Mannitol is a polyol, a molecule similar to sugar but is unable to form a ring
structure as it only contains hydroxyl functional groups. It is widely used as
an artificial sweetener and in the pharmaceutical industry it is widely used as
a cryoprotectant to preserve the functionality of proteins during freeze drying
or sub zero storage. In order to understand how cryoprotectants protect the active
ingredients it is important to understand their physical behaviour during the processes
the formulation are subjected to. Often a range of cryoprotectants are tested
and the ‘best’ is chosen based on the extent to which the active ingredient has
retained its functionality. Ideally it would be preferable to understand the
process in order to allow design of the formulations, reaching a more rapid optimisation.
Mannitol is an unusual cryoprotectant in
that it forms polymorphic crystals during freezing, in contrast to other common
cryoprotectants such as sucrose and trehalose. Anhydrous mannitol can form a
number of different crystalline structures, alpha, beta and delta (gamma was
found to be a mixture of polymoprhs, not a true one) as well as amorphous, and
there is the added complication of a hemi hydrate form too.
For mannitol it was observed that frozen
solutions of mannitol formed an opaque ring round the edge of the frozen drop
(Fig 1). The samples were assessed using X-Ray crystallographic techniques and
found that the samples were crystallising the mannitol one specific polymorph,
beta. The issue with the X-ray methods were that it was slow (unable to
determine any kinetic changes during annealing) and could not be used to
indicate the localisation of the crystalline material.
Raman spectroscopy was chosen to shed
further light on this issue as it is sensitive to the physical
structure of the sample, including crystalline packing and conformational
changes within the individual molecules. For this study only the CH an OH
stretching regions were analysed as these contained sufficient information to discriminate
the polymorphs and to provide insight into the changes in ice structure associated
with the annealing. The Raman signals for each of the mannitol polymorphs is
unique, as is the signal from amorphous and hemi-hydrate (Figure 2). In
fact the amorphous signal contains some
structure, with a doublet of bands at the high and low wavenumber sides
of the CH
region, dependant on the temperature of the sample. All amorphous
manitol share a dominant broad bell shaped curve (labelled as
amorphous/common). At high temperatures a doublet of peaks appears
at the low wavenumber side of the amorphous curve, with the modes
arising from the hydroxyls shifting to a very high position. The
position of OH stretching modes is inversely related to
the strength of the hydrogen bonding, suggesting that in the
molten mannitol the hydrogen bonding between the molecules is very
weak. In contrast the OH modes in the solution phase amorphic mannitol
exhibit a shift towards a lower wavenumber, and a considerable increase
in intensity. This would suggest stronger hydrogen bond with an
increase in OH modes involved - i.e. the formation of a solvation
Principal component analysis of the
annealing dataset revealed two main principal components associated with the
mannitol, as shown in Figure 3. The main mannitol PC was dominated by the beta
form, but also showed significant contribution from the delta form. The next PC
was equivalent to the subtraction of the signal arising from the beta polymorph
from that of the delta polymorph, revealing that PC measured the balance of
these two polymorphs. The other polymorphic forms were present in much smaller
quantities. Figure 4 shows the trends with annealing time for these two PCs. It
was known from calorimetry that some kind of change of physical structure
change was occurring over the first 40 minutes (see bottom trends on Figure 4),
which then tapered off. The Raman results unequivocally demonstrate that this
is not due to crystallisation from amorphous mannitol, but that the mannitol
spontaneously crystallises, and does so predominantly in the delta form (PC4
centre data). The thermal events observed in the DSC experiments do in fact
arise from a delta to beta polymorphic transition.
Curiously, the edge trend differs significantly
from the centre of the drop, with almost all the mannitol crystallising as beta
form initially, then some delta creeps in, presumably as delta crystals from
the interior are pushed outwards. Ultimately though, all the mannitol gradually
converts to the beta form (based on overnight annealing, data not shown).
The top row in Figure 5 shows an animation
of the distribution changes of the crystalline mannitol (green) in the frozen
drop as it is annealed. At the edge there is a clear directionality in the
trend, whereas in the centre the changes are more chaotic and are non-directional
as would be anticipated from random diffusion. The second row highlights the conversion from
delta to beta and it is readily appreciable that in the first frame (03 min)
the edge is a mixture of red and green (delta and beta), while the centre is
almost exclusively delta, in agreement with figure 4 above. Over time the delta
converts to beta, reflected in a conversion from red to green in Figure 5
This data demonstrated the first direct
measurement of mannitol crystallisation behaviour in a frozen solution,
incorporating assessment of time dependency, spatial distribution, diffusion
and polymorphic transitioning. In addition to the insights into the mannitol
the Raman data also unveiled key discoveries regarding the role of the
solvation shell in the process
1. Beattie, J. R., et
al., European Journal of Pharmaceutics
and Biopharmaceutics 2007, 67,
2. Campbell Roberts,
S. N., et al., Journal of Pharmaceutical
and Biomedical Analysis 2002,
28, (6), 1135-1147.
3. Burger, A., et al.,
J. Pharm. Sci. 2000, 89, (4), 457-468.
4. Reinhard, V., Applied Spectroscopy 2005, 59, (3), 286-292.
5. Telang, C., et al.,
Pharmaceutical Research 2003, 20, (12), 1939-1945.
6. Okumura, T.;
Otsuka, M., Pharmaceutical Research 2005, 22, (8), 1350-1357.
7. Vehring, R., Applied Spectroscopy 2005, 59, (3), 286-292.
8. Roy, A., et al., Free Radical Research 2004, 38, (2), 139-146.
9. Auer, M. E., et
al., Journal of Molecular Structure 2003, 661, 307-317.
10. Roberts, S. N. C.,
et al., Journal of Pharmaceutical and
Biomedical Analysis 2002, 28,
11. Braun, D. E., et
al., International Journal of
Pharmaceutics 2010, 385, (1-2), 29-36.
12. Hulse, W. L., et
al., Drug Development and Industrial
Pharmacy 2009, 35, (6), 712-718.
13. Ye, P.; Byron, T., American Laboratory 2008, 40, (14), 24-27.