Pharmaceutical Development and Technology, 2(3), 257-264 (1997)

RESEARCH ARTICLE

Glass Fragility and the Stability of

Pharmaceutical Preparations-Excipient
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

Selection
Ross H. M. Hatley
Quadrant Healthcare plc, Maris Lane, Trumpington, Cambridge, CB2 2SY, UK

Received April 15, 1996; Accepted February 18, 1997

ABSTRACT

The objectives of this study were, first, to calculate the zero mobility temperatures, To, of treha-
lose and sucrose by the Pikal method from the width of the glass transition and compare these to
For personal use only.

the literature, obtained by enthalpy relaxation measurement, and second, to compare the To val-
ues and physicochemical properties of trehalose to those of sucrose in t e r n of potential to stabi-
lize labile actives in the glassy state. Differential scanning calorimetry and coulometric Karl-Fischer
analysis were used. The glass transition temperatures, Tg,for the two carbohydrates at circa 0.7%
moisture were 101 "C and 64°Cfor trehalose and sucrose, respectively. Anhydrous amorphous tre-
halose had a Tg of 116°C. The To values were found to be 44 and 3.5"C for trehalose and sucrose,
respectively. The Tg- To value for sucrose was compared, and found to be in good agreement with
that found by enthalpy relaxation measurements. Trehalose was found to be resistant to crystalli-
zation above the glass temperature. The study supports the validity of the calculation method pro-
posed by Pikal for T,. It has been proposed in the literature that To is a better measure of stability
than Tg. Trehalose has a signijicantly higher To than sucrose and thus would work more effectively
in stabilizing a labile active.
KEY WORDS: Carbohydrate glass; Fragility; Glass transition; Stability; Sucrose; TD*Trehalose.

INTRODUCTION (2). The glassy preparation derives its stabilizing prop-

erties from its high viscosity. A considerable body of lit-
Ideally, it is desirable that a formulation of a drug erature now exists on the formulation of pharmaceuti-
substance is in the liquid state, since this offers the most cals with respect to formation of glassy products by
convenient and cost-effective method of manufacture. drying (3-15). In essence the principal is to add a glass-
However, this is often not possible because the liquid forming excipient (typically a carbohydrate, polymer,
state can give rise to stability problems and the dry state protein, or amino acid) to a solution of the active ma-
is therefore the necessary condition for the stabilization terial and to dry this solution under conditions that re-
of many drug substances, particularly parenterals (1). sult in the formation of an amorphous glassy matrix in
The most stable dry state condition is the crystalline which .the active is protected from degradation.
state (l), but many potentially useful actives, e.g., pro- The formation of an effective stabilizing system re-
teins, do not easily crystallize. For these materials the quires not only that the active material be dried into the
formation of the amorphous (glassy) state is a necessity glassy state such that the amorphous product is physi-

257
Copyright 01997 by Marcel Dekker, Inc.
 258 Hatley

cally stable, but also that the active within it be chemi- below Tg there is a continuing loss of configurationa1
cally stable. This may appear a trivial statement but, as entropy. At some temperature, To, configuration entropy
will be shown, physical stability of the amorphous for- reaches zero. This temperature is the zero mobility tem-
mulation and chemical stability of an active within it are perature since, in the absence of entropy, molecular
not the same thing; the predominant influences may motion effectively stops.
differ. The principal factor responsible for stability is Glasses are classified as fragile or strong according
widely believed to be the glass temperature (T,) of the to their fragility parameter (D), which is derived from
dried material, with the chemical stability of an active the structural relaxation time (24). Strong glasses have
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

material being a direct function of the glass temperature a large D value, while fragile glasses have a small D
of the dried product (16). The Tg of a formulation is a parameter. In turn, To is related to Tgand D by:
function of the Tg values of the various components in
To = Tg/(1+(0.0255 x D))
the mix (17,181. The higher the T, value. the more
stable the product will be at a given storage temperature. Consequently a strong glass (large D) has a large Tg
If Tg is exceeded during storage, then the matrix losses - To and a fragile glass (small 0) has a small Tg - To
its rigidity and becomes a deformable “rubber” from The potential importance of this difference in terms of
which the product can physically collapse, individual excipient selection for the stabilization of labile materials
components crystallize, or the active degrade (19). Of- is addressed in this paper.
ten these changes damage the active irreversibly. In this study, two carbohydrate excipients are exam-
Although the relationship between the physical stabil- ined: sucrose, which is commonly used in a number of
ity of the amorphous formulation and T , is established, pharmaceutical formulations, and trehalose, which can
it has now been shown that T, is not the upper limit of have superior stabilizing properties (10). The objectives
chemical stability-some reactions still proceed below were to:
For personal use only.

this temperature (20). The amount of degradation var- Calculate To values of trehalose and sucrose by the
ies depending on the glass former (3). The rates of these method described by Pika1 from the width of the
reactions are orders of magnitude lower than those glass transition (25) and compare the values ob-
found either above the glass temperature or in solution, tained for sucrose with those in the literature ob-
( 2 1 ) but are very significant over the time scales em-
tained by enthalpy relaxation measurement (26) in
ployed during storage. Examples include the stability of order to assess the validity of Pikal’s method.
catalase. which cannot be related to the glass tempera- Compare the physicochemical properties of treha-
ture of the stabilizing excipient (22), and the poor stor- lose to those of sucrose in terms of their potential
age stability of hGH in a dextran formulation, even stabilizing effects on labile actives in the glassy
though it has a high glass temperature (20).
state.
The mechanisms of degradation below Tg have not
been fully elucidated, but they may result from side The reason for the stabilizing properties of trehalose
chain flexibility of a labile active in the glass or diffu- is a matter of debate (6) and this investigation provides
sion of small molecules, e.g.. water or oxygen, through the first assessment of the differences in the physico-
the glassy matrix causing hydrolysis or oxidation. Mo- chemical properties between the two excipients, which
lecular motion has been detected in glasses (23). An will assist in the resolution of this debate.
additional degradative mechanism that could occur is
interaction between the labile material and the glass-
forming excipient-for example, the glycation of lysine MATERIALS AND METHODS
vasopressin in the presence of reducing sugars (24).
Angel1 has recently proposed the concept of “strong” Carbohydrate glasses were formed by drying aque-
and “fragile” glasses (24). Strong glasses show a broad ous solutions. Samples were either freeze-dried or am-
glass transition by differential scanning calorimetry bient-temperature-dried under vacuum, as described in
(DSC). indicating a gradual loss of rigidity as configu- Ref. 6. (Primary drying was performed at -40°C for 80
rational entropy increases with increasing temperature. h and followed by a ramped 2.5”C/min secondary dry-
By contrast, fragile glasses show a sharp glass transi- ing to 25°C. All temperatures refer to the shelf tempera-
tion, indicating a sudden change from the glassy to a ture.) Samples were sealed under vacuum and stored at
more fluid rubbery state. As temperature is reduced 25°C until required.
 Glass Fragility and Stability 259

Thermal analysis was performed on a Perkin-Elmer as the deviation of an extrapolation of the baseline above
DSC-7. Samples of circa 10 mg were loaded in a dry and below the glass transition from the measured curve.
room into hermetically sealed aluminum sample pans. The midpoint was the midpoint of the glass transition as
Sample mass was determined gravimetrically on a determined by the computer software supplied with the
Mettler AE240 balance. Samples were loaded into the DSC.
DSC at 30"C, cooled to -60°C at lO"/min. Finally, the The moisture contents for the samples were 0.79%
samples were heated again at lO"/min. The power-time (trehalose) and 0.73 % (sucrose).
curve was recorded throughout. From the values in Table 1, the fragility constant D
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

Karl Fisher analysis was performed on a Metrohm and To were calculated according to the procedure de-
684KF coulometer. In a dry room (RH <20%)vials scribed in Refs. 27 and 28, using Moynihan's findings
were opened and 2 ml of formamide was added to a that the width of the glass transition is related to the
known mass of material. The moisture content of 0.25 effective activation energy (for viscous flow), and that
ml portions was determined. The moisture content was the activation energy for viscous flow is almost indis-
corrected against a formamide blank and related to the tinguishable from the activation energy for structural
initial sample mass on a wlw basis. relaxation, AH* (28), combined with the calculations of
Angel1 (27).
The accuracy of the calculation of D by this method
RESULTS is highly dependent on the accuracy of measurement of
ATg at the glass transition and relies on equations origi-
Representative DSC power-time curves for sucrose nally calculated from measurement on high Tg inorganic
and trehalose are shown in Figs. 1 and 2. The glass glasses. There may therefore be deviations in the abso-
transition temperatures are 101°C and 64°C for treha- lute values of To calculated. Pikal, however, finds this
For personal use only.

lose and sucrose respectively. In Fig. 3, a DSC power- to be less than 10°C when he compares the To value he
time curve for trehalose dried to <0.1% moisture is found for sucrose with that obtained via enthalpy relax-
shown, in which the heating has been performed over ation (25). In the study reported here the inter-sample
an extended temperature range. The glass temperature variance in the measurement of Tg was 0.15"C, giv-
is 116°C. Note that there is no evidence of crystalliza- ing a maximum variance in the width of the glass tran-
tion above Tg. In Table 1, the onset (T&, midpoint sition of f 0.3"C.
(Tg,.,), and endpoint (TgJ glass transition temperatures, To data were calculated according to the method de-
and ATg for amorphous sucrose and trehalose are pre- scribed in Ref. 25. The data obtained are contained in
sented. Onset and endpoint Tg values were determined Table 2. The low D value for both carbohydrates indi-

11.4 --

11.2 --
11 *-

10.2 I I I 1 I I I I1 I I I I ; I I I I i I I I I I I i t ; I I I I 1 1 I 1 1 / I f I I I I
~ M

Figure 1. A representative differential scanning calorimetic power-time curve for the heating of amorphous sucrose. Sample mass
4.300 mg. Heating rate lO"/min. Tgis 64°C at 0.73% moisture.
 260 Hatley

E
W
12.8 1 I

b-
3
0
d
c..
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

4
w

TEMPERATURE ("C)
Figure 2. .4representative differential scanning calorimetic power-time curve for the heating of amorphous trehalose. Sample
mass 7.800 mg. Heating rate IO"/min. T , is 101°C at 0.79% moisture.

cates they are both fragile (column 3), with trehalose water has a great influence; acting as a plasticizer and
being more fragile than sucrose. This difference in fra- depressing the Tgof the preparation (18). In this study,
gility, although slight if the intrinsic errors in measure- Tg of anhydrous trehalose glass (Fig. 3) is 116°C. This
For personal use only.

ment of the width of the glass transition are taken into is in agreement with recent literature (29), in which a
account, when combined with the intrinsically higher value of 120°C is quoted. The often reported value of
glass temperature (column 5) of trehalose, results in a 77°C (30) is probably a consequence of the trehalose
higher value of To (column 4) for trehalose. not being thoroughly dried-a similar transition tempera-.
ture is also reported by Green and Angel1 for trehalose
DISCUSSION glass derived from the melt of the dihydrate (31). We
found that 0.75 % moisture depressed the glass tempera-
Tg Values
ture by 20°C (c.f. Figs. 2 and 3) indicating that in very
The Tgof a formulation is a function of the Tg val- dry materials, Tg is extremely sensitive to moisture con-.
ues of the various components in the mix (17). Residual tent.

22 , I

TEMPERATURE ("C)
Figure 3. A representative differential scanning calorimetic power-time curve for the heating of anhydrous amorphous treha-
lose. Sample mass 16.7 mg. Heating rate lO"/min. Tg is 116°C at a moisture content of <0.1%.
 Glass Fragility and Stability 261

Table I
rhe Onset (Tg0), Midpoint (T8,,,), and Endpoint (T,) Glass Transition Temperatures
and AT8 for Amotphous Sucrose and Trehalose with Moisture Contents of 0.73
and 0.79 %, Respectively

Carbohydrate Tgo(K) Tgrn(K) Tgc(K) ATg (K)

Trehalose 370 374 377 7
Sucrose 333 337 340.5 7.5
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

Fragility and Crystallization tective effect of the amorphous matrix that the sucrose
glass provides will be permanently lost, even if-the
The more fragile the glass, the more sensitive it is to product is subsequently returned to lower temperatures.
changes around the glass transition. Above the glass It could be argued that this is not significant because a
transition, softening occurs much more rapidly for a sucrose-based product could be formulated to have a
given increase in temperature (23) than found with glass temperature much in excess of the temperatures it
stronger glasses. This would be expected to make the will experience during storage. Of course, this is pos-
glass more likely to devitrify just above Tg.The greater sible, but relies either on formulation with additional
fragility of trehalose compared to sucrose would suggest excipients with higher glass temperatures (e.g., human
that softening of the glass, as it exceeds Tg,would make serum albumin) or very thorough drying. It should be
For personal use only.

trehalose more prone to crystallization than sucrose. noted that a water content of 2-3% is typical of many
Surprisingly, this is not the case. Studies performed products after 1-2 months storage, as moisture ingress
under the same experimental conditions on sucrose (30) from the stopper into the product takes place (33).
show that sucrose crystallizes 30-40°C above the glass Moisture content of this magnitude reduces the glass
temperature. Trehalose does not crystallize under the temperature of sucrose to between 28 and 40°C (30)-
same conditions. Indeed, we have found that samples of temperatures that are commonly experienced during
trehalose glass can be held at 180"C, well in excess of storage. Because trehalose does not easily crystallize
Tg,for periods in excess of 48 hr with no crystallization. above Tg , trehalose-based formulations are much more
In the DSC, where heating rates of 5 or 10"C/min resistant to temperature abuse than might otherwise be
are used, crystallization of sucrose takes place 30-40°C expected.
above Tg . However, crystallization above Tg is a func-
tion of both T - Tg and time; thus, if a sucrose glass is To Values and Their Measurement
heated more slowly, then crystallization will occur much
closer to Tg . The implications for stability during stor- Tg has been said to be the safe storage temperature
age are significant. Many products are subject to tem- for an active. This has recently been shown not to be
perature abuse during transport and storage (32), and the case with many reported examples of degradational
under such conditions Tgmay be exceeded. Thus, if a processes occurring below T' (6, 24, 34). For this deg-
sucrose-based product is subjected to even 1 hr at a radation to occur there must be mobility or molecular
temperature 5°C above the glass transition, there is a freedom below Tg. By definition, To is the zero mobil-
high risk that crystallization will take place and the pro- ity temperature, thus this, not Tg, represents the safe

Table 2
AH*/RT,, D, To (K),and Tg (K)for Trehalose and Sucrose
(Trehalose Is the More Fragile and Has To at a High Temperature)
Carbohydrate AH*IRT, D To (K) TQ (K)
Trehalose 258.50 6.988 317 374
Sucrose 217.92 8.57 276.5 337
 262 Hatley

storage temperature for an active material in the amor- 7, and a To of 44°C. If we consider the case of a stron-
phous state if ull degradation is to be stopped. Conse- ger glass with a Tgof, say, 167"C, by conventional ex-
quently, below the glass transition, fragility is an advan- pectations this would be predicted to be a better stabi-
tage as the glass more quickly increases in viscosity for lizer, since the Tg value is higher. Nevertheless, if ATg
a given temperature drop compared to a stronger glass. is set at 20°C (which is not an unrealistic value), then
Thus for a fragile glass, the difference Tg - To is much calculation yields a D value of 22.3 and a To of 7°C.
less and T, occurs at a higher temperature than with a The stronger glass has a higher Tg but To, the zero
strong glass. It should be remembered, however, that mobility temperature at which reactions cease, is 37°C
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

not all degradative processes proceed between To and lower than trehalose. The material with the high Tgmay
Tg. Translational movement of proteins is an example not, therefore, be the better stabilizer.
that correlates highly with Tg and does not occur be- In summary, the stability of an active in an amor-
tween T,, and Tg (25). There are insufficient data avail- phous matrix may be governed by To, the zero mobil-
able at present to determine where between To and Tg ity temperature, not Tg. To ensure that degradation does
different reaction processes cease. It could be speculated not take place it would be necessary to make excipient
that the smaller the reacting species, the closer to To the selection on the basis of To, not Tg. If stability is gov-
system needs to be stored to prevent degradation. erned by To then, to work most effectively in stabiliz-
This study has shown that the To value for sucrose ing an active, a glass-forming excipient should have
is 3.5"C. The Tg - T,, difference is 61.5"C; this is in both a high glass temperature and be fragile, since these
good agreement with the value of 60°C found by two properties combine to produce excipients with high
Hancock et al. (26). Because the calculations of To values. Trehalose has been shown to have both a
Hancock et al. were based on a more direct approach high Tg value and fragility, resulting in a To in excess
using enthalpy relaxation, this agreement helps validate of 40°C. It therefore appears that trehalose fits these
For personal use only.

the more indirect but simpler approach of using the requirements more effectively than sucrose, whose low
width o f the glass transition. The possible errors quoted Tg results in a To of 3.5"C.
earlier may therefore be extreme. (It is also interesting Calculation of To by the method proposed by Pikal
to note that the sucrose sample used in this study con- from the width of the glass transition agrees well with
tained a small amount of residual moisture, whereas the values calculated from enthalpy relaxation. Although
sample used by Hancock et al. was dry. It would appear further comparative data on other excipients are re-
then, that low moisture content does not affect fragility, quired, this study suggests that Pikal's procedure may
since To - Tg remains constant.) To ensure the stability provide a rapid and effective method for the routine
of a product stabilized with sucrose, it is advisable to screening of formulations with respect to stability pre-
store the product below its To temperature of 3.5"C. diction.
The To value of trehalose at 44°C is sufficiently high
that room temperature storage, with guaranteed stabil-
ACKNOWLEDGMENTS
ity, is possible. It is interesting to note that Pikal rates
the stabilizing properties of trehalose as greater than
The author would like to thank Dr. Mike Pikal of Eli
dextran (34). Dextran, although not analyzed here in
Lilly for making available details of the calculation of
detail, has a broad glass transition indicating a strong
To from the width of the glass transition. Drs. Trevor
glass. Thus the ranking of Pikal appears to correspond
Gard and Julian Blair are thanked for their constructive
to fragility-the more fragile the glass, the better it is
comments during the preparation of the manuscript.
as a stabilizing agent.

To and Tg as Predictors of Stability

REFERENCES
To form an amorphous matrix with good stabilizing J. I. Wells, Pharmaceutical Preformulation: The Physi-
properties, a material should have a high Tgbut also be cochernical Properties of Drug Substances, Ellis
a fragile glass. High Tg without fragility may be unsat- Horwood Limited, Chichester, 1988.
isfactory. as is fragility without a high Tg. This is best M. J. Pikal, Freeze-drying of proteins. I. Process de-.
demonstrated by way of an example. sign, Biopharm., 3, 18-27 (1990).
Trehalose with a glass temperature of 101°C has M. J. Pikal, Freeze-drying of proteins. 11. Formulation
been shown here to have a AT, of 7"C, a D value of selection, Biopharm., 3, 26-30 (1990).
 Glass Fragility and Stability 263

4. M. J. Pikal, Freeze-drying of proteins: process, formu- 18. B. C. Hancock and G. Zografi, The relationship be-
lation and stability, in Formulations and Delivery of tween the glass transition temperature and the water
Proteins and Peptides, (J. L. Cleland and R. Langer, content of amorphous pharmaceutical solids, Pharm.
eds.), American Chemical Society Symposia Series Res., 1 1 , 471-477 (1994).
N0.567, 1994, pp. 120-133. 19. M. P. W. M. te Booy, R. A. de Ruiter, and A. J. de
5. F. Franks, Freeze-drying: from empiricism to predict- Meere, Evaluation of the physical stability of freeze-
ability, Cryoletters, 11, 93-1 10 (1990). dried sucrose-containing formulations by differential
6. R. H. M. Hatley and C. A. L. S . Colaco, Stabilisation scanning calorimetry, Pharm. Res., 9, 109-1 14 (1992).
of biological molecules, cells and organisms, in Protein 20. M. J. Pikal, K. M. Dellerman, M. L. Roy, and R. M
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

Labelling: Bioconjugation Techniquesfor the Biomedi- Riggin, The effects of formulation variables on the sta-
cal Sciences (M. Aslam, ed.), Macmillan Press, London. bility of freeze-dried human growth hormone, Pharm.
(in press). Res., 8, 427-436 (1991).
7. E. M. Gribbon, R. H. M. Hatley, T. Gard, J. Blair, J. 21. W. L. Kerr, M. H. Lim, D. S. Reid, and H. Chen,
Kampinga, and B. J. Roser, Trehalose and novel hydro- Chemical reaction kinetics in relation to glass tempera-
phobic sugar glasses in drug stabilization and delivery, tures in frozen food polymer solutions, J. Sci. Food
in Chemical Aspects of Drug Delivery Systems (D. R. Agric., 61, 51-56 (1993).
Karsa and R. A. Stephenson, eds.), RSC, Cambridge, 22. J. F. Carpenter, S. J. Perstrelski, T. J. Anchordoguy,
1996, pp. 138-145. and T. Arkawa, Interactions of stabilizers with proteins
8. F. Franks and R. H. M Hatley, Stable enzymes by water during freezing and drying, in Formulations and Deliv-
removal, in Stability and Stabilization of Enzymes (W. ery of Proteins and Peptides (J. L. Cleland and R.
J. J. van den Tweel, A. Harder, and R. M. Buitelaar, Langer, eds), American Chemical Society Symposia
eds.), Elsevier, Amsterdam, 1993, pp. 45-54. 567, 1994, pp. 134-147.
9. F. Franks, R. H. M. Hatley, and S. F. Mathias, Mate- 23. C. A. Angell, Formation of glasses from liquids and
rials science and the production of shelf stable biopolymers, Science, 267, 1924-1935 (1995).
For personal use only.

biologicals, Biopharm., 14, 38-5 (1991). 24. E. Tarelli, P. H. Corran, B. R. Bingham, H. Mollison,
10. D. Blakeley, B. Tolliday, C. A. L. S. Colaco, and B. and R. Watt, Lysine vasopressin undergoes rapid
Roser, Dry, instant blood typing plate for bedside use, glycation in the presence of reducing sugars, J. Pharma-
Lancet, 336, 854-855 (1990). ceutical and Biomedical Analysis, 12, 1355-1361 (1994)
11. C. A. L. S. Colaco, S. Sen, M. Thangavelu, S Pinder, 25. M. J. Pikal, D. R. Rigsbee, and M. L. Roy. The rela-
and B. Roser, Extraordinary stability of enzymes dried tionship between glass transition temperature and stabil-
in trehalose: simplified molecular biology. Biotech, 10, ity of freeze-dried human growth hormone, Symposia
1007-1011 (1992). Abstract for The 10th AAPS Annual Meeting, Nov. 5-
12. C. A. L. S . Colaco, C. J. S. Smith, S. Sen, D. H. 9. 1995, Miami Beach Florida.
Roser, D. C. Chilvers, and B.J. Roser, The chemistry 26. B. C. Hancock, S. L. Shamblin, and G. Zografi, Mo-
of rational protein drug stabilization, in Formulations lecular mobility of amorphous pharmaceutical solids
and Delivery of Proteins and Peptides (J. L. Cleland and below their glass transition temperatures, Pharm. Res.,
R. Langer, eds.), American Chemical Society Sympo- 12, 799-806 (1995).
sia Series, 567, 1994 pp: 222-240. 27. C. A. Angell, Relaxation in liquids, polymers and plastic
13. B. J. Roser, Trehalose drying: a novel replacement for crystals-patterns and problems, J. Non-Cryst. Solids,
freeze-drying, Biopharm. 5, 44-53 (1991). 131, 13-31 (1991).
14. R. H. M. Hatley, F. Franks, S. Brown, G. Sandhu, and 28. C. T. Moynihan, Correlation between the width of the
M. Gray, Stabilisation of a pharmaceutical drug sub- glass transition region and the temperature dependence
stance by freeze-drying: a case study, Drug Stability 1, of the viscosity of high-T, glasses, J. Am. Ceram. SOC.,
2-12 (1996). 76, 1081-1087 (1993).
15. R. H. M. Hatley, The effective use of differential scan- 29. Y. Roos, Melting and glass transitions of low molecu-
ning calorimetry in the optimisation of freeze-drying lar weight carbohydrates, Carbohydrate Res., 238, 39-
processes and formulations, J. Biol. Standardization, 74, 48 (1993).
105-122 (1992). 30. F. Franks, B. Aldous, and T. Auffret, Using water
16. F. Franks. Freeze-drying: from empiricism to soluble glasses as a strategy for bioproduct stabilization,
predicability . The significance of glass transitions, J. GEN, August, 20 (1995).
Biol. Standardization, 74, 9-19 (1992). 31. J. L. Green and C. A. Angell, Phase relations and vit-
17. R. H. M. Hatley and F. Franks. Applications of DSC rification in saccharide water solutions and the trehalose
in the development of improved freeze-drying processes anomaly, J. Phys. Chem., 93, 2880-2882 (1989).
for labile biologicals, J. Thermal Analysis, 37, 1905- 32. R. Gray, The bug in the box, Pharm. Tech. Europe, 6,
1914 (1991). 20-32 (1994).
 264 Hatley

33. M. .I.Pikal, M. L. Roy, and S. Shar, Mass and heat 34. M. J. Pikal, Freeze drying of proteins,. in Pepride and
transfer in vial freeze-drying of pharmaceuticals: role of Prorein Delivery, (V. H. L Lee, ed.) 2nd Edition,
the vial, J . Pharm. Sci., 73, 1224-1237 (1984). Marcel Dekker, New York, 1996.
Pharmaceutical Development and Technology Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
For personal use only.