(13f) On-Line Estimation of Diastereomeric Resolution with the Use of Fbrm, Raman Spectroscopy, and Atr-Ftir

                    The practical resolution of racemic
compounds into pure enantiomeric components is a very important process
research activity in the pharmaceutical industry today. One commonly used
technique for such resolution (separation) is the diastereomeric formation
route where the racemic compound reacts with a chiral acid or base to form a
mixture of two diastereomers that by definition have different physical
properties (i.e. solubilities, and pharmacological activities) and thus can be
separated through traditional crystallization techniques. The capability of
on-line monitoring of the optical purity of the crystals during diastereomeric
crystallization will help to develop a robust crystallization procedure.

   The
first objective of the presentation is to examine an alternative on-line
measurement to replace slurry density used in a Partial Least Square Regression
(PLS) model to predict solid composition of one of the two diastereomers[1].  The second objective is to present a PLS
model that quantifies the concentration of both diastereomers in liquid
phase. 

Our recent
research[2]
showed that Raman spectroscopy is capable of differentiating diastereomers in a
crystallization slurry, and the estimation model accuracy increases provided
the changing process parameters of temperature and slurry density are included
in the model.  However, slurry density
is not easily measured on-line without a sampling loop.  It is essential to find an alternative
on-line measurement from which to infer slurry density.  The choice for the alternative online
measurement is ATR-FTIR and FBRM. 
ATR-FTIR is often used to study how solute concentration changes in
solution phase over time during crystallization.  It serves as an indirect measure of the solid phase dynamic.  One the other hand, FBRM is a particle size
counter and it gives a more direct measure of slurry density in measuring the
total number of crystals.  Different
combinations of two or all of the measurements are incorporated into a new PLS
estimation model and quantify the solid composition of the diastereomers.

The second
objective is achieved by building a calibration model using ATR-FTIR to
quantify both the total solute concentration and percent composition of the
diastereomer in the solution phase.  The
PLS models will be a prelude to understand and model the thermodynamics and
kinetics of crystallization process of interest.