Crystallization Guide

Crystallization and Precipitation

Guide to Effective Process Development

Author: Des O’Grady PhD, METTLER TOLEDO

Crystallization is recognized as a common unit operation used to isolate and purify

product in a number of industries. However, the scientific fundamentals of crystallization
are typically not as well-known as they are for other unit operations such as distillation
or absorption. The crystallization step directly influences product attributes such as
crystal size, shape, yield and purity which offers great potential to improve both the
quality of the final product and the efficiency of the process.

In the first part of this two-part white paper the fundamentals of crystallization will be intro-
duced and common-sense guidelines for the design of a high quality crystallization process
will be presented. Using case studies and references from industry and academia, key
crystallization topics such as solubility, supersaturation and crystallization kinetics will be
explained with examples of how each can be utilized to make informed decisions regarding
effective crystallization process development. The role of technology in crystallization devel-
opment, from effectively controlling crystallization process parameters to monitoring crystal
size and supersaturation in situ, will also be presented.

Contents
1 Introduction to Crystallization and Precipitation
2 Common Ways to Reduce Solubility and
Drive Crystallization
3 Supersaturation: The Driving Force for Crystallization
and Nucleation Growth
4 The Importance of Crystal Size and Shape Distribution
5 Conclusion
1 Introduction to Crystallization and Precipitation
Crystallization Guide

Crystallization touches every aspect of our lives from the foods we eat and the medicines we
take, to the fuels we use to power our communities. The majority of agrochemical and
pharmaceutical products go through many crystallization steps during their development and
manufacture, key food ingredients such as lactose and lysine are delivered to humans and
animals as crystals and the unwanted crystallization of gas hydrates in deep sea pipelines is
a major safety concern for the petrochemical industry.

Scientists and engineers working in many industries around the world are required to under-
stand, optimize and control crystallization processes every day. The purpose of this guide is
to introduce key crystallization concepts and highlight the many resources available for those
working in this exciting field.

Crystallization: A process whereby solid crystals are formed from another phase,
typically a liquid solution or melt

Crystal: Solid particles in which the constituent molecules, atoms or ions are arranged
in some fixed and rigid repeating three-dimensional pattern or lattice

Precipitation: It is a little bit difficult to define the term precipitation. For some, it
is simply a very fast, perhaps uncontrolled, crystallization process. For others it is
crystal formation resulting from a chemical reaction. It can also vary by industry; in
the pharmaceutical industry “crystallization” is common and in the chemical industry
“precipitation” is the vernacular.

The prevalence of crystallization processes in industry

can likely be attributed to the fact that crystallization Product Performance1
acts as both a separation and purification step. In a “Crystallization of an API [ac-
tive pharmaceutical ingredient
single step, crystal product of the desired purity can
– the crystal product] in par-
be created and then isolated. Despite this obvious ticular is critical for product
advantage, crystallization processes still need to qualities such as chemical
be understood and controlled to the desired crystal purity and correct polymor-
product quality and to ensure an efficient and cost- phic form, which need to be
effective crystallization process. strictly controlled to meet set
specifications”.

While crystals have many important attributes the

crystal size distribution probably has the greatest
Process Performance1
impact on the quality and effectiveness of the final “The API crystallization
product (and the process needed to deliver it). process and crystal proper-
Crystal size and shape directly influence key steps ties have a significant effect
downstream from the crystallizer with filtration and on downstream processing.
drying performance being particularly susceptible to For example, excess fines or
wide particle size distribution
changes in these important attributes. Similarly the
may cause slow filtration
final crystal size can also directly influence the quality and inefficient drying, which
of the final product. In a pharmaceutical compound, may be a major bottleneck
bioavailability and efficacy are often related to particle of the entire manufacturing
size with smaller particles often desired for their process”.
enhanced solubility and dissolution characteristics.

2 Crystallization and Precipitation

The crystal size distribution of ice plays a vital role in the taste and mouthfeel of ice cream
with crystals smaller than 50 μm better than crystals larger than 100 μm. For agrochemicals
it is vital to ensure that particles are small enough to be sprayed without blocking nozzles
while large enough not to drift into neighboring fields.

Crystal size distribution is influenced by a number of variables which can be optimized

and controlled during crystallization development and manufacturing. In turn, crystal size
distribution influences both product and process performance. Figure 1-1 highlights how
choosing the correct input properties and process parameters for a crystallization can
influence the transformations that occur and the product and process attributes that result
once the crystallizaton is complete.

Input Properties Process Parameters

• Concentration • Agitation
• Solvent Type • Cooling Rate
• Seeded vs. Unseeded • Hold Time
• Scale • Seeding Protocol
• Temperature • Anti-solvent Addition
• Impurity Profile • Temperature Cycle

Transformation
• Nucleation (Primary/Secondary)
• Growth
• Phase Separation
• Form Convertion/Habit Shift
• Attrition
• Agglomeration

Product Performance Process Performance

• Particle Size • Filterability
• Purity • Drying
• Yield • Formulability
• Dissolution Rate • Scale-ability
• Flowability
• Robustness
• Repeatability

Figure 1-1. Relationship between common crystallization parameters, transformations and attributes2

1. Kim S. et al., “Control of the Particle Properties of a Drug Substance by Crystallization Engineering and the Effect
on Drug Product Formulation” Organic Process Research & Development, 9, 894-901 (2005).
2. Adapted from Powder Technology Volume 189, Issue 2, 31 January 2009, Pages 313–317.

Crystallization and Precipitation 3

2 Common Ways to Reduce Solubility and
Crystallization Guide
Drive Crystallization

Crystallization is generally achieved by reducing the solubility of the product in a saturated

starting solution by cooling, adding antisolvent, evaporation or some combination of these
methods. Another common method used to induce crystallization is via a chemical reaction
where two or more reactants are mixed to form a solid product insoluble in the reaction
mixture; a common example of this would be the reaction of an acid and a base to form
a salt.

The method chosen to crystallize product can vary greatly depending on a number of
factors. For example, protein crystals are temperature-sensitive ruling out cooling and
evaporation and leaving antisolvent addition as the most common crystallization method.
For many crystallization processes, cooling can be advantageous as it is reversible; the
saturated solution can be reheated in the event of a non-optimal operation. Many large scale
industrial crystallization processes rely on evaporation to produce common products such as
potassium nitrate and ammonium chloride.

Regardless of the method chosen a common feature is starting with a saturated solution.

Saturated Solution and Solubility: At a given temperature there is a maximum quantity

of solute that can be dissolved in a given solvent. At this point, the solution is saturated.
The quantity of solute dissolved at this point is the solubility.

Units: Solubility is usually reported as:

- g of solute/100 g of solvent
- g of solute/L of solution
- mole fraction
- mole %

Solubility curves (Figure 2-1) are 80

commonly used to illustrate the
70
relationship between solubility, Solvent A
Good Solvent
Solubility (g Solute / 100 g Solvent)

Solvent B
temperature and solvent type. By 60 Solvent C

plotting temperature vs. solubility,

50
scientists create the framework
needed to develop the desired 40

crystallization process. 30

In Figure 2-1, the solubility of 20

Good Antisolvent
the given material in Solvent A is 10
high – meaning more material
can be crystallized per unit mass 0
0 20 40 60 80
of solvent. Solvent C has a low Temperature (°C)
solubility at all temperatures, Figure 2-1. Solubility of model compound in three solvents
indicating it could be a useful
antisolvent for this material.

4 Crystallization and Precipitation

Solubility curves also reveal the theoretical yield for a given crystallization process. In Figure
2-1 if a saturated solution containing 50 g of product per 100 g of Solvent A is cooled from
60 °C to 10 °C then it is clear that 10g of product per 100 g of solvent will remain in solution.
In other words exactly 40 g of product per 100 g of solvent should crystallize. This allows
scientists to compare the theoretical yield with the actual yield and define the efficiency of the
crystallization process.

Once an appropriate solvent is chosen, the solubility curve becomes a critical tool for
the development of an effective crystallization process. With this information the starting
concentration and temperature (or antisolvent ratio) can be chosen and the first important
decisions regarding how the crystallization will be developed can be made. For example if
seeds are to be used to induce crystal nucleation it is critical to add them at a temperature
lower than the solubility point – otherwise they will dissolve.

Example Case Study3

Point of Nucleation Point of Dissolution
ParticleTrack (a probe based
instrument that tracks the rate and
degree of change to particle size 12000 55
and count as particles exist in
process) can be used to measure 10000
the solubility curve and MSZW by 45
accurately identifying the point of 8000
Counts per Second

dissolution (point on the solubility

Temperature (°C)
curve) and point of nucleation 6000 35
(point on the MSZW) at various
solute concentrations. In Figure 4000
2-2, an undersaturated solution is Temperature (°C)
25
Particle Count (0-20 µm)
cooled at a slow, fixed rate until 2000 Particle Count (50-250 µm)

the point of nucleation is measured

by ParticleTrack, indicating a point 0 15
on the MSZW. Next, the solution 0 20 40 60 80 100
Time (mins)
is heated slowly until the point of
dissolution is measured indicating Figure 2-2. Points of nucleation and dissolution during cooling and heating
a point on the solubility curve.

Solvent is then added to the system

to reduce the concentration and the
process is repeated. In this way,
the solubility curve and MSZW
can be measured rapidly over a
wide range of temperatures. This
process can be automated using
a synthesis workstation, such
as EasyMax® or OptiMax™ by
introducing a feedback loop where
the ParticleTrack signal can be used
Figure 2-3. Automated crystallization workstation with tight control over process param-
to initiate the heating, cooling and eters and probe based tools for real time unattended characterization (ParticleTrack and
dilution steps. ParticleView probes monitoring in an EasyMax synthesis workstation)

Crystallization and Precipitation 5

In Figure 2-4, the solubility curve
Crystallization Guide
40
and metastable zone width(s) for Solubility Data

potassium aluminum sulphate is MSZW (C.15 °C per min)

MSZW (C.7 °C per min)

shown. While the solubility curve

is thermodynamically fixed for a 30
given solvent-solute system the

Concentration (g hydrate/ 100 g water)

MSZW is a kinetic boundary and
can change depending on process 20
parameters such as cooling rate,
agitation or scale. Characterizing
the MSZW under a range of process
10
conditions can help scientists
understand how a crystallization
process may behave at different
0
scales - or in the event of a process 0 10 20 30 40 50 60
upset. Variability in the MSZW under Temperature (°C)
different conditions may indicate Figure 2-4. Solubility curve and MSZW for potassium aluminum sulphate.
that the system may not behave
consistently in terms of nucleation point and kinetics. Such a result may justify investigating the possibility of
seeding the process in order to fix the nucleation point for every experiment or batch.

Dynamic approaches to solubility determination, such as this one, are sometimes limited in their accuracy
since a fast heating rate means the exact point of dissolution can be overestimated. Static methods, such as
gravimetric analysis may offer more accuracy – but are more time-consuming and cumbersome to implement.
Many techniques can be used to measure solubility curves1-4 and recent research aimed at predicting solubility
in different solvents is showing promise5. The references below offer a good starting point for more in depth
study.

1. Howard K. Zimmerman, The Experimental Determination of Solubilities, Jr. Chem. Rev., 1952, 51 (1), pp 25–65
2. Granberg & Rasmusson, Solubility of Paracetamol in Pure Solvents, J. Chem. Eng. Data, 1999, 44 (6), pp
1391–139.
3. Barrett and Glennon, “Characterizing the Metastable Zone Sidth and Solubility Curve Using Lasentec FBRM and
PVM,” Trans ICHemE, vol. 80, 2002, pp. 799-805.
4. Mark Barrett, Mairtin McNamara, HongXun Hao, Paul Barrett, Brian Glennon, Supersaturation tracking for the
development, optimization and control of crystallization processes, Chemical Engineering Research and Design,
Volume 88, Issue 8, August 2010, Pages 1108-1119.
5. Abraham, M. H., Smith, R. E., Luchtefeld, R., Boorem, A. J., Luo, R. and Acree, W. E. (2010), Prediction of solubility
of drugs and other compounds in organic solvents. J. Pharm. Sci., 99: 1500–1515.

6 Crystallization and Precipitation

3 Supersaturation: The Driving Force for Crystal
Nucleation and Growth

Crystallization scientists gain control of crystallization processes by carefully controlling the

prevailing level of supersaturation during the process.

Supersaturation: The difference between the actual concentration and the solubility
concentration at a given temperature is defined as the supersaturation (ΔC)

Figure 3-1 illustrates the concept

of supersaturation and expands
on the role of the metastable
zone width, the kinetic boundary
at which crystallization occurs. Nucleation Cooling

Concentration
When a saturated solution is
gradually cooled crystals will Metastable Limit ∆C (Supersaturation)
“crash out” at a given temperature.
It seems intuitive to think that this
temperature may be the same as Solubility

the solubility temperature – the

temperature at which the last Temperature
crystal dissolved during heating. Figure 3-1. Solubility curce and MSZW (Metastable Zone Width)
However, when a saturated solution
is cooled the system enters a metastable region where the solution becomes supersaturated,
in other words, more of the solute is in solution that the solubility curve predicts. As cooling
continues, a certain temperature will be reached where crystal nucleation will occur. This
point is called the metastable limit and the difference in temperature between this point and
the solubility curve is called the metastable zone width. Once the metastable limit is reached
and crystallization starts supersaturation is consumed and eventually the liquid phase
concentration will reach equilibrium at the solubility curve.

g b
It is critical to understand the
g G = k ∆C B = k ∆C b
concept of supersaturation because
it is the driving force for crystal G = Growth Rate B = Nucleation Rate
nucleation and growth, and as
kg = Growth Constant kb = Nucleation Constant
such, will ultimately dictate the final
g = Growth Order b = Nucleation Order
crystal size distribution. Nucleation
is the birth of new crystal nuclei ΔC = Supersaturation ΔC = Supersaturation
– either spontaneously from Table 3-1. Equations that define relationship between supersaturation, nucleation
solution (primary nucleation) and growth rate.1
or in the presence of existing
crystals (secondary nucleation). Crystal growth is the increase in size (or more accurately
“characteristic length”) of crystals as solute is deposited from solution. The relationship
between supersaturation, nucleation and growth is defined by a well known set of (somewhat
simplified) equations (Table 3-1).

Crystallization and Precipitation 7

For organic crystallization systems, the value of the
Crystallization Guide
Crystal Size
growth order (g) is typically between 1 and 2 and the

Nucleation Rate, Growth Rate, Crystal Size

value of the nucleation order (b) is typically between 5
and 10. When we plot these equations for a theoretical
organic crystallization process the importance of
supersaturation becomes clear. At low supersaturation,
crystals can grow faster than they nucleate resulting
Nucleation Rate
in a larger crystal size distribution. However, at higher Growth Rate
supersaturation, crystal nucleation dominates crystal
growth, ultimately resulting in smaller crystals. Figure
3-2, relating supersaturation to nucleation, growth
and crystal size clearly illustrates how controlling Supersaturation
supersaturation is vitally important when it comes to Figure 3-2. Plot showing the relationship between supersaturation,
creating crystals of the desired size and distribution. growth and nucleation rate

Example Case Study5

Peak Height Associated
with Benzoic Acid
Supersaturation can be monitored 0.05
Sample in 0.075 g Benz / g Sol.
in situ and in real time by Background and Sample 0.04
1 g Water / g Ethanol Solution
measuring the liquid phase
concentration with ReactIR (ATR- 0.03

Peak Height
FTIR). In this example, researchers 0.02
looked at the cooling crystallization
of benzoic acid from ethanol water 0.01

mixtures and utilized RectIR to 0

monitor supresaturation in situ 2650 2150 1650 1150 650
and in real time. A novel approach Wavenumber (cm-1) -0.01

is to eliminate the need for a

calibration step and simply monitor Figure 3-3. Supersaturation study utilizing in situ ReactIR monitoring.
a peak in the ReactIR spectrum
that is characteristic of the solute
cocentration (Figure 3-3). By Point of Dissolution
heating a crystal suspension slowly 15000
a solubility trace (peak height as

Counts per Second (ParticleTrack)

a function of temperature) can be 0.42 12500

created. As temperature increases,

Peak Height (ReactIR)

ReactIR Solubility Curve

10000
crystals dissolve and solute goes FBRM Fines #/sec
0.38
into solution. If the heating rate 7500

chosen is slow enough it can

5000
be assumed that the system is 0.34
effectively at the solubility point at 2500

each temperature, and that a plot of

0.30 0
temperature vs. peak height is an 0 10 20 30 40 50 60
effective way to trace the solubility Temperature (°C)

curve without calibration(Figure Figure 3-4. Plots Peak Height, Temperature and Counts to determine the point of
3-4). While this approach may not dissolution.
be as accurate as using a static
method (outlined in section 2) the simplicity and real time nature of this method make it a
valuable approach that can be deployed routinely during crystallization development.

8 Crystallization and Precipitation

Once a solubility trace is determined, crystallization processes can be run under varying
process conditions and the effect of these conditions on supersaturation can be observed.
Faster cooling rates result in nucleation at lower temperatures and the highest level of
supersaturation throughout the process. A very slow cool down results in a higher nucleation
temperature and low supersaturation throughout the process. A one-hour cubic cool down
(slow at first and fast at the end) has a medium level of supersaturation throughout.

The influence of varying supersaturation on crystal size and shape distribution can be clearly
observed by comparing ParticleView (a probe based real-time microscope) images for each
experiment. Higher supersaturation results in the smallest crystals – since nucleation will be
favored over growth. The opposite is true for the slowest cool down.

100 µm 100 µm 100 µm

1 Hr Cool Down 1 Hr Cubic Cool Down 15 Hr Cool Down

Point of Nucleation Solubility Curve

1 Hr Cool Down
0.39 1 Hr Cubic Cool Down
15 Hr Cool Down

0.38
Peak Height

0.37

0.36

0.35

0.34
0 5 10 15 20 25 30
Temperature (°C)

Figure 3-5. Temperature vs. peak height for three experiments with different cooling profiles

The fundamentals of crystallization and the relationship between supersaturation and

crystallization kinetics are covered in great detail in these references.

1. Jaroslav Nývlt, Kinetics of nucleation in solutions, Journal of Crystal Growth, Volumes 3–4, 1968, Pages
377-383.
2. John Garside, Industrial crystallization from solution, Chemical Engineering Science, Volume 40, Issue 1, 1985,
Pages 3-26.
3. D. O’Grady, M. Barrett, E. Casey, B. Glennon, The Effect of Mixing on the Metastable Zone Width and Nucleation
Kinetics in the Anti-Solvent Crystallization of Benzoic Acid, Chemical Engineering Research and Design, Volume
85, Issue 7, 2007, Pages 945-952.
4. Noriaki Kubota, Masaaki Yokota, J.W. Mullin, Supersaturation dependence of crystal growth in solutions in the
presence of impurity, Journal of Crystal Growth, Volume 182, Issues 1–2, 1 December 1997, Pages 86-94.
5. Mark Barrett, Mairtin McNamara, HongXun Hao, Paul Barrett, Brian Glennon, Supersaturation tracking for the
development, optimization and control of crystallization processes, Chemical Engineering Research and Design,
Volume 88, Issue 8, August 2010, Pages 1108-1119.

Crystallization and Precipitation 9

4 The Importance of Crystal Size and
Crystallization Guide
Shape Distribution

This set of ParticleView images (Figure 4-1) neatly illustrates a.

the complex size, shape and structure of various crystals.
From large round “boulders” to beautifully delicate “dendrites”,
crystal product is often varied, posing challenges to effective
separation and downstream manipulation.

Filtration or centrifugation is typically the step that comes after

crystallization and the crystal size and shape can greatly af-
fect the efficiency of this unit operation. It is not much use to
150 µm
design a crystallization that is complete in one hour if it ends
up taking 24 hours to filter! Looking at the ParticleView images
some clues as to how these different crystal products will filter b.
can be gathered:

a. These crystals will likely filter quickly and consistently. The

larger boulders will leave plenty of space for the filtrate to
pass through rapidly.
b. Flat plates like these can be some of the most difficult to fil-
ter. Plates tend to stack on top of each other creating a layer
of crystals that the filtrate cannot get through. This leads to
150 µm
long and potentially variable filtration times, depending on
how the crystals are discharged from the crystallizer.
c. This is another case where filtration times can be long.
c.
Small crystals will plug the gaps left by the larger crystals
making it difficult for the filtrate to pass through the bed
of crystals. This is a common problem because many
crystallization processes are designed with a fast cool, or
antisolvent addition step at the end of the crystallization (to
increase yield) that leads to excessive secondary nucle-
ation. Additionally, in many cases the agitation is increased
at the end of the batch to help with discharge and this leads
to crystal breakage. 150 µm

d. This image is more common than many would expect, at

least in organic crystallization systems that are seeded. A
structure such as this would be difficult to observe using
d.
an offline microscope as it will be crushed during sampling
and preparation. However, ParticleView reveals a beautiful
dendritic structure. A dendrite such as this often forms when
a crystallization is seeded with milled seed. Imperfections
on the crystal surface lead to crystal growth from these ar-
eas and long crystal branches growing from a seed core. It
is difficult to predict how something like this will filter but it
is likely to break apart resulting in variable filtration times. 150 µm

Figure 4-1: ParticleView real-time microscope images

10 Crystallization and Precipitation

Filtration is just one aspect of crystallization where particle size is important. For many
products, the crystal size impacts the effectiveness of the product; for example the rate
of absorption of a pharmaceutical drug in the body or the burn rate of a highly energetic
material. Other aspects of the process can also be influenced by particle size and shape – for
example flowability. Table 4-1 highlights some other areas where crystal size plays a key role
in various applications.

Application Product Quality Process Quality

Active Pharmaceultical Speed of action (small crystals disolve faster Drying rates (large crystals have a lower surface
Ingredients allowing medicines to work faster) area to volume ratio and may dry more slowly)
Energetic Materials Burn rate (small crystals burn faster and with Safe handling (small crystals can be volatile and
higher engery) unsafe)
Agrochemicals Segregation (large crystals sink to the bot- Yield (small crystals can bind or block filters
tom of formulation tanks and are not sprayed resulting in multiple washes and the reduction in
evenly) yield)
Sugar Flowability (uniform crystal size makes for Poor contol (small crystals, resulting from exces-
better product flowability) sive nucleation, can destabilize a process)
Amino Acids Bioavailability (small crystals dissolve Throughput (wide crystal size distributions can be
faster) difficult to stir and transport through a process
Bulk Chemicals Price (crystals of different size can be sold at Centrifugation (small crystals can reduce centrifu-
different prices for different purposes) ge performance and cause breakdowns)
Table 4-1. Crystal size vs product and process quality for various applications

Example Case Study

12500 Increased Fines Generation 70
(Secondary Nucleation)
Counts (Fines, 0-20 µm)

10000

Temperature (°C)
The process conditions chosen 60
Temperature (°C)
can greatly influence crystal size, 7500
Fines (#/sec, 0-20 µm)

either by affecting supersaturation

5000
and kinetics (Chapter 3) or by more 50
Increased Cooling Rate
physical mechanisms such as 2500
crystal breakage or agglomeration,
0 40
which are both influenced by 04:00 08:00 12:00 16:00 20:00 24:00
mixing conditions in the crystallizer. Relative Time (hr:min)
Figure 4-2. ParticleTrack data: time vs. temperature and counts (0-20μm) showing faster
Figures 4-2 and 4-3 show how cooling causes secondary nucleation

increasing the cooling rate at the

600
end of a crystallization experiment Cooling Ramp 1
Cooling Ramp 2
100 µm

results in a secondary nucleation 500

event and a more rapid increase in
Counts per Second

400
fine crystal counts (less than
More fines, 100 µm
20 μm). An increase in cooling rate 300
smaller crystal size
generates more supersaturation -
200
which is consumed by nucleation
rather than growth. Careful control 100
of cooling rates at every point in a
0
crystallization is vital to ensure the 1 10 100 1000
target crystal size distribution is Chord Length (µm)
obtained. Figure 4-3. ParticleTrack data with confirmation from ParticleView images show how more
small crystals appear after the second cooling ramp

Crystallization and Precipitation 11

In Figure 4-4, the influence of a high shear wet mill is shown on crystal size and shape.
Crystallization Guide
Initially rod-like crystals are present but as wet-milling progresses the crystals size decreases
and the crystal count increases. ParticleTrack and ParticleView monitor this progression and
can support targeting an endpoint that delivers crystals of the most appropriate size and
shape in a repeatable way.

a. Begin Milling Endpoint Not Yet Reached

90
13000

80

Mean Chord Length (µm)

12000
Counts per Second

11000 70
Particle Count (Fines)
Mean Dimension

10000
60

9000
50
00:00 05:00 10:00 15:00 20:00 25:00 30:00
Relative Time (hr:min)

b.
100 µm 100 µm

Figure 4-4: a. Time vs. Counts and mean chord length for a wet-milling process; b. ParticleView images taken at
7 mins (left) and 30 mins (right) for the same process.

12 Crystallization and Precipitation

5 Conclusion

In the preceding chapters some of the fundamental aspects of crystallization development

have been covered - with special attention paid to manipulating crystal size and shape
distribution. Sound understanding of the fundamentals is vital to ensure an effective
crystallization process can be developed and scaled up to manufacturing, however a
number of topics have not been addressed. In a follow up white paper, a number of other
crystallization challenges and solutions will be addressed including seeding, scale-up and
how to choose the correct process parameters to ensure target process and product attributes
are obtained.

White Paper: Best Practice for Inline Particle

Size Characterization

This white paper outlines how scientists and engineers can improve
process understanding, product quality and process performance by
applying inline particle size and count measurements. By implementing
inline technologies, measurement errors and variability caused by offline
sampling and sample preparation methods are eliminated.

www.mt.com/wp-E25

White Paper: Understand Crystallization with

In Situ Microscopy

This White Paper demonstrates that in situ microscopy offers a faster

alternative to traditional offline visualization methods and discusses how
GlaxoSmithKline, Merck, Sintef, and University College Dublin (UCD) use
in situ microscopy to gain a unique perspective of their crystallization
processes.

www.mt.com/wp-PVM

Crystallization and Precipitation 13

Crystallization Guide

Appendix A: ParticleTrack with

FBRM® (Focused Beam
Reflectance Measurement)

Measurement for optimization in real time – ParticleTrack is a precise and sensitive

technology which tracks changes to particle dimension, particle shape, and particle count.
Over a wide detection range from 0.5 to 2000 µm, measurements are acquired in real time
while particles are forming and can still be modified enabling process optimization and
control. No sampling or sample preparation is required – even in highly concentrated (70 %
and higher) and opaque suspensions.

www.mt.com/ParticleTrack

Figure c. Chord Length

Distributions
Laser Source

Laser Return

Optics Module
1 2 3 4

Figure b.

Figure a. Figure d. Trended

Statistics
Sapphire Window

How does ParticleTrack work?

The ParticleTrack probe is immersed into a another edge. Thousands of individual chord
dilute or concentrated flowing slurry, droplet lengths are typically measured each second
emulsion, or fluidized particle system. A to produce the Chord Length Distribution
laser is focused to a fine spot at the sapphire (CLD) (Figure c). The CLD is a “fingerprint” of
window interface (Figure a). A magnified the particle system, and provides statistics
view shows individual particle structures will to detect and monitor changes in particle
backscatter the laser light back to the probe dimension and count in real time (Figure d).
(Figure b). These pulses of backscattered
light are detected by the probe and translated Unlike other particle analysis techniques,
into Chord Lengths based on the calculation ParticleTrack measurement makes no
of the scan speed (velocity) multiplied by the assumption of particle shape. This allows
pulse width (time). A chord length is simply the fundamental measurement to be used
defined as the straight line distance from to directly track changes in the particle
one edge of a particle or particle structure to dimension, shape, and count.

14 Crystallization and Precipitation

Appendix B: ParticleView with PVM® (Particle and
Vision Microscope)

Vision for understanding and optimization – ParticleView is a real-time probe based vision
tool which provides instant critical insight into crystal, particle, and droplet systems.
ParticleView enables chemists and engineers to detect and understand process changes that
could take months to discover with traditional offline microscopy techniques.

www.mt.com/ParticleView

Light Source

Camera

PVM Optics
Particle System

Sapphire Window

How does ParticleView work?

ParticleView uses a high
resolution camera and internal
illumination source to obtain
high quality images even
in dark and concentrated 200 µm

suspensions or emulsions. With

no calibration needed and easy
data interpretation, ParticleView
quickly provides critical knowledge
of crystal, particle, and droplet
behavior.

Left: ParticleView inline image;

Right: Offline microscope

Crystallization and Precipitation 15

References and Further Reading
1. Kim S. et al., “Control of the Particle Properties of a Drug Substance by Crystallization
Engineering and the Effect on Drug Product Formulation” Organic Process Research &
Development, 9, 894-901 (2005).
2. Howard K. Zimmerman, The Experimental Determination of Solubilities, Jr. Chem. Rev.,
1952, 51 (1), pp 25–65.
3. Granberg & Rasmusson, Solubility of Paracetamol in Pure Solvents, J. Chem. Eng. Data,
1999, 44 (6), pp 1391–139.
4. Barrett and Glennon, “Characterizing the Metastable Zone Sidth and Solubility Curve Using
Lasentec FBRM and PVM,” Trans ICHemE, vol. 80, 2002, pp. 799-805.
5. Mark Barrett, Mairtin McNamara, HongXun Hao, Paul Barrett, Brian Glennon,
Supersaturation tracking for the development, optimization and control of crystallization
processes, Chemical Engineering Research and Design, Volume 88, Issue 8, August
2010, Pages 1108-1119.
6. Abraham, M. H., Smith, R. E., Luchtefeld, R., Boorem, A. J., Luo, R. and Acree, W. E.
(2010), Prediction of solubility of drugs and other compounds in organic solvents. J.
Pharm. Sci., 99: 1500–1515.
7. Jaroslav Nývlt, Kinetics of nucleation in solutions, Journal of Crystal Growth, Volumes
3–4, 1968, Pages 377-383.
8. John Garside, Industrial crystallization from solution, Chemical Engineering Science,
Volume 40, Issue 1, 1985, Pages 3-26.
9. D. O’Grady, M. Barrett, E. Casey, B. Glennon, The Effect of Mixing on the Metastable Zone
Width and Nucleation Kinetics in the Anti-Solvent Crystallization of Benzoic Acid, Chemical
Engineering Research and Design, Volume 85, Issue 7, 2007, Pages 945-952.
10. Noriaki Kubota, Masaaki Yokota, J.W. Mullin, Supersaturation dependence of crystal
growth in solutions in the presence of impurity, Journal of Crystal Growth, Volume 182,
Issues 1–2, 1 December 1997, Pages 86-94.

Mettler-Toledo AutoChem, Inc.

7075 Samuel Morse Drive
www.mt.com/Crystallization
For more information
Columbia, MD 21046 USA
Telephone +1 410 910 8500
Fax +1 410 910 8600
Email [email protected]

Subject to technical changes

©11/2014 Mettler-Toledo AutoChem, Inc.