ADVANCED USES OF FREEZEDRYING MICROSCOPY (FDM)
FOR PRODUCT FORMULATION
AND LYO-CYCLE DEVELOPMENT
Kevin R. Ward Ph.D. MRSC
Director of Research and Development,
Biopharma Technology Ltd.
Winchester, UK.
Vials of freeze-dried product
Poor
Good
OK
Poor
The product in the “Poor” vials has become soft and dense
during freeze-drying, because it has become warmer than its
“Critical Temperature”!
2
“How do we know what the Critical
Temperature is for our product?”
• The “Critical Temperature” will be:
– The eutectic temperature (Teu) for crystalline materials
– The collapse temperature (Tc) for amorphous materials
(somewhere at or above the glass transition temperature)
– The lower of the above temperatures for mixed systems
(depending on whether micro-collapse is acceptable)
• We can analyse the critical temperature of a formulation
before freeze-drying it, using, for example:
– Freeze-Drying Microscopy (FDM)
– Impedance (Zsinφ) and Thermal Analysis
3
Freeze-drying microscopy (FDM)
FDM is the study of
freeze-drying at the
microscopic level
FDM allows
determination of
collapse, melting
and “qualitative
phenomena” such
as skin formation
4
What is a Freeze-Drying Microscope?
• Effectively a ‘micro
freeze-dryer’ where
the freeze-drying of a
small sample may be
observed
• First designs in the
mid- 1960s
• Now manufactured
commercially
5
FDM set-up with LN2 cooling system – BTL Lyostat2
Freeze-Drying Stage
Camera
Compound
Microscope
Liquid
nitrogen
pump
Temperature
controller
To vacuum
pump
Liquid
nitrogen
Dewar
Vacuum gauge
6
Sample Preparation for FDM
• Sample loading takes about 60 seconds.
• Routine analysis usually takes 20 - 30 minutes, or up to 60
minutes if heat annealing is used.
Block
Sample
holder
Side
Door
7
Sample Format in Lyostat2
Objective Lens (usually 10 x)
Glass cover slip (13 mm dia.)
Metal Spacer (70µm thick)
2µl of sample
Quartz cover slip (16 mm dia.)
Aperture
Temperature-Controlled Block
Light Source (from below)
8
Sample Loading and Cooling
• Ideally the raw formulation is used
• Sometimes necessary to use samples that
have previously been frozen or lyophilised
• After loading the sample, the Lyostat2 is set
to cool to the desired temperature
• The sample is allowed to cool and freeze
(Note: for eutectic materials, there will be more than
one freezing event!)
9
Sublimation
front
Temp/time
table
INITIAL FDM IMAGE
• When sample reaches
the holding temperature
and has been observed
to freeze, vacuum
pump is switched on
and drying begins.
Dried
sample
Frozen
sample
On-line
plot
• Sublimation interface
can be seen moving
through the frozen
sample.
10
INTERPRETATION OF EVENTS
Collapsed material
• Increasing or decreasing
the temperature of the
sample allows you to
view its freeze-drying
characteristics.
• By examining the freezedried structure behind
the interface, the
collapse temperature of
the material can be
determined.
Sublimation
front
Frozen
sample
• The temperature may be
cycled in order to
evaluate Tc more closely
11
INTERPRETATION OF EVENTS
Regained
structure
Frozen
sample
• Sample structure lost
when collapse
temperature was
exceeded.
• Structure regained as
sample was re-cooled
to below its collapse
temperature.
Collapsed
sample
Sublimation
front
12
INTERPRETATION OF EVENTS
Sublimation
front
Frozen
sample
• 100% structure has
been regained by
lowering the sample
temperature.
• Sample temperature
was again increased to
above its collapse
temperature, causing
the sample to collapse.
Dried
sample with
structure
Collapsing
again on
reheating
13
‘Micro-collapse’ (see e.g. Wang, 2004)
Macroscopically
similar but is it:
Wetter?
Less stable?
More difficult to
reconstitute?
Below Tc of
amorphous phase
Above Tc of
amorphous phase
A similar effect may also be observed due to the melting of
crystalline component(s) onto a rigid amorphous structure
(depending on which has the lower critical temperature)
14
FDM of 2% Mannitol + 1% Glucose
Frozen
material
(Drying front)
Regions of
(micro) collapse.
Just glucose?
Regions with good
dried structure.
Just mannitol?
-41oC, around Tc for glucose. Micro-collapse or just poor solute mixing?
15
So, what else can FDM tell us?
• Eutectic melting temperature
16
NaCl Below Eutectic Temperature
Dry
Frozen
17
NaCl Above Eutectic Temperature
Note
changes in
appearance
of frozen
structure
Eutectic liquid
18
So, what else can FDM tell us?
• Eutectic melting temperature
• May give some indication of skin (crust)
formation potential of a formulation
19
20
21
So, what else can FDM tell us?
• Eutectic melting temperature
• May give some indication of skin (crust)
formation potential of a formulation
• Whether heat-annealing may be of benefit
– To increase ice crystal size – and what
conditions are required for this (above Tg’?)
– To encourage some components to crystallise
22
Effect of annealing on ice crystal size
Sample cooled to -40°C,
then warmed to -10°C
Same sample after a further
15 minutes at -10°C
Experiments can be carried out to compare rates of change at different
temperatures, in order to establish what annealing temperature might
be most efficient to use in the freeze-dryer.
23
FDM setup with polarised light
Camera
Analyser
Sample
Polariser
24
Effect of annealing on solute behaviour:
FDM with polarised light function
KCl solution quench
cooled below -40°C
No sign of crystals
(no light rotation)
Drying at -18°C
Polariser shows presence
of crystals (white areas)
25
Effect of annealing on solute behaviour:
FDM with polarised light function
Start of Eutectic Melt
at -11°C
Eutectic Melt complete
but ice still present.
∴Crystals were indeed
due to eutectic KCl26
Further applications of FDM
• It is possible to examine differences in
relative drying rates:
– For different formulations
– For a specific formulation at different
temperatures
Ref: Zhai, S., Taylor, R., Sanches, R. and N.K.H.
Slater (2003). Measurement of Lyophilisation
primary drying rates by freeze-drying microscopy.
Chem. Eng. Sci. 58, 2313-2323
27
Use of Phase Contrast / DIC
• Phase Contrast may be used to “artificially
colour” different parts of a sample
• The following sequence of slides shows a
solution of mannitol:
– In the liquid state
– After initial freezing
– Following annealing
28
Mannitol solution at 20oC (liquid)
29
Mannitol solution cooled to -30oC
Amorphous solid plus small layer of excluded liquid at edge
30
Warming to
o
-25 C
(note changes in appearance already)
31
Annealing to
o
-5 C
Significant crystallisation, moving in from edge of sample
32
CONCLUSIONS
•FDM can provide a visual indication of:
–
–
–
–
Collapse temperature (Tc)
Eutectic temperature (Teu)
Skin formation potential
Annealing effects: on ice structure, solute
crystallisation, critical temperature
– Relative rates of drying for different formulations, or
for the same formulation at different temperatures
All the above information can be useful for formulation
& cycle development
33