2
880 Crystal Growth & Design, Vol. 10, No. 7, 2010
Zhang et al.
Seed crystals of pure (R)-MA ranged from 212 to 300 μm were
prepared using a GA-6 Gilsonic Autosiever (Gilson, Worthington,
OH). To avoid dissolving in the solution, seed crystals were added
from a head nozzle to the crystallizer at the 2 °C below the starting
temperature of each crystallization batch. A M400LF focused beam
reflectance (FBRM) system (Lasentec, Redmond, WA, US) was
used to determine the seeding effect.
2.2. Batch Crystallization. The crystallization experiments were
performed in a jacketed 250 mL glass crystallizer. A Teflon-coated
stirrer bar was used for magnetic stirring (25 ꢀ 8 mm), which was
kept at a constant stirring rate of 300 rpm to ensure that the crystals
were well distributed in the solution. The temperature in the crystal-
lizer was controlled by a Julabo FP50 bath circulator (Allentown,
PA, US). The FBRM system was used to monitor the nucleation of
fresh nuclei or the dissolution of the solids in the solution. An in situ
ATR-FTIR (Hamilton Sundstrand, CA) was used for collecting of
infrared spectra. The IR spectra were related to the total solute
concentration in the solution using a calibration curve. The optical
rotation of the solution which indicates the concentration difference
of the two enantiomers was measured continuously by an Autopol
IV polarimeter (Rudolph Research Analytical, Hackettstown, NJ,
US). Crystal-free solution with a volume of 20 mL was drawn from
the crystallizer every 4 min using a DOSE IT P910 peristaltic pump
Figure 1. Illustration of a typical ternary phase diagram for a
racemic compound forming system.
from the racemic mixtures; pure enantiomer can only be
produced by direct crystallization when the solution compo-
(
IBS Biosciences, Switzerland) with the help of a membrane filter
and heated during the transport to an online polarimeter. After
measurement, the solution was pumped back into the crystallizer.
Signals from ATR-FTIR and polarimeter were then converted to
obtain the individual concentration of each enantiomer in the
solution.
A typical experimental run was performed as follows. The
aqueous solution with different initial ee of (R)-MA for the batch
crystallization was prepared according to the solubility determined
from the experiment. The prepared solution was heated up and
maintained 30 min at the starting temperature to ensure that all of
the crystals were dissolved. The temperature of the solution was
then quickly lowered by 2 °C, and at this temperature, a predeter-
mined amount of homochiral seed crystals of (R)-MA were added to
initiate the experiment. The solution, being stirred at a rate of
3,14
sition exceeds the eutectic composition of the system.
The
principle of direct crystallization process for resolution of a
racemic compound can be illustrated in a typical ternary
solubility phase diagram as shown in Figure 1. The vertexes
of the triangle represent the pure components: the solvent
(on top), the (þ)- and (-)-enantiomers (left and right). The
area within the triangle can then be divided into a number of
15
domains. Mixtures in the upper region comprise single
phase solutions. On reducing the water content, a number of
regions emerge in which solids and solutions are in equili-
brium. The regions on the left- and right-hand sides are the
two-phase areas in which crystals of pure enantiomers are in
equilibrium with saturated solutions having compositions on
3
00 rpm, was then linearly cooled to the ending temperature at a
rate of 0.02 °C/min and maintained at that temperature for 20 min.
At the end of each experiment, the slurry was quickly filtered
through a 1 μm filter paper under reduced pressure. Subsequently,
crystals were air-dried at 60 °C overnight and the size distribution
was measured. The experimental conditions for the crystallization
batches are summarized in Table 2. Run 2 and Run 3 were replicated
and the repeatability of the batch crystallization processes was
found to be very good.
0
0
lines AE or A E . In the middle two-phase region, the racemic
compound is in equilibrium with solutions of compositions
0
on line ERE . The remaining regions are the three-phase
areas in which mixtures of pure enantiomers and racemic
compound crystals are in equilibrium with solutions of fixed
composition. It follows that crystallization in two-phase
regions would only yield one solid form, that is, pure enantio-
mer or racemic crystals, respectively, while crystallization in
the three-phase region will yield products that are mixtures of
pure enantiomer and the racemic compound crystals. Hence,
pure enantiomers can be obtained by direct crystallization
when the process is controlled within the left- or right-hand
side of the two-phase regions.
2.3. Analytical Methods. An Agilent 1200 series HPLC system
(
(
Agilent Technologies, Palo Alto, CA) with ChiralCel OD-H column
Chiral Technologies, West Chester, PA) was used to measure the
optical purity of final products. The mobile phase was a mixture of
hexane and isopropyl alcohol (85:15, v/v), with 0.1% of acetic acid
used as a modifier. HPLC analyses were performed at 25 ( 0.1 °C
with an elution flow rate of 0.7 mL/min. The detection wavelength
was 254 nm.
The optical purity of final products was also verified using a
Mettler Toledo 822e DSC system, together with the STARe soft-
ware. Final products were examined by differential scanning calori-
metry (DSC) from 25 to 150 °C at a heating rate of 5 °C/min.
The powder X-ray diffraction (XRD) patterns of the final pro-
ducts were also tested using a Rigaku-MiniFlex powder diffracto-
meter and were compared with those of enantiomeric pure (R)-MA
and (R,S)-MA. Samples were scanned from a diffraction angle (2θ)
of 5° to 55° with a step size of 0.05° and a counting time of 1 s for each
step. The crystal size distribution of the final product was measured
using a Malvern Mastersizer (Malvern Instruments, UK) with a
Scirocco 2000 sample handling unit.
In order to obtain pure enantiomers, the starting composi-
tion of the enantiomeric enriched solution should be selected
based on the temperature difference (Tstart - Tend) and the
eutectic composition at the ending temperature in case of
cooling crystallization. On the basis of the solubility data, the
starting and the ending temperatures of a batch crystal-
lization process were first determined. The eutectic point
0
E of the ending temperature (Tend) was then fixed on the
ternary phase diagram. Connecting the point corresponding
0
to pure emantiomer to point E gives a line which intersects
the solubility curve of the starting temperature (Tstart) at
point P. The composition of point P corresponds to the
lowest initial enantiomeric excess value (ee ) of the desired
0
3
. Theoretical Aspects
enantiomer.
3.2. Modeling and Simulation. A simplified dynamic model
for an ideally mixed batch crystallizer was used assuming no
3.1. Direct Crystallization from Enantiomeric Enriched
Solution. For racemic compound forming systems, crystal-
lization alone is not capable of obtaining pure enantiomers
1
6-18
crystal agglomeration, abrasion, and breakage.
It was