Best Practices for

Crystallization Development
A Review of Modern Techniques

New Technologies for

Crystallization Development
Benjamin Smith, Mettler-Toledo AutoChem, Inc.

Crystallization and precipitation are Today every crystallization and precipita- high quality information, accelerate un-
critical processing steps in chemical tion step has an opportunity for improved derstanding, and establish knowledge for
development. They can serve as understanding and quality. Chemists crystallization development and transfer.
purification and separation steps, and use established inline Process Analytical
have implications on the yield, purity Technology (PAT) techniques to under- This paper demonstrates the methodol-
and particle size distribution. Even stand what is changing during the process ogy chemists use to identify operating
though crystallization has advanced and gain knowledge to ensure the desired parameters such as temperature, solvent
significantly over the past decade, many size, shape and form is isolated. In the addition rates and seeding to improve
chemists have such short deadlines that past, understanding crystallization pro- crystallization, batch repeatability, and
they must base everyday decisions on past cesses was considered time consuming, crystal size and shape distribution.1 By
experience rather than understanding and reserved for specialized groups, who accelerating process understanding,
the crystals in situ. Due to the complexity focused on the most important process more robust crystallization processes
of crystallization, a process may be steps. Today new generations of intuitive are developed with higher yield and
developed simply by crashing solids out process analytical tools provide a rapid higher purity. Examples include a 10%
of solution and transferring a non-robust understanding of changes (nucleation, yield improvement2, elimination of a
process with inconsistencies in the yield, growth, oiling out, agglomeration and costly process impurity17, and increased
purity and particle size distribution. supersaturation) from within the crystal- monthly throughput by 20%4.
lizer. These tools make it easy to gain
 Contents

A Crystallization Workstation:
Optimized Space for Process Understanding 3

Integrated Experiment Platform 3

Establishing Solubility and Metastable Zone Width to

Accelerate Crystallization Development 4

Immediately Understand What is Changing,

and Establish Direction for the Next Experiment 4

Quickly Understand Agglomeration and

Oiling Out Conditions to Improve Purity 5

Accelerate Development by Understanding

Changes to Nucleation and Secondary Nucleation 6

Optimize Seeding and Mixing Conditions to

Produce Fine or Coarse Crystals 7

2
 Common questions during crystallization development are easily
answered with intuitive process monitoring tools by tracking changes
from within the crystallization vessel:
A Crystallization Workstation:
Optimize Space for Process Understanding
“Did the solids crash out?”
During crystallization development, chemists often produce
crystals rapidly without time for a full Design of Experiment “Did a solvent oil out?”
(DoE). There is very little time for thorough process optimiza- “When did the crystals begin to agglomerate?”
tion, yet it is a perfect time to screen design parameters and
“Is it easily transferred to our manufacturing facility
determine the solubility, solvent, and temperature profile. It
or Contract Manufacturing Organization (CMO)?”
is an ideal point to establish a direction which avoids future
disturbances such as impurities, undesired polymorph forms, or “Did I seed at the right temperature?”
particle size and shape distributions that are difficult to process “Will the product have the purity and yield we require?”
downstream. When disturbances like these occur they require
“Is the process consistent batch to batch?”
costly re-designs which can be prevented if caught earlier in
crystallization development. “What will be the filtration rate?”
“What is the solubility?”
Traditional round bottom or jacketed laboratory reactor vessels
provide a manually controlled temperature and mixing environ-
ment. They are time consuming to set up, not repeatable, and
are challenging to configure with in-process analytical tools.
Established small volume crystallization workstations (such as
EasyMax™ or OptiMax™) provide a platform where chemists
quickly and efficiently carry out experiments day and night with
tight control over temperature, mixing, dosing, and pH control.
These vessels are easy to use, highly repeatable, and quickly
integrate with process analytical tools.
What is EasyMax™?
Designed to replace the jacketed lab reactor,
EasyMax™ and OptiMax™ are fast and easy to
set up. They capture experimental data to deliver
an enhanced understanding of the process, allow-
Integrated Experiment Platform ing users to optimize crystallization design with
precise control over critical process variables such
By providing a crystallization development platform for rapid as temperature, mixing conditions, and anti-solvent
laboratory data acquisition, EasyMax™ and OptiMax™ syn- addition rates.
thesis workstations, combined with a standardized software
interface, simplify and accelerate process optimization. Seamless
60
data acquisition and control within a single software suite allows 14000 1517cm-1 Peak Area
0.30
crystal size and shape, solubility, and supersaturation to be 55
12000
quickly interpreted and understood. Synchronized data from 50 0.28 Tr
inline process analytical technologies along with temperature,
Peak Height (A.U.)

10000
45
mixing, pH, and anti-solvent addition enables users to quickly
Counts/sec

0.26
Ref. counts/sec No Wt
8000
convert data to information and make insightful decisions about
°C

40 Paracetamol 1517
Ref Tr
0.24
the next experiment to perform. 6000
35
0.22
4000
Everyday development chemists must quickly identify the correct 30
Total Counts
input and process parameters and understand crystal transfor- 25
0.20 2000
mations to ensure product quality and process performance. A
20 0.18 0
crystallization workstation with inline process analytical tools 02:00:00 04:00:00 06:00:00 08:00:00 10:00:00
enables users to quickly establish direction in everyday crystal- Relative Time
lization and precipitation processes.
Figure 1. Synchronized data from process
analytical technologies

3
 Establishing Solubility and Metastable Zone Width to
Accelerate Crystallization Development
When a crystallization workstation is integrated with in situ mid-IR (such as
ReactIR™), it provides a hardware and software solution for rapid and highly sensitive

Concentration
measurements of the solute concentration and solubility curve, without interference
from the suspended crystals. Knowledge of the solubility curve sets the direction for all Metastable
future crystallization development and enables chemists to maximize yield, purity, and
the particle size distribution. Solubility

Early crystallization development also requires fundamental knowledge of the real-time

solution concentration relative to the equilibrium solubility. The kinetic limit between Temperature

the nucleation point and the solubility curve is the metastable zone width (MSZW). The Figure 2. Solubility curve and MSZW (Metastable
MSZW is essential to successful crystallization development and provides the funda- Zone Width)
mentals for knowledge transfer during the later stages of development. A crystallization
Barrett, M. et al, Chemical Engineering Research
workstation coupled with an inline particle characterization tool (such as FBRM®) and Design6
provides a real-time integrated measurement of the nucleation and growth associated
with the mid-IR measurement of real-time solute concentration.

Powerful yet simple software tools synchronize the operating conditions (temperature,
pH, mixing rate, and solvent dosing) from the crystallization workstation with data
from all process analytical technologies to provide informative reports which show the
impact of the variables on the process.
t = 1:16:15

Immediately Understand what is Changing,

and Establish Direction for the Next Experiment
While developing a crystallization, in situ measurement techniques allow users to
quickly identify the size and shape of crystals, particles, or oil droplets. Inline particle
vision and measurement techniques are especially insightful by offering an eye into a
vessel or pipeline at elevated temperatures and concentration where supersaturation is
high and offline sampling is impossible. For example in Figure 4, inline images (cap-
tured with PVM® technology) provide immediate understanding of crystal morphology
dynamics without the need for sampling. Users can quickly understand the temperature t = 1:04:01
and solvent conditions at the exact point of transition, and with this information, PVM®
users make immediate, real-time decisions regarding the next experiment and the
direction of development.

Figure 3. PVM® images showing changes in crys-

tal morphology

4
 Figure 4. Inline PVM® images
tracking two batches of an
identical organic crystal
molecule with and without
agglomeration

100µm 100µm

Figure 5. Inline PVM® images

help identify conditions which
caused the immiscible phase
(drop) formation during a fast
precipitation process.

100µm 100µm

Quickly Understand Processing Conditions to Improve Purity

Crystal purity is a common concern and rapid agglomeration may trap impurities within
the crystal structure. Consequently, avoiding agglomeration is frequently preferred.
Inline PVM® images quickly reveal the process parameters affecting crystal shape and
the extent of agglomeration. By providing clear insight into the crystal morphology
and agglomeration kinetics, PVM® enables users to quickly identify correct seeding,
temperature, and supersaturation parameters. This ensures the development of a robust
What is PVM®?
crystallization process by avoiding agglomeration and ensuring the desired form and
purity.7 PVM® (Particle Vision and Measurement) is a probe
based vision tool which enables users to study and
immediately understand crystallization with inline
Any chemist or engineer involved with crystallization development for some period of high resolution digital images capturing crystals as
they naturally exist in-process, eliminating the need
time will experience unexpected events such as phase separation (oiling out), which is
for offline sampling. With PVM®, users observe real
often a source of impurities. Oiling out is usually impossible to see by eye and representa- time movies which simply replay the crystallization
tive sampling is typically unobtainable at elevated temperature and supersaturation. process. When implementing PVM® in a crystal-
lization workstation virtually no data analysis is
Inline tools offer insight, which is impossible with traditional techniques. Many case necessary since the images themselves explain the
studies have highlighted the ability to use PVM® to identify oiling out and to quickly particle size and shape changes, and are synchro-
nized with process temperature, mixing, solvent
investigate the root cause of unexpected events7,8,9. In a single experiment, PVM® provides
addition.
information which would have taken excessive time and effort to detect using traditional
analytical techniques. PVM® is used to reduce process development time by months to In everyday crystallization development, PVM®
enables users to immediately visualize the crystalli-
ensure projects are on time and meet purity requirements. zation, understand what is changing, and establish
direction for the next experiment.

5
 Accelerate Development by Understanding Changes
to the Particle Size and Particle Count
Tracking and quantifying inline changes to crystal dimension and count provides a
rapid understanding of the particle system’s response to changing process parameters.
For example, probe based particle measurement provides a fast understanding of crystal
growth and nucleation rates while determining seeding, temperature, and solvent
parameters11.

By understanding how the particle system responds to changing process parameters,

FBRM® users quickly establish conditions to ensure crystal product meets quality
requirements and process performance, repeatability, and stability goals.12, 13, 14, 15
What is FBRM®?
1000 FBRM® (Focused Beam Reflectance Measurement)
300µm measures a fingerprint distribution of the particle
800
system that is sensitive to changes in dimension,
shape and count. Real-time measurements track
Counts (no. weight)

the rate and degree of change to particles and par-

600 175µm ticle structures as they naturally exist in the process
– eliminating the need for offline sampling11.
400
100µm

200

0
0 10 100 1000
Chord Length (µm)

Figure 6. (left) FBRM® measures bimodal distribution; (right) Inline

PVM® image confirmation

70

80000
60
Temperature 100µm
Counts (no. weight)

60000 G400 3/sec 0-20µm 50

Temp (°C)

40
40000
30

20000 20

10
0

0
0 00:15 00:30 00:45 1:00 1:15
Relative Time

100µm
Figure 7. Inline FBRM® measurements track the growth of
the seed and subsequent nucleation of fine crystals during
cooling.10

6
 Figure 8. FBRM® tracks the nucle-
ation kinetics when seed crystals are
added at varying temperatures and
supersaturation.16
Counts/sec (1-10µm)

0.25g
Seed
Added

Seeding Temp
19°C
27°C
33°C

Time

Optimizing Conditions to Produce Fine or Coarse Crystals 2500 3.2mm Pipe

1.6mm Pipe
Nucleation Events

During crystallization, a typical goal is to maximize yield while improving filtration 0.78mm Pipe
0.3 mm Pipe
2000
rates and avoiding downstream bottlenecks17, 18, 19. FBRM® technology provides immedi-
ate process understanding by presenting which process conditions produces fine-small Counts (1-5µm range)
1500
particles and which produce large-coarse particles.20
1000

Rate
For example, FBRM® enables users to understand the effect of antisolvent addition rates
ation
and mixing rates by tracking the number of fine particles (in the range of 1-5µm) over 500
Nucle

Reduction in Fines
time and providing immediate indication of nucleation, growth rates, and endpoints. with Improved Mixing

Inline particle characterization allows researchers to modify addition velocity condi- 0

00:00 00:03 00:06 00:09 00:12 00:15
tions and increase mixing rates to minimize undesirable nucleation and eliminate Time (hr:min)
filtration bottlenecks during laboratory development and scale-up. By improving the
mixing conditions, yield losses caused by excessive fines in the filter, centrifuges, and Figure 9. FBRM® tracks the reduction in fines
dryers (dust) are eliminated. resulting in improved mixing rates

7
 ®
M
PV

™
IR
act
Re

METTLER TOLEDO: World Leaders in Technology for

®

Understanding Crystallization Development FB

RM
METTLER TOLEDO is the world leader in expanding the role of process ana-
lytical technologies for more effective and efficient development of everyday
iC
crystallization processes. Crystallization is a critical process for the purifica- Software
tion and isolation of chemical compounds in the manufacture of many fine ALR

chemical and pharmaceutical products. The results of the crystallization step

have far reaching impacts on overall process efficiency and final product
quality. It is also a difficult process to understand without inline tools which Our People
provide immediate insight from within crystallization workstations. Chemists METTLER TOLEDO has a global net-
have many opportunities to apply inline process analytical measurements work of Technology and Application
with integrated crystallization workstations to immediately understand the Consultants with extensive research
and industry experience supporting
crystallization and make rapid decisions regarding the direction of development
organic chemistry, chemical develop-
and future process viability. Chemists use these tools to reduce development
ment and scale-up.
time, minimize disturbances in later phases of development, and ensure a
crystal product is produced that meets necessary specifications for crystal size Email: [email protected]
and shape distributions. Phone: 410-910-8500

Our process analytical instrumentation includes technologies for Website

in-process crystal population measurement of crystal count and We are online at the
dimensions (FBRM®), in-process crystal imaging (PVM®), and METTLER TOLEDO website
in situ supersaturation monitoring (ReactIR™). FBRM®, PVM®, (www.mt.com/crystallization),
and ReactIR™ technologies are available for 30ml to 2000ml where you can find additional detailed
laboratory development, 2-100 liter kilolab or 100-50,000 liter information on our products and appli-
manufacturing scale, and continuous flow pipelines. cations – including an extensive list of
upcoming and on-demand webinars.
Our Crystallization Workstation Vessels (including EasyMax™,
Blog
Optimax™, and RC1e®) provide an integrated hardware platform
Chemical Research, Development and
with state of the art control of critical process variables – includ- Scale-up is our Blog highlighting the
ing mixing rate, cooling rate and anti-solvent addition rate. latest publications and providing expert
commentary from our own internal
Our iC software suite ties it all together with an integrated software experts and from academic and indus-
package that provides precise control of the laboratory reactor try professionals.
with synchronized data from in situ analytics for understanding
the crystallization. Customer Community
Our Customer Community Site provides
owners and users of our technologies
with free access to archived citation
lists, application reports, case stud-
ies, and extensive training materials
– including immediate access to all of
our on-demand webinars.

Social Media
Get real-time updates through Facebook
and Twitter on the latest developments
in chemical synthesis, chemical engi-
neering and scale-up.
8
References
1. Airiau, C., Applications of Chemometrics in GlaxoSmithKline, 2009.
http://madd.polytech-lille.net/Download/EPIC09/Airiau.pdf
2. A. Ridder, How to Control Particle Size on Plant Scale: High Shear Wet Milling or Traditional
Optimization, Proceedings of the 15th International Process Development Forum, Annapolis,
2008.
3. Scott et al, Organic Process Research & Development 9, (6): 890-893, 2005.
http://pubs.acs.org/doi/abs/10.1021/op050081p
4. K. Wood-Kaczmar, The Applications of FBRM® to Resolve Batch Processing Problems in Drug
Manufacture, A Case History, Proceedings of 2001 Lasentec User Conference, Barcelona, 2001.
http://us.mt.com/us/en/home/supportive_content/application_editorials.M-2-130_applica-
tion_note.twoColEd.html
www.mt.com/crystallization
5. Wu, H. et al, International Journal of Pharmaceutics, Quality-by-Design (QbD): An Integrated
Process Analytical Technology (PAT) Approach for a Dynamic Pharmaceutical Co-precipitation
Process Characterization and Process Design Space Development. Benjamin Smith
http://www.ncbi.nlm.nih.gov/pubmed/21138762 [email protected]
6. Barrett, M. et al, Chemical Engineering Research and Design, Supersaturation Tracking for the
Development, Optimization and Control of Crystallization Processes.
http://www.sciencedirect.com/science/article/pii/S0263876210000638
7. Desikan, et al, Organic Process Research & Development, (9): 933-942, 2005.
http://dx.doi.org/10.1021/op0501287
8. Deneau & Steele, Organic Process Research & Development, (9): 943-950, 2005.
http://dx.doi.org/10.1021/op050107c
9. Lafferrère L. et al., Crystal Growth and Design, 4, (6): 1175-1180, 2004.
http://www.cinam.univ-mrs.fr/pro_perso/veesler/pdf/demixion.pdf
10. Barrett, M., Hao, H., Glennon, B. Solid State Pharmaceutical Cluster.
11. Leyssens, T. et al. Org. Process Res. Dev., 15 (2), pp 413–426, 2011.
http://pubs.acs.org/doi/abs/10.1021/op100314g
12. Barrett, M. et al., Chemical Engineering Science, (66): 2523–2534.
13. Barrett, P. et al., Organic Process Research & Development, (9-3): 348-355, 2005.
http://dx.doi.org/10.1021/op049783p
14. Birch M. et al., Organic Process Research & Development, (9): 360-364, 2005.
http://dx.doi.org/10.1021/op0500077
15. Sarenteas K. et al., Organic Process Research & Development, (9): 911-922, 2005.
http://dx.doi.org/10.1021/op050101n
16. O’Grady AN_kinetics_crystallization-AN-092407. http://us.mt.com/global/en/home/support-
ive_content/application_editorials/small-scale.rxHgAwXLlLnPBMDSzq--.ExternalFileComponent.
html/kinetics_crystallization-AN-092407.pdf
17. Mousaw P., et al,, Organic Process Research & Development, (12): 243-248, 2008.
http://dx.doi.org/10.1021/op700276w
18. Kim, et al, Organic Process Research & Development, (9): 894-901, 2005.
http://dx.doi.org/10.1021/op050091q
19. B. Johnson, C. Szeto, O. Davidson, and A. Andrews, Optimization of Pharmaceutical Batch
Crystallization for Filtration and Scale-up, AIChE Annual Meeting, Los Angeles, 1997.
20. Black S. et al., Organic Organic Process Research & Development (9): 943-950, 2005.

Internet: http://www.mt.com/crystallization

Subject to technical changes

© 08/2011 Mettler-Toledo AutoChem, Inc.
7075 Samuel Morse Drive
Columbia, MD 21046 USA
Telephone +1 410 910 8500
Fax +1 410 910 8600
Email [email protected]