chalker,
B. Kelson, A.A. Malik, T. E. Clement, H. Palandoken,
A tale of two epoxides
We were making one epoxide and throwing the other
away in the waste stream.
One day someone came along and wanted the other
Oh Boy, we said. We can sell our waste for lots of
Very Clean Commercial Process Currently in
Operation at Multi-metric Ton/month Scale
Et3N CH2N 2
BocNH COOH BocNH R BocNH N2
O O O
HCl [H] Base
BocNH BocNH Cl BocNH
We Need both S,S- and R,S- Diastereomers
Ideally, you control the reduction chemistry
to get the desired diastereomer.
BocNH COOH BocNH CCH2Cl
Cl BocNH Cl
So We Tried Hundreds of Reagents
and Reaction Conditions:
Some Selected Reductions of CMK to CMA
Reagent(s) Solvent(s) Temp Time R,S:S,S
Li(OtBu)3AlH Et2O 0C 3Hrs 8:1
(+)-Dip Chloride(1.4eq) THF 5C-RT 12Hrs 5:1
K-Selectride THF Reflux 2 Hrs 2:1
R-Alpine Borane(Conc.) THF Reflux 9Dys 1:1
L-Selectride THF R.T. 1 Hr 0.9:1
Most conditions NaBH4/CeCl3(anh.) THF RT 2 Hrs 0.8:1
give mixtures NaCNBH3 THF RT 36Hrs 0.7:1
(+)-2-Butanol/NaBH4 THF RT 1 Hr 0.6:1
of R,S- NaBH4/(-)-2-Butanol THF RT 30Min 0.6:1
and S,S- NaBH4/L-Tartaric Acid THF 5C 1 Hr 0.6:1
NaBH4/D-Tartaric Acid THF RT 30Min 0.5:1
in high isolated BH3-t-butylamine THF R.T. 1 Hr 0.5:1
yield. LAH THF 25 C 1 Hr 0.5:1
THF*BH3 EtOH/THF R.T. 2 Hrs 0.2:1
Al(iOPr)3 IPA 50C 3 Days 0.05:1
With the reduction well understood:
It should now be easy to get R,S- or S,S- at any “almost
purity” you liked from 95:5 to 10:80.
But, Typical demand is for >99.8% de
So, it becomes just a purification problem.
However - Real Life is Often Not So Clean
S,S- is highly crystalline and easily purified
Typical purity is 99.8% with <0.1% R,S-
R,S- is lower melting and more soluble
Forms a 94:6 eutectic with S,S-
Can not be purified to 99% by crystallization either at
chloromethyl alcohol or epoxide stage.
So - How to Make Pure R,S-
by Most Efficient Route?
We looked at alternative
Alternate Routes Gave High-purity R,S- but at
Inversion route 90% yield 99% de
H H H H
Br BocN H
BocN H CCH2Br BocN H O
BocN H COOH OH BocN H H
Olefin Route 65% yield 98% de
We Began to Look at Industrial Scale Counter
Current (SMB) Chromatography
as an Alternative.
Is it real?
Does the technology really exist and will it work
Can we win?
If we use the technology, will it make the desired product with
the desired quality.
Is it worth it?
Can we use the technology cost effectively against other
Counter Current Separation
Weaker Retained Enantiomer Stronger Retained Enantiomer
Wind (Mobile Phase)
M. Negawa, F. Shoji, Journal of Chromatography 1992, 590, 113
t0 t0 + T / 2
ELUENT EXTRACT ELUENT EXTRACT
RAFFINATE FEED RAFFINATE FEED
t0 + 1 T t0 + 1 T + T / 2
The diastereomers move at different rates on the stationary phase. By moving
the location of feed and removal lines by solenoids, a continuous separation of
enantiomers takes place.
Analytical scale HPLC Separation of
R,S- and S,S-CMA
Normal Phase Separation on Silica Gel
Modeled from Analytical Chromatogram
Feed, Raffinate and Extract
In Operation on Reverse Phase
Selected Separation Conditions
CSP k’1 Solubility % productivity g/day/kg
Chiralpak® AD CMA 0.9 2.6 1.0 100
Zorbax C-18 CMA 0.91 1.43 0.05 22
Zorbax C-18 Epoxide 0.46 1.6 2.6 130
Silica Gel Epoxide 0.85 1.3 25 400
Chiralpak® AD Epoxide 0.9 2.1 25 100
Various solvents and column media can be used
Separations scale from analytical data
Preparation of multi-kilograms within short period of time
Lab SMB for training, customer demonstrations, and process
1-15 kg quantities.
Semi-works (“8-200”) unit capable of three to five metric tons per year
Phase III product support
Early commercial quantities
Commercial scale (“6-800”) unit to support a commercial product
launch ( 40-75 metric tons per year)
Operational July 2000
Selected Milestones in Counter-Current
1960’s UOP - Industrial separation of xylenes
1985 Prochrom and Separex started
1987 FDA approves first drug purified by batch prep HPLC
1992 Daicel Patents for Chiral Separations
1993 First implementation of short column/non-rotary valve equipment by NOVASEP
1994 First “real” operating machine with multi-ton capacity (8 x 200 at NOVASEP)
First commercial batches (600 g) of CSP by Daicel
1997 First 6 x 450 plant begins operation.
1999 Large scale plant constructed at Aerojet for commercial production
“Apologies to any historians present”
Licosep Lab Unit
Aerojet Lab in Sacramento CA
Control Screen for Lab Unit
SMB Facility is Under Construction
April 30, 1999
Commercial Separations Suite
Modular Building Design
2 Large-scale SMB units on one side
Reactor suite, solvent recycle and solids isolation on
Class 100,000 Building
All operations are conducted under cGMP
Goal is to combine chemical steps with
Derivatization and purification
Is it worth it?
Cost of separations vary with chromatographic
Single enantiomer separations are cost competitive at
throughputs rates of 1 kg/day/kg of phase.
Multiple component mixtures can be separated.
High containment of materials and very high recoveries
Continuous separations are producing R,S-epoxide
Purity of 99.8 % de and 96% recovery are obtained.
“waste” from S,S- process can be mined for R,S- by this method.
Chromatography gives separations at a lower cost than inversion
The separation technology coupled with reduction control gives
flexibility in product mix to meet commercial demand.
Roger-Marc Nicoud, NOVASEP
Tom Lewis, Chiral Technologies
Rob Miotke, Aerojet Fine Chemicals