Thermal Analysis of Frozen Solutions: Multiple Glass
Transitions in Amorphous Systems
GREGORY A. SACHA, STEVEN L. NAIL
Baxter BioPharma Solutions, Bloomington, Indiana 47403
Received 17 November 2008; revised 27 January 2009; accepted 2 February 2009
Published online 21 April 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21737
ABSTRACT: Frozen aqueous solutions of sucrose exhibit two   glass transition-like 
thermal events below the melting endotherm of ice when examined by DSC, but the
physical basis of these events has been a source of some disagreement. In this study, a
series of sugars, including sucrose, lactose, trehalose, maltose, fructose, galactose,
fucose, mannose, and glucose were studied by modulated DSC and freeze-dry microscopy
in order to better understand whether sucrose is unique in any way with respect to this
behavior, as well as to explore the physical basis, and the pharmaceutical significance of
these multiple transitions. Double transitions were found to be a common feature of all
sugars examined. The results are consistent with both thermal events being glass
transitions in that (1) both events have second-order characteristics that appear in
the reversing signals, (2) annealing experiments reveal that enthalpy recovery is
associated with each transition, and (3) Lissajous plots indicate that no detectable
latent heat of melting is associated with either transition. The data in this study are
consistent with the idea that the lower temperature transition arises from a metastable
glassy mixture containing more water than that in the maximally freeze-concentrated
solute. Freeze-dry microscopy observations show that for all of the sugars examined, it is
the higher temperature transition that is associated with structural collapse during
freeze-drying. There is no apparent pharmaceutical significance associated with
the lower-temperature transition. ß 2009 Wiley-Liss, Inc. and the American Pharmacists
Association J Pharm Sci 98:3397 3405, 2009
Keywords: freeze-drying; microscopy; collapse; carbohydrates
0
INTRODUCTION a single Tg when examined using differential
scanning calorimetry (DSC). Sucrose is an exci-
Studying the thermal behavior of frozen solutions pient used in parenteral pharmaceutical formula-
provides valuable information for the develop- tions that presents an exception, displaying two
ment of lyophilized formulations. Predominantly glass transition-like thermal events below the ice
amorphous formulations commonly exhibit a melting endotherm.1 4 The midpoint of one event
glass transition of the maximally freeze-concen- occurs at approximately 488C (T1) and the
0 0
trated solute(s) (Tg). The Tg designates the midpoint of the other event occurs at approxi-
product temperature at or above which viscous mately 358C (T2). The physical significance of
flow of the freeze-concentrated mixture will occur these transitions has been a matter of some
during primary drying. Most formulations exhibit controversy in the literature. T1 is often consid-
0
ered the true Tg and T2 has been referred to as
the incipient melting of ice or as a second glass
Correspondence to: Gregory A. Sacha (Telephone: 812-355-
transition.1 5 A better understanding of these
2046; Fax: [phone]; E-mail: [e-mail])
events is needed to determine their significance
Journal of Pharmaceutical Sciences, Vol. 98, 3397 3405 (2009)
ß 2009 Wiley-Liss, Inc. and the American Pharmacists Association with respect to quality attributes of lyophilized
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009 3397
3398 SACHA AND NAIL
formulations. The purpose of this study was to The collapse temperatures for solutions of
examine the effects of annealing and concentra- sucrose, lactose, and maltose were measured
tion on the thermal events occurring at T1 and T2 using freeze-dry microscopy. Five microliters of
in frozen mono- and disaccharide solutions and to solution was placed between two glass cover slips,
determine the relationship between the tempera- then placed on a freeze-dry microscope stage
tures of these events and collapse as observed by (Linkam Scientific Instruments, UK) attached to
freeze-dry microscopy. a polarized light microscope (Olympus BX51,
McCrone Microscopes and Accessories, West-
mont, IL) and a digital camera (PAXcam 2þ,
Villa Park, IL). Samples were frozen to 408C
MATERIALS AND METHODS using a temperature controller. A vacuum of
100 mTorr was pulled and the samples were
Carbohydrate solutions were prepared gravime- allowed to freeze-dry until a sublimation front and
trically by weighing the appropriate amount of dried material were visible. The vacuum set-point
solid material and adding the appropriate weight was maintained while slowly increasing the
of purified water. The solutions were mixed temperature of the samples in increments of 2
with gentle heating until there was no visual 38C/min and holding the samples at the respective
evidence of solid material in the solution. Sucrose, temperatures to evaluate the appearance of
D-fructose, D-maltose monohydrate, D-galactose, the partially dried solids after movement of the
and L( )-fucose were obtained from Sigma sublimation front. The experiments were termi-
Aldrich (St. Louis, MO). D-Trehalose dihydrate nated after the appearance of collapse.
and D-mannose were obtained from Fluka (Buchs
SG, Switzerland). Lactose monohydrate and
D(þ)-glucose monohydrate were obtained from
EMD (Gibbstown, NJ). Anhydrous dextrose was RESULTS
obtained from J T Baker (Phillipsburg, NJ). All
materials were obtained as research grade. Modulated DSC experiments demonstrated that
Thermal characteristics of the solutions were two   glass transition-like  thermal events were
measured using a DSC (Model Q2000, TA Instru- common features of all sugars examined in this
ments, New Castle, DE) in the modulated mode. study (Figs. 1 and 2). The same results were
The DSC was calibrated using indium and observed for samples that were quickly cooled to
mercury standards. The solutions were added to 808C and for samples that were cooled to 808C
aluminum, Tzero pans, and hermetically sealed at 0.58C/min (Data not shown). All further
with a Tzero press and Tzero lids. Most experi- experiments used a rapid cooling cycle to reduce
ments were conducted by cooling the solutions to the time for each experiment.
808C as quickly as possible and holding the
sample isothermally for 2 min before modulating
the temperature at 18C every 120 s. The samples
were held isothermally for 5 min after starting the
temperature modulation and then warmed using
a ramp rate of 0.58C/min. Samples were exposed to
annealing steps at temperatures above and below
their glass transition temperatures for 10 h. The
data were evaluated using TA Universal Analysis
software (TA Instruments, New Castle, DE).
Stepwise quasi-isothermal MDSC cycles were
conducted using 20% sucrose solutions. Modu-
lated heat flow data were collected while holding
the sample isothermally for 30 min at the desired
temperature. The sample was equilibrated at the
desired temperature for 10 min prior to collecting
the 30 min isothermal data. The cycles were
Figure 1. Reversing heat flow curves for 20% solu-
conducted for the temperature range of 55 to 08C
tions of fructose (A), galactose (B), sucrose (C), maltose
in increments of one degree. (D), and trehalose (E).
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009 DOI 10.1002/jps
THERMAL ANALYSIS OF FROZEN SOLUTIONS 3399
Table 1. Values for T1 and T2 for Sucrose, Trehalose,
Lactose, and Maltose Measured Using MDSC
Concentration (%) T1 (8C) T2 (8C)
Sucrose
2 49 36
4 45 34
5 48 36
6 46 35
8 46 35
10 49 35
20 49 35
30 48 35
40 47 35
50 50 39
55 55 39
Figure 2. Reversing heat flow curves for 20% solu-
58 62 41
tions of fucose (A), glucose (B), mannose (C), and lactose
59 ND 43
(D).
60 ND 43
80 ND 42
Trehalose
The lower temperature event (T1) and higher
2ND 33
temperature event (T2) can be difficult to
5 41 32
detect when studying solutions at concentrations
10 43 32
typically used in lyophilized formulations, which
20 42 32
are typically 2 5%. The intensity of the thermal
Lactose
event occurring at T1 is usually too weak to be
2ND 31
observed in 2% solutions (Tab. 1). Increasing the
5 43 31
concentration of a solution increases the intensity
10 42 30
of the thermal events, but can also lead to changes 20 43 30
Maltose
in the appearance and position of the events as the
2ND 33
viscosity of the solution increases. Experiments
5 40 33
were conducted using sucrose solutions to deter-
10 42 32
mine the maximum concentration that could be
20 42 32
used without observing significant changes in the
midpoints of the thermal events. The concentra-
tion of sucrose solutions can be as high as 40%
without affecting the midpoints for T1 or T2 explained that excess enthalpy decreases during
(Fig. 3). Concentrations greater than 40% shifted annealing and the enthalpy recovered during
the midpoint of the events to lower temperatures. heating increases with increasing annealing time.
No events for T1 were detected for sucrose This leads to correspondingly larger endothermic
concentrations of 59 80% within the temperature peaks. Roos also noted that several days or weeks
range utilized for the experiments. The midpoints of annealing may be required for solutions with
of the events occurring at T1 and T2 remained high solute concentrations to obtain maximum ice
consistent for concentrations between 2% and 20% formation.7 Therefore, 10 h of annealing time was
for all sugars examined in this study. The used at higher temperatures to maximize ice
midpoints of the events for a few of the sugars formation and to maintain consistency between
are provided in Table 1. The sugar solutions in experiments. The low temperature annealing
this study were prepared as 20% (w/w) solutions to conditions were chosen to ensure that the
obtain the optimal intensities for the thermal annealing temperature was below the tempera-
events. ture of the onset of T1 for each of the solutions.
Annealing experiments were conducted above Likewise, the high temperature annealing condi-
and below T1 for each of the sugar solutions for up tions were chosen to ensure that the annealing
to 10 h at the specific annealing temperature. The temperature was above the upper temperature
long annealing time was chosen to attempt to range for T1. Figures 4 6 display the thermo-
maximize enthalpic recovery. Tant and Wilkes6 grams for 20% solutions of sucrose, trehalose, and
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009
3400 SACHA AND NAIL
Figure 5. Nonreversing heat capacity curves for 20%
Figure 3. Reversing heat flow curves for sucrose
solutions of trehalose with no annealing (A) and after
solutions with concentrations of 20% (A), 40% (B),
annealing at 358C (B) and 528C (C) for 10 h.
55% (C), 59% (D), and 80% (E).
fructose without annealing and after annealing latent heat flow. The release of latent heat as a
above and below T1. The nonreversing heat result of melting will disfigure the ellipse into a
capacity thermograms for sucrose 20% (w/w) circle or a figure eight. Ellipses were obtained at
include annealing above the temperature range 48 and 358C (Fig. 7), indicating no measurable
for T2. flow of latent heat. Some distortion of the ellipse
Stepwise quasi-isothermal MDSC (SQI-MDSC) appeared at 108C with the majority occurring at
experiments were used to test for the flow of latent 6to 38C (Fig. 8). The ellipse returned at 28C.
heat in solutions of sucrose 20% (w/w) in the Freeze-dry microscopy experiments are com-
temperature range of 55 to 08C. Lissajous plots monly used in combination with DSC experiments
were created using the data obtained isothermally to thermally characterize solutions intended for
with modulation and plotting the data using lyophilization. DSC experiments provide data on
modulated heat flow as a function of modulated the physical nature of the thermal events and the
heating rate. An ellipse is formed when there is no temperatures at which they occur. Freeze-dry
Figure 4. Nonreversing heat capacity curves for
sucrose 20% solutions without annealing (A), with
annealing at 258C for 10 h (B), with annealing at Figure 6. Nonreversing heat capacity curves for 20%
398C for 10 h (C), and with annealing at 608C for solutions of fructose with no annealing (A) and after
10 h (D). annealing at 508C (B) and 658C (C) for 10 h.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009 DOI 10.1002/jps
THERMAL ANALYSIS OF FROZEN SOLUTIONS 3401
Figure 7. Lissajous plots for sucrose 20% at 488C
(black) and at 358C (green). The ellipses indicate no
flow of latent heat.
microscopy is used to visually identify the onset of
viscous flow to provide information on the
practical significance of the thermal event(s).
Freeze-dry microscopy studies were conducted
to identify the onset of collapse for 2%, 5%, and
10% (w/w) solutions of sucrose, lactose, trehalose,
and maltose. The onset of collapse matched well
with the temperature of the midpoint for T2
determined by DSC for the carbohydrates tested.
Examples are provided for sucrose and lactose in
Figures 9 and 10.
Figure 9. Freeze-dry photomicrographs of sucrose
2% (A) and sucrose 10% (B) solutions showing the
onset of collapse of the dried layer at 35 and 348C,
respectively.
DISCUSSION
mussen in 1968 and continue to be a topic of some
Multiple thermal events in frozen solutions of
controversy.1,2 The subject is important to phar-
sucrose were first observed by Luyet and Ras-
maceutical and food science because it affects
processing and storage conditions for many
products. Studies from both disciplines have
focused on demonstrating the presence of both
thermal events and attempting to understand
their physical-chemical basis.3 5,8,9 MDSC experi-
ments conducted in the study reported here
demonstrate that multiple thermal events are
associated with all of the mono- and disaccharides
examined (Figs. 1 and 2). Few references exist
on the occurrence of multiple transitions in
other mono- and disaccharides or on the practical
significance of the transitions in lyophilized
formulations. Blond and Simatos observed two
thermal events in 55% galactose solutions, but
hypothesized that they were either two distinct
glass transitions or a single glass transition
Figure 8. Lissajous plots for sucrose 20% at 108C
associated with an enthalpy relaxation.10 Car-
(black) and at 48C (green). Distortion of the ellipse in
Green indicates the flow of latent heat (melting). rington and co-workers noted two thermal events
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009
3402 SACHA AND NAIL
20% (w/w) (Tab. 1). The midpoints for T1 and T2
begin to decrease for sucrose at concentrations
greater than 40% (w/w) (Tab. 1 and Fig. 3). This
suggests that water is becoming trapped in the
amorphous material as the viscosity increases and
the plasticizing affects of the water are leading to
decreases in the values for T1 and T2. In all cases,
the intensity of the event at T2 is greater than the
intensity of the event at T1 and the event at T1
becomes difficult to detect at pharmaceutically
relevant concentrations, which are typically 2
10%.
Most studies agree that the event occurring at
T1 is a glass transition, but differ in opinion with
regard to the nature of the event(s) occurring at
T2.1,3,4,15 Rasmussen and Luyet suggested that
TAM was related to a change in the interface
between the ice crystals and the solution. Ablett
et al. proposed that T2 is caused by the dissolution
of ice.16 Levine and Slade proposed that T2 is the
glass transition of the maximally freeze concen-
0
trated solute, Tg. Chang et al.5 examined 10%
sucrose solutions using MDSC and plotted the
data using reversing and nonreversing heat flow
as functions of temperature. They observed
enthalpy recovery and merging of T1 with T2
after annealing, although the detection of the
events was limited by the concentration of the
solution. Aubuchon et al.9 investigated frozen 40%
Figure 10. Freeze-dry photomicrographs of lactose
sucrose solutions using MDSC, where the focus of
2% (A) and lactose 10% (B) solutions showing the onset
the investigation was on the technique itself. Both
of collapse of the dried layer at 298C.
low-temperature events were shown to be con-
in fructose solutions, but referred to the event at sistent with second order transitions. Similarly,
T2 as the onset of the melting of ice.11 Arvani- Inoue and Suzuki studied enthalpy relaxation in
toyannis et al.12 observed two thermal events in frozen sucrose solutions at concentrations of 40%
frozen mixtures of glucose and fructose below the and 80% using conventional DSC.17 Annealing
ice melting endotherm, but did not discuss their temperatures in the range of 70 to 558C were
physical meaning. Studies conducted on solutions used for up to 6 days. While these investigators
of trehalose reported only a single glass transition show a thermogram of frozen 40% sucrose that is
within a temperature range of 15 to 808C.13,14 similar to the results reported here in that two
Luyet and Rasmussen defined the event occurring enthalpy recovery endotherms appear to be
at T1 as Tg for sucrose solutions and defined two present, they interpret the higher-temperature
events occurring on either side of the midpoint for transition as the onset of ice melting.
T2 as ante-melting (TAM) and incipient (TIM) Interpretation of the results can be complicated
melting of ice.1,2 They observed that the tempera- by the poor resolution of the events when plotted
tures for TAM and TIM were independent of the using heat flow signals. The events are easily
starting concentration for the solutions and observed when the data are plotted using non-
noticed that the temperature for Tg decreased reversing heat capacity as a function of tempera-
as the concentrations for the sucrose solutions ture. The data were examined using nonreversing
increased from 50% to 70%.2 The results from our heat capacity because it clearly displayed
study agree with this observation and demon- the decrease in heat capacity associated with
strate that the temperatures for the midpoints of the kinetically dependent event of the formation of
T1 and T2 remain consistent for the carbohydrates ice while also displaying the increases in heat
examined in this study for concentrations up to capacity associated with molecular mobility.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009 DOI 10.1002/jps
THERMAL ANALYSIS OF FROZEN SOLUTIONS 3403
Experiments conducted using 20% sucrose, tre- when sucrose solutions were annealed at tem-
halose, and fructose solutions were plotted using peratures between T1 and the onset of the melting
nonreversing heat capacity as a function of of ice defined as occurring at 348C. Chang and
temperature and revealed two, easily identifiable co-workers made similar observations when
events that correspond to the two events observed annealing 10% solutions of sucrose and attributed
in the reversing heat flow curves (Figs. 4 6). The T1 to a metastable glassy mixture associated with
thermograms display two areas with increasing the glass transition.5 The data in this study agree
heat capacity that abruptly decrease as the with their conclusions except when annealing a
temperature increases. Increasing heat capacity 20% sucrose solution at 258C, which is above the
is indicative of molecular mobility that may temperature range for T2. The expectation was
correlate with the release of unfrozen water from that the transition at T1 would completely merge
the freeze-concentrated solute. A typical melting with the transition at T2 as a result of complete
event would display an increase in heat capacity devitrification of the metastable glass. However,
that would remain elevated as the temperature the midpoint for T1 remained at the same
increased. The decreases in heat capacity temperature as the midpoint for an unannealed
observed in the nonreversing heat capacity solution (Fig. 4). This suggests that viscous flow of
thermograms is consistent with crystallization the maximally freeze-concentrated solute may
of unfrozen water. Aubuchon et al.9 also reported have resulted in blending of the solute with the
a decrease in heat capacity when experimenting metastable glass. Cooling and re-heating the
with sucrose solutions. Complex heat capacity solution would form the same thermogram as
plots were used to compare slowly cooled (18C/ an unannealed solution.
min) and quench cooled samples. Quench cooled Aubuchon et al. utilized stepwise quasi-isother-
samples were characterized by a single low mal MDSC (SQI-MDSC) to demonstrate the
temperature transition, whereas a two-step tran- melting point of indium using Lissajous plots.9
sition was evident for a slowly cooled sample. They mentioned that the method was also used for
Stepwise isothermal heat capacity measurement sucrose solutions and that the Lissajous figures
for the two cooling rates showed that the heat showed no indication of melting at the higher
capacity was relatively stable throughout the temperature transition. However, they did not
temperature range of the two transitions, whereas present any data or report the temperature at
the heat capacity decreased significantly during which the onset of melting was detected. Lissajous
the time interval of each isothermal temperature figures are created using data obtained isother-
for the quench cooled sample, presumably due to mally with modulation and plotting the data using
crystallization of previously unfrozen water. modulated heat flow as a function of modulated
Annealing studies were conducted at tempera- heating rate. An ellipse is formed when there is no
tures above and below the temperature range for latent heat flow. The release of latent heat as a
T1. Enthalpy recovery was observed for the event result of melting will disfigure the ellipse into a
at T1 after annealing the carbohydrate solutions circle or a figure eight. This study utilized SQI-
at temperatures below T1 for up to 10 h. The MDSC to examine sucrose 20% solution in the
magnitude of T2 appeared similar before and after temperature range of 55 to 08C. Ellipses were
annealing at temperatures below T1. The lack of observed at 48 and 358C (Fig. 7), indicating no
change in the magnitude for T2 may be the result flow of latent heat. Some distortion of the ellipse
of annealing at a very low temperature. Chang appeared at 108C with the majority occurring at
and Baust18 observed enthalpy relaxation for 6 to 38C (Fig. 8) and the ellipse returned at
glycerol solutions when annealed at temperatures 28C. The lack of latent heat flow at or near T2
below the glass transition and noted that the (approximately 358C) indicates that the event is
magnitude of the enthalpy recovery endotherm at a glass transition. Melting is not observed until
the glass transition decreased with lower anneal- the temperature of the sample reaches at least
ing temperatures. 68C.
The midpoint for T1 increased and began to Interpretation of calorimetric data differs
merge with T2 when the solutions were annealed widely between researchers, and, thus far, no
0
at temperatures between the two events (Figs. 4 visual data demonstrating T1 or T2 as Tg have
6). This is likely the result of the devitrification of been reported. Freeze-dry microscopy is com-
the glassy mixture. Roos and Karel19 observed monly used to identify the onset of collapse
that maximum ice formation would only occur in pharmaceutical solutions during the develop-
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009
3404 SACHA AND NAIL
ment of lyophilization cycles. The onset of in frozen solutions of glycerol, ethylene glycol,
glucose, and sucrose. Biodynamica 10:319 331.
collapse occurs when the temperature of the
2. Luyet B, Rasmussen D. 1968. Study by differential
sample reaches or exceeds the temperature for
0
thermal analysis of the temperatures of instability
the maximally freeze-concentrated solute, Tg.
of rapidly cooled solutions of glycerol, ethylene
Freeze-dry microscopy studies confirmed that
glycol, sucrose and glucose. Biodynamica 10:167
the onset of collapse for 2%, 5%, and 10% (w/w)
191.
solutions of sucrose, lactose, trehalose, and mal-
3. Sahagian ME, Goff HD. 1995. Thermal, mechanical
tose corresponds closely to T2 (Figs. 9 and 10). This
and molecular relaxation properties of stabilized
supports the conclusion that the event at T2 is
sucrose solutions at sub-zero temperatures. Food
a glass transition and that it represents the
Res Int 28:1 8.
maximally freeze-concentrated solute. There is
4. Sahagian ME, Goff HD. 1994. Effect of freezing rate
widespread agreement that the event at T1 is a on the thermal, mechanical and physical aging
properties of the glassy state in frozen sucrose
glass transition. The data reported here is
solutions. Thermochim Acta 246:271 283.
consistent with the idea that T1 arises from a
5. Chang L, Tang X, Pikal MJ, Milton N, Thomas L.
metastable glassy mixture with a higher water
1999. The origin of multiple glass transitions in
content than the glass represented by T2. How-
frozen aqueous solutions. Proc NATAS Annual Conf
ever, while T2 corresponds closely to the onset of
Therm Anal Appl 27:624 628.
collapse in freeze-drying, there is no apparent
6. Tant MR, Wilkes GL. 1981. An overview of the
pharmaceutical significance associated with the
nonequilibrium behavior of polymer glasses. Polym
event at T1.
Eng Sci 21:874 895.
7. Roos Y. 1995. Characterization of food poly-
CONCLUSIONS
mers using state diagrams. J Food Eng 24:339
360.
Frozen, binary solutions of sucrose in water
8. Izzard MJ, Ablett S, Lillford PJ, Hill VL, Groves IF.
exhibit two thermal events below the ice melting
1996. A modulated differential scanning calori-
endotherm. One event occurs at a low tempera-
metric study: Glass transitions occurring in sucrose
ture designated as T1 and the other occurs at a solutions. J Therm Anal 47:1407 1418.
9. Aubuchon S, Thomas L, Theuerl W, Renner H.
higher temperature, T2. The two thermal events
1998. Investigations of the sub-ambient transitions
were found to be common features among all
in frozen sucrose by modulated differential scan-
mono- and disaccharide solutions studied. Modu-
ning calorimetry (MDSC). J Therm Anal 52:53
lated DSC data support that both events are glass
64.
transitions in that (1) both events have second-
10. Blond G, Simatos D. 1991. Glass transition of the
order characteristics that appear in the reversing
amorphous phase in frozen aqueous systems. Ther-
signals, (2) annealing experiments reveal that
mochim Acta 175:239 247.
enthalpy recovery is associated with each transi-
11. Carrington AK, Goff HD, Stanley DW. 1996. Struc-
tion, and (3) Lissajous plots indicate that no
ture and stability of the glassy state in rapidly and
detectable latent heat of melting is associated
slowly cooled carbohydrate solutions. Food Res Int
with either transition. The data in this study are 29:207 213.
consistent with the idea that the lower tempera- 12. Arvanitoyannis I, Blanshard JMV, Ablett S,
Izzard MJ, Lillford PJ. 1993. Calorimetric study
ture transition arises from a metastable glassy
of the glass transition occurring in aqueous gluco-
mixture containing more water than that in the
se:fructose solutions. J Sci Food Agric 63:177
maximally freeze-concentrated solute. Freeze-dry
188.
microscopy observations show that for all of the
13. Chen T, Fowler A, Toner M. 2000. Literature
sugars examined, it is the higher temperature
review: Supplemented phase diagram of the
transition that is associated with structural
trehalose-water binary mixture. Crybiology 40:277
collapse during freeze-drying. This indicates that
282.
the higher temperature event should be consid-
14. Miller DP, de Pablo JJ, Corti H. 1997. Thermophy-
ered the glass transition of the maximally freeze-
sical properties of trehalose and its concen-
0
concentrated solute, Tg. trated aqueous solutions. Pharm Res 14:578
590.
REFERENCES
15. Levine H, Slade L. 1986. A polymer physico-chemi-
cal approach to the study of commercial starch
1. Rasmussen D, Luyet B. 1969. Complementary hydrolysis products (SHPs). Carbohydrate Polym
study of some non-equilibrium phase transitions 6:213 244.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009 DOI 10.1002/jps
THERMAL ANALYSIS OF FROZEN SOLUTIONS 3405
16. Ablett S, Izzard MJ, Lillford PJ, 1992. Differential 18. Chang ZH, Baust JG. 1991. Physical aging of glassy
scanning calorimetric study of frozen sucrose and state: DSC study of vitrified glycerol systems. Cryo-
glycerol solutions. J Chem Soc Faraday Trans biology 28:87 95.
88:789 794. 19. Roos Y, Karel M. 1991. Amorphous state and
17. Inoue C, Suzuki T. 2006. Enthalpy relaxation of delayed ice formation in sucrose solutions. Int
freeze concentrated sucrose-water glass. Crybiol- J Food Sci Technol 26:553 566.
ogy 52:83 89.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 9, SEPTEMBER 2009