Mass-Directed Preparative SFC with Open-Bed Fraction Collection
Rui Chen, Ph.D.
Waters Corporation, USA
A p p lic at ion ben efit s
INTRODUC TION
The Gas-Liquid Separator (GLS) in the Prep 100
In recent years, improving productivity and cost-savings have become major initiatives
SFC MS Directed System enables open-bed fraction
in pharmaceutical discovery and development. The introduction of new technologies
collection at high flow rates. Combined with the
for high-throughput synthesis has enabled preparation of up to 48 compounds in the
AutoPurification LC MS system in high throughput
same time frame previously required for a single compound1. Consequently, there has
purification environments, the orthoganility
been increasing pressure on development and implementation of efficient purification
between RPLC and SFC offers potential cost-savings
platforms to support the overall high throughput process.
by recovering more compounds from medicinal
chemistry for ensuing research and development.
Mass-directed reverse phase liquid chromatography (RPLC) purification has revolutionized the workflow for chemists in drug discovery by enabling one fraction collection
per injection, largely due to the specificity that MS detection offers. As a result,
mass-directed RPLC has quickly become the most commonly used technique in high
throughput purification environments. Considerable effort has been put into streamlining the process so that minimal user intervention is required. However, a major
limitation in the process is the long dry-down time of aqueous fractions downstream.
Adopted primarily for chiral analysis and purification by the pharmaceutical industry in the early 1990s, supercritical fluid chromatography (SFC) has experienced
Wat e rs solut ions
a remarkable resurgence in the past decade. The reduction in solvent consumption
Prep 100 SFC MS Directed System
and collection in relatively small volumes of volatile organic solvents in preparative
2767 Sample Manager
SFC has led to significant savings on operational costs. For example, Ripka et al.
calculated that 20,000 samples purified by SFC instead of RPLC would realize a
3100 Mass Detector
48 times reduction in solvent consumption2.
2998 Photodiode Array Detector (PDA)
A mass-directed preparative SFC system, like its counterpart in RPLC, is an ideal platform
Gas-Liquid Separator
MassLynx Software
TM
for high-throughput purification of diverse libraries of drug-like compounds. However,
open-bed fraction collection has been a challenge for SFC due to aerosol formation
caused by the depressurization of CO2, particularly at high flow rates. Kassel et al.
demonstrated the feasibility of mass-triggered SFC fraction collection at 15 mL/min3 and
Zhang et al. reported their results with flow rates up to 30 mL/min4. The challenge of
K e y wo rds
managing aerosols in collection at high flow rates has been overcome with the develop-
Open-bed fraction collection
ment of Waters® Prep 100 SFC MS Directed System5.
CO2 expansion
In this application note, we present our research results to illustrate one of the
SFC
Purification
key innovations of the Prep100 SFC MS Directed System: the Gas-Liquid Separator
(GLS). Some comparative results on library compound purification using both
AutoPurification® LC/MS and Prep 100 SFC MS Directed systems are also presented
to highlight the orthogonality of the two techniques.
1
Figure 1. Schematic of the Prep 100 SFC MS Directed System.
SYSTEM CONFIGURATION
All preparative experiments were performed on a Prep 100 SFC MS Directed System,
as shown in Figure 1, which consists of: Waters 2767 Sample Manager, 3100 MS
Detector, 2998 PDA Detector, 515 Pumps and Waters CO2 Pump and Co-solvent
Pump, Analytical-2-Prep Column Oven, Automated Back Pressure Regulator
(ABPR), GLS, and Tunable Splitter. The system was controlled by Waters
MassLynx™ Software.
RESULTS AND DISCUSSION
One of the major challenges in implementing open-bed fraction collection at high
flow rates in SFC is CO2 expansion. After exiting the ABPR, high pressure liquid
CO2 decompresses to its gaseous form at a volume ratio of approximately 1:500.
This rapid expansion causes aerosol formation, which can lead to sample loss and
cross contamination. This challenge is addressed by a proprietary GLS, as shown in
Figure 2. The geometry of the GLS, the dimension of the initiator, and its angle with
respect to the inner wall effectively eliminate the majority of gaseous CO2. Another
key aspect of the GLS implementation is to preserve the peak integrity during a
2
Mass-Directed Preparative SFC with Open-Bed Fraction Collection
Figure 2. Schematic
of the GLS.
gradient before and after the GLS, such that time delay between detectors and collector tip across a gradient can
be accounted for correctly. To accomplish that, a make-up flow of methanol with a reverse gradient is introduced
to ensure a constant flow passing through the GLS, especially at the point in the gradient when the co-solvent percentage is low. To prove the efficiency of the GLS, two detectors, one placed in the main flow path before the GLS
and the other after the GLS in place of the collector tip, were used to compare the pre- and post-GLS peak profiles.
Figure 3 shows the two traces at different modifier percentages from the two UV detectors. It is evident that peak
profiles were well maintained passing through the GLS. The maximum band broadening due to GLS is 1.2 s (results
not shown) at the peak tail, representing a less than 2% sample loss.
UV
MS triggered
collection
portion
Collector
tip
Same
portion at
collector tip
Figure 3. UV traces from
two detectors at 10%, 30%,
and 50% co-solvent.
10% Co-solvent
30% Co-solvent
50% Co-solvent
The effectiveness of the GLS is demonstrated by the high recovery and high purity of collected fractions. For example,
Aurigemma et al. reported their evaluation of the Prep 100 SFC MS Directed System5 . In their experiment, a mixture
containing 8 mg/mL each of (A) flavone, (B) carbamazepine, (C) acetaminophen, and (D) sulfamethazine was injected
and collected over three replicate injections, as shown in Table 1. Overall, sample recovery and purity was greater
than 95% (with the exception of peak C in injection 1) during gradient elution. In particular, the high purity of two
closely eluting components (C and D) indicated a negligible cross contamination between peaks.
In a separate experiment6, similar evaluations were performed using a mixture of caffeine, sulconazole, and
bendroflumethazide. On average, greater than 90% recovery and greater than 99% purity were achieved for
all compounds using UV triggering, MS triggering, and UV+MS triggering, respectively.
An additional built-in advantage of the Prep 100 SFC MS Directed System is the software continuity from the
analytical platform, Resolution™ SFC MS System, capable of both pre-purification scouting and post-purification
analysis. Both platforms are controlled by the MassLynx Software package. After the analytical scouting run,
data is processed and the retention time for the ion of interest is used to match one of the pre-defined focused
gradients on the preparative system. The purification strategy is automatically determined and exported to the
preparative platform. The whole process is automated by AutoLynx™ and FractionLynx™ application managers,
part of MassLynx Software control.1 Furthermore, since both Prep 100 SFC MS Directed and AutoPurification
LC/MS systems are controlled by MassLynx Software, it should be effortless for existing AutoPurification
LC/MS users to adapt Prep 100 SFC MS Directed into their workflow.
3
Mass-Directed Preparative SFC with Open-Bed Fraction Collection
RPLC
100
Intensity (%)
0%
130
207
251
323
m/z
354
0
0
1
Time (min)
2
Absorbance (AU)
400
Absorbance (AU)
80
0
0
2
3
130
206
323
m/z
377
2
4
Time (min)
6
8
200
100
Time (min)
395
300
20
1
187
320
m/z
0
500
40
0%
125
0
3
60
321
100%
0%
100
0
349
100%
321
Intensity (%)
SFC
100
349
100%
0
2
4
Time (min)
6
8
Figure 4. RPLC and SFC
chromatograms illustrating
the different selectivity
between RPLC and SFC.
Implementing the Prep 100 SFC MS Directed System into high-throughput purification environments provides a
cost-savings that results from reduction in solvent consumption and dry-down time2,7. In addition, the orthogonality between SFC and RPLC offers increased cost-savings by pushing more compounds from medicinal chemistry
through the pipeline. SFC is loosely considered a normal phase chromatographic technique and complements RPLC
in selectivity. For example, Figure 4 shows the comparison between mass-directed RPLC and SFC in purifying a
real-world pharmaceutical compound directly from combinatorial synthesis; in which, the impurity was resolved
from the target compound in SFC but not in RPLC. In our previous study8, we reported a 30% increase in success
rate when combining the two techniques compared to using each technique alone. In the same study, we also
reported a 20% reduction in overall batch processing time for SFC compared to RPLC as a result of shorter drydown time post-purification.
Injection 3
Injection 2
Injection 1
Inj.
4
Peak
UV Purity
TIC Purity
Recovery (TIC)
A
100.0
100.0
99.0
B
100.0
100.0
97.0
C
100.0
100.0
86.0
D
98.7
99.1
98.0
A
100.0
100.0
97.0
B
100.0
100.0
96.0
C
100.0
100.0
97.0
D
100.0
99.2
97.0
A
100.0
100.0
97.0
B
100.0
99.5
95.0
C
100.0
100.0
99.0
D
100.0
99.4
98.0
Mass-Directed Preparative SFC with Open-Bed Fraction Collection
Table 1. Purities and
recoveries of SFC collections.
CONC LUSIONS
n
References:
The GLS, the key innovation in the Prep 100 SFC MS Directed System, has been able
to overcome one of the main historical challenges in SFC open-bed fraction collection
n
Greater than 90% recovery and greater than 95% purity were achieved in two
independent evaluations, which indicates the effectiveness of the GLS
n
The level of automation available through the software between the analytical
and preparative platforms enables a streamlined analytical to purification
process with minimal user intervention
n
The orthogonality between mass-directed RPLC and SFC can potentially offer
users additional cost-savings by recovering more compounds from medicinal
chemistry for ensuing research and development
1. McClain, RT et al., Journal of Liquid Chromatography &
Related Technologies, 2009; 32: 483–499.
2. Ripka, WC, Barker, G, and Krakover, J., Drug Discovery Today,
2001; 6(9): 471-477.
3. Wang, T. et al., Rapid Commun. Mass Spectrom,
2001; 15: 2067-2075.
4. Zhang, X. et al., Journal of Combinatorial Chemistry,
2006; 8: 705-714.
5. Aurigemma, C. et al., poster presentation, LabAutomation,
Palm Springs, CA, USA (2009).
6. Chen, R. and Subbarao, L., unpublished results, 2009.
7. White, C, and Burnett, J., J. Chromatog A., 2005; 1074, 175.
8. Mich, A, Matthes, B, Chen, R, and Buehler, S., LCGC Europe,
the Application book, March 2010.
Waters and AutoPurification are registered trademarks of Waters
Corporation. T he Science of W hat’s Possible, Prep 100, MassLynx,
Resolution, AutoLynx, and FractionLynx are trademarks of Waters
Corporation. All other trademarks are the property of their
respective owners.
©2010 Waters Corporation. Produced in the U.S.A.
May 2010 720003468en AG-PDF
Waters Corporation
34 Maple Street
Milford, MA 01757 U.S.A.
T: 1 508 478 2000
F: 1 508 872 1990
www.waters.com