Full text of DOI:10.1111/j.1365-2621.2000.tb10263.x

JFS: Food Engineering and Physical Properties
The Contribution of Water to the Specific Heat
Change at the Glass-to-Rubber Transition
of the Ternary System BSA-Water-NaCl
C. INOUE AND M. ISHIKAWA
ABSTRACT: We studied the glass-to-rubber transition of a BSA-water-NaCl system using DSC. The Tg was 178-187 K
and decreased with increasing NaCl content from 1:9 to 6:4 NaCl:BSA. The specific heat change (⌬Cp*, J/((g BSA)K))
at Tg increased as the weight ratio increased from 1:9 to 5:5 NaCl:BSA, as did the change in calculated unfrozen
water (UFW). The UFW was also correlated with the amount of NaCl in the glassy region with a correlation coeffi-
cient of approx. 6 (mol H2O)/(mol NaCl), which is comparable to the hydration number of Na+. These results suggest
that UFW could contribute to the ⌬Cp at Tg of the BSA-water system.
Key Words: glass transition, specific heat change, unfrozen water, bovine serum albumin, NaCl
Introduction
creases in the unfrozen water content (Sartor and others 1995).
REEZING IS WIDELY USED AS A FOOD PRESERVATION METHOD. AS The average mean-square displacements for allnonhydrogen at-
food is cooled, water crystallizes, but a fraction of water does oms in hydrated deoxymyoglobin above 200 K were larger than
F
not crystallize and remains as a solvent in the concentrated liq- that of freeze-dried myoglobin powder, as determined from the
uid phase down to a very low temperature (Franks 1985). Various Mšssbauer effect on the iron (Parak and others 1982). These ob-
undesirable changes may occur in this liquid phase (Franks servations and others (Slade and others 1989) indicate the im-
1985), such as the formation of low-molecular-weight products portance of water in the glass transition of protein-water sys-
(Shenouda 1980), discoloration (Bito 1965), and changes in rheo- tems. The reported Tg’s of other protein-water system areis
logical properties of tissues (LeBlanc and others 1988). This dete- (Slade and others 1989) are in good agreement with that of our
rioration can be due to the destruction of cell tissue by ice-crystal tuna system (Inoue and Ishikawa 1997). However, a detailed elu-
growth and to biochemical reactions in the freeze-concentrated cidation of the mechanism of the glass transition of protein solu-
liquid phase at temperatures higher than the glass-transition tions is not available.
temperature (Tg) (Franks 1985). These unfavorable changes
With regard to the mechanism of the glass transition of pro-
would be minimized, if the ice-crystal growth is minimized, and tein solutions, one can ask a simple question. Which molecule is
the freeze-concentrated fraction is kept as an amorphous solid the major contributor to the glass transition? In other words,
below its Tg. Therefore the Tg and the relative amount of water in which molecule loses more degrees of freedom of motion at the
the vitreous phase are the key parameters that govern the quali- Tg and “vitrifies”? From the analysis of protein structures, protein
ty and stability of frozen foods (Slade and others 1989).
molecules are known to lose mobility at a certain temperature
Most research concerning the glass transition of solutions of when frozen. The average mean-square displacements for all
organic materials has focused on the glass-transition tempera- nonhydrogen atoms in freeze-dried myoglobin powder increase
ture (Roos and Karel 1991; Slade and others 1989), and less at- above 200 K, as determined from the Mšssbauer effect (Parak
tention has often been paid to the relative amount of water in and others 1982). The scattering of the atom positions in met-
the vitreous phase. Quantitative information about the glassy re- myoglobin crystals, which are obtained by X-ray diffraction, is
gion of the system can be obtained from the magnitude of the smaller at 80 K than at 220 to 300 K (Frauenfelder and others
specific heat change ( Cp) at the Tg. We have found that the addi- 1979). The absorption spectrum of zinc-substituted myoglobin
tion of NaCl to tuna flesh or its filtrate influences the Cp as well begins to depart from the calculated spectrum for the glassy
as the Tg (Inoue and Ishikawa 1997). The Cp of tuna filtrate is state above 180 K (Ahn and others??? 1995). These temperatures,
governed by the NaCl:organic solute ratio and is independent of taken as Tg values, are in good agreement with those of a tuna fil-
moisture (Inoue and Ishikawa 1997). Although interpretation is trate system (Inoue and Ishikawa 1997) and other calorimetric
difficult due to the complexity of natural samples, this observa- observations (Sartor and others 1995; Chang and Randall 1992).
tion suggests that protein-NaCl interactions might be the major The protein itself could be a candidate molecule, accounting for
factor that influences the glass transition. To examine this hy- the specific heat change at the glass transition. However, quanti-
pothesis, experiments with model solutions are necessary.
tative information such as its contribution to the Cp is not avail-
The glass transition of protein-water systems has been inves- able.
tigated by several methods. With a calorimetric method, it has
On the other hand, there is also evidence that water mole-
been found that an anhydrous myoglobin crystal does not show cules contribute to the specific heat change at the glass transi-
any step-wise change in specific heat (Miyazaki and others tion (Sartor and others 1995). So far, the question is still open.
1993). But fully hydrated powders made from proteins such as The purpose of the present study was to answer the question,
myoglobin, hemoglobin, and lysozyme exhibit step-wise chang- which molecule is the major contributor to the glass transition of
es in specific heat at around 170 K (Sartor and others 1995). The a protein solution? Our results give some insight into the mecha-
C
p of myoglobin, hemoglobin, and lysozyme increases with in- nism of the glass transition of protein solutions and hence could
© 2000 Institute of Food Technologists
Vol. 65, No. 7, 2000—JOURNAL OF FOOD SCIENCE 1187

Contribution of Water to the Specific Heat Change . . .
ice
be useful for improving of food-freezing technology.
Hmice = Hm0 ϩ (Cpice (Tmice) Ϫ Cpwater(Tmice)(T m0 Ϫ Tm
)
(2a)
Materials and Methods
OVINE SERUM ALBUMIN (BSA, LOW-SALT TYPE) WAS PUR-
chased from Wako Pure Chemical Industry, Osaka, Japan.
where Hm0 (333.88 J/(g H2O K)) is the integral heat of fusion of
0
B
pure ice at T
(273.15 K) in the literature (Perry and Chilton
m
1973) and Cpice(Tmice) J/(g H2O) and Cpwater(Tmice) J/( H2O K) are
the specific heats of ice and water at the observed melting peak
temperature (Tmice, K), respectively. Cpice and Cpwater were calcu-
lated by interpolation and extrapolation of the reported specific
heats of ice and water (Perry and Chilton 1973), respectively.
The aw eutectic in Eq. 1 was calculated from the following equa-
tion:
Binary systems, BSA-water mixtures, were prepared by mixing
weighed BSA and distilled water to moisture ranges from 60 to 80
wt % water. Ternary systems, the BSA-water-NaCl mixtures, were
prepared by mixing weighed BSA, NaCl (anhydrous, special
grade, Kokusan Chemical Works, Tokyo), and the appropriate
amount of distilled water to make the moisture level be between
60 and 90 wt % water. The mixtures were placed in a refrigerator
at 278 K, until they became transparent.
aweutectic = (qeutectic/⌬Hmeutectic) ␣
(3)
We used a differential scanning calorimeter, Model DSC-7
(Perkin-Elmer Corp., Norwalk, Conn.), to observe the thermal
behavior of the samples. Liquid nitrogen was used as a coolant.
Hermetically sealable 20-_l aluminum DSC pans (Perkin-Elmer)
were used in all measurements, with an empty aluminum pan as
the reference. The instrument was calibrated every d for heat
flow and temperature, using distilled water (melting point, 273 K;
Hm0, 333.88 J/(g H2O), Perry and Chilton 1973) before measure-
ments were made. Calibration was carried out from 243 K to 303 K
at a cooling rate of 50 K/min and a heating rate of 30 K/min. A
portion of each sample was weighed into a DSC pan and sealed
just before analysis. Sample weights were in the range of 4.78 to
13.44 mg. The sample was cooled at 50 K/min from 293 to 118 K,
held at 118 K for aminute, and then heated from 118 to 303 K at
30 K/min; this procedure was repeated at least twice for all sam-
ples. Three aliquots of all solutions were measured.
The endothermic step-wise change in the DSC heating curve
was regarded as the glass-to-rubber transition. In the present
study, the temperature of this step-wise change was determined
using 2 methods: We determined the temperature at which the
extended lower baseline intersected an extension of a line that
was tangential to an inflection point of the step-wise curve (Tg
set) and the temperature at the maximum of the derivative of the
DSC heating curve (Tgmiddle).
where qeutectic J/(g sample) is the observed integral heat of fusion
of the eutectic crystal, ␣ (0.7669) is the weight fraction of water in
the eutectic mixture (Landolt–Börnstein 1972), and Hmeutectic J/
(g eutectic mixture) is the integral heat of fusion of the eutectic
crystal as a function of temperature. Hmeutectic J/(g eutectic mix-
eutectic
ture) was determined from a linear approximation of Hm
against temperature (Landolt-Bšrnstein 1972), as follows:
⌬Hmeutectic = 1.8 ϫ Tmeutectic Ϫ 214.6
(3a)
where Tmeutectic is the melting peak temperature of the eutectic
crystal. The awrecrystallization in Eq. 1 was calculated from the follow-
ing equation, in the same manner as aw eutectic in Eq. 3:
awrecrystallization = (qrecrystallization / Hmrecrystallization) ␣
(4)
where qrecrystallization J/(g sample) is the observed integral heat of
recrystallization, and Hmrecrystallization J/(g eutectic mixture) is the
heat of recrystallization at the recrystallization peak temperature
(Tmrecrystallization), evaluated from the following equation:
on-
⌬Hmrecrystallization = 1.8 ϫ Tmrecrystallization Ϫ 214.6
(4a)
The amount of NaCl per BSA in the glass(aNglass, (g NaCl)/
(BSA)) was calculated from the following equation:
The change in the specific heat capacity through the
glass-to-rubber transition
We determined the Cp J/(g sample K) by computing the dis-
tance between the baseline curves before and after the step-
wise change (Saito 1990; Inoue and Ishikawa 1997).
aNglass = aN Ϫ (aNeutectic Ϫ aN
)
(5)
recrystallization
The unfrozen water content (UFW, (g H2O)/(g BSA)) was cal- where aN (g NaCl)/(g BSA) is the total amount of NaCl in the sam-
culated from the following equation:
ple, aN
(g NaCl)/(g BSA) and aN
(g NaCl)/(g
eutectic
recrystallization
BSA) are the amounts of NaCl in the eutectic crystal and recrys-
(1) tallized solid, respectively. aNeutectic and aNrecrystallizationwere calcu-
lated from the next 2 equations,
UFW = aw aw ice Ϫ (aweutectic Ϫ aw
)
recrystallization
where aw (g H2O)/(g BSA) is the total amount of water in the sam-
ple, awice, aweutectic, and awrecrystallization are the amounts of water (g
H2O)/(g BSA) contributing to the enthalpy change at the melting
of ice eutectic crystal and recrystallization exotherm, respective- and
ly. The recrystallization peak was not observed in the case of low-
er-NaCl samples (NaCl:BSA_2:8); therefore, awrecrystallization = 0 at
NaCl:BSA_2:8. We regarded the recrystallized solid as being com-
posed of the eutectic mixture because the recrystallization peak
aNeutectic = (qeutectic/ Hmeutectic) (1 Ϫ ␣)
(6)
aNrecrystallization = (qrecrystallization / Hmrecrystallization) (1 Ϫ ␣) (7)
The moisture of the as-received BSA, determined using a
was observed whenever the eutectic peak was observed. The thermogravimetric method (383 K, 60 min, TGA-7, Perkin-Elmer)
awice was calculated from the following equation:
was 5.16 wt % water.
awice = (qice/⌬Hmice) bBSA
(2)
Results and Discussion
HE COMPOSITION OF BINARY BSA-WATER AND TERNARY BSA-
water-NaCl systems tested is shown in Figure 1. The points
on the lines parallel to AA’ represent samples with the same
moisture content but different weight ratios of NaCl and BSA.
Points on the radial lines from the water apex represent samples
with the same weight ratio of NaCl and BSA but with different
where qice J/(g sample) is the observed integral heat of fusion of
ice per sample, bBSA (gBSA)/(g sample) is the weight fraction of
BSA in the sample, and ⌬Hmice J/(g H2O) is the integral heat of fu-
sion of pure ice as a function of temperature. Hmice J/(g H2O) was
calculated from the following equation:
T
1188 JOURNAL OF FOOD SCIENCE—Vol. 65, No. 7, 2000

Contribution of Water to the Specific Heat Change . . .
moisture contents.
Typical DSC heating curves for 80 wt % water samples are de-
In all samples, 2 DSC heating curves were obtained from re- picted in Figure 2 and 3 (Figure 3 is an enlargement of Figure 2).
peated scans. The maximum difference between repeated scans Step-wise changes were observed at 175 to 195 K in all DSC
was less than 0.08 mW, which corresponds to 1.6 ϫ 10Ϫ2 J/(g K) in curves (Figure 3). We consider this step-wise change to be the
specific heat units. This showed that all changes detected in DSC glass-to-rubber transition because an exothermic peak was ob-
heating curves in the present study were reproducible, and that served between the step-wise change and the melting peak in
detectable irreversible changes did not occur during the cooling some DSC curves. We can assign this exothermic peak to the re-
and heating operations.
crystallization of the supercooled liquid fraction (for example,
see Figure 2).
NaCl is known to influence Tg in a tuna filtrate system (Inoue
and Ishikawa 1997). In Figure 4, the Tgmiddle’s of the binary and
Figure 1—Initial concentrations of BSA-water and BSA-water-NaCl
systems
Figure 3—DSC heating curves for a BSA-water-NaCl system. This is
an enlarged version of Figure 2
Figure 4—Glass-to-rubber transition temperature (Tg
) of BSA-
Figure 2—Typical DSC heating curves for a BSA-water-NaCl system. water and BSA-water-NaCl systems. Standard deviamtiidodnle of three
Curve (a) is for a BSA-water system, curves (b) to (g) are for BSA- samples is shown. Tuna filtrate-NaCl systems data were from Inoue
water-NaCl systems.
and Ishiwaka (1997).
Vol. 65, No. 7, 2000—JOURNAL OF FOOD SCIENCE 1189

Contribution of Water to the Specific Heat Change . . .
the ternary systems were plotted against the weight ratio of NaCl obtained for tuna filtrate systems in our previous work (Inoue
and BSA. The glass to glass-to-rubber transition temperature, Tg- and Ishikawa 1997).
middle, of the binary BSA-water system was in the range of 188 to
BSA-water systems can be used to model the protein-water
197 K. The Tgmiddle of the ternary system BSA-water-NaCl system system of tuna filtrate systems. Chang and Randall (1992) re-
was about 10 K lower than that of the binary system. The Tg of ported that the Tgmiddle of the BSA-water system is 192 K at 90 wt
the ternary systems decreased slightly as the NaCl: BSA ratio in- % water, which is in good agreement with our results: 1 K at 80 wt
creased, which was similar to the behavior of that in the tuna fil- % water, 190 K at 70 wt % water, and 185 K at 60 wt % water (Table
trate system. The values of Tgmiddle in the present work were 5 to 1). The Tgmiddle’s of 6 protein solutions were determined to be in
20 K lower than those of tuna filtrate systems. Moisture content the range of 186 K to 206 K by Chang and Randall (1992). Some of
middle
did not appear to affect Tg
tems.
of the binary and ternary sys- the Tgmiddle values in Table 1 were calculated to be in the range of
170 to 177 K, by adding half of the width of the glass-to-rubber
onset
The specific heat change at the glass-to-rubber transition, transition temperature range to the Tg
values reported by
Cp* J/(g BSA K), of the binary and ternary systems is plotted Green and others (1994) and Sartor and others (1992, 1995). The
against the weight ratio of NaCl and BSA in Figure 5. The Cp* of available data for the Tgmiddle of fully hydrated protein-water sys-
the ternary systems increased with increasing weight ratio of tems fall within the range of 170 to 206 K.
NaCl and BSA, and it was not affected by moisture content in
The effect of the particular protein species on the glass-to-
samples with low NaCl:BSA ratios (0:10 to 5:5). The Cp* de- rubber transition temperature in fully hydrated systems is not
creased as the NaCl:BSA ratio increased between 6:4 and 7:3.The evident. The Tgmiddle decreases slightly with increasing moisture
Cp* of the BSA-water and BSA-water-NaCl systems was in good content for lysozyme (Sartor and others 1995) and BSA (present
agreement with that of tuna filtrate (Figure 5).
work) in fully hydrated systems. This could suggest that the
Calculated unfrozen water (UFW, (g H2O)/(g BSA)) content is glass-to-rubber transition of protein-water systems is not direct-
plotted against NaCl:BSA in Figure 6. The UFW content of the ly related to specific higher-order structures but is governed by
ternary system increased with increasing NaCl between 0:10 and lower-order structures, such as hydrophilic groups on the surface
5:5 for NaCl:BSA with 60 to 80 wt % water, which was similar to the of protein molecules. According to the method of Berlin and oth-
trend observed in Cp* (Figure 5). The data for the 90-wt-% water ers (1970) and Mrevlishvili and Privalov (1969), the UFW content
samples were not plotted because of their high sensitivity to the of our BSA-water system was 0.65 Ϯ 0.13 (g H2O)/(g BSA) (Table
choice of the baseline.
2). This result was higher than the UFW contents reported by
We investigated the glass transition of the protein-water sys- Berlin and others (1970) and Mrevlishvili and Privalov (1969). Al-
tem for numerous proteins (Table 1). Unfortunately, these data though their experimental details were not given by Berlin and
cannot be compared directly because of variations in the defini- others (1970) and Mrevlishvili and Privalov (1969), our high cool-
tion of the glass-transition temperature. There are 3 ways to de- ing rate of 50 K/min might have resulted in the larger UFW we
fine the glass-to-rubber transition temperature (Tg) from a DSC- measured. The UFW content of the tuna filtrate system, evaluat-
heating curve: the onset point, the mid-point, and the end point ed using the same method as in the present study, had a wider
(Seyler 1994). Among these, the onset point or the mid-point variance (Table 2). The complexity of components in that real
combined with the width of the transition is widely used. In the food system might have caused such a wide scatter of UFW val-
present study, the onset point (Tgonset) for the BSA-water system ues.
was 161 to 177 K, and the mid-point (Tgmiddle) was 184 to 192 K.
These were a bit lower than, but still comparable to, the results
Figure 6—Unfrozen water (UFW) content of BSA-water and BSA-wa-
ter-NaCl systems. Calculated UFW of BSA-water and BSA-water-NaCl
with from 60 to 80 wt % moisture was plotted against the weight
Figure 5—⌬C * of BSA-water and BSA-water-NaCl systems. Standard ratio of NaCl:BSA. Standard deviation of three samples is shown. 90
deviation of p3 samples is shown. Tuna filtrate-NaCl systems data wt % water samples are not shown, because of their high sensitivity
were from Inoue and Ishiwaka (1997).
to the choice of the baseline.
1190 JOURNAL OF FOOD SCIENCE—Vol. 65, No. 7, 2000

Contribution of Water to the Specific Heat Change . . .
Table 1—Tg of BSA-water and other protein-water systems
Table 2—Unfrozen water content of BSA-water and tuna filtrate sys-
tems
Sample
Moisture (wt %)
Tgonset (K)
Tgmiddle (K)
UFW g/(g protein) rate of cooling (K/min)
reference
BSA
60
70
80
90
90
90
90
172 Ϯ 2
174 Ϯ 6
177 Ϯ 13
189 Ϯ 1
189 Ϯ 3
192 Ϯ 5
192
186
193
206
195
187
0.65 Ϯ 0.13
0.49 Ϯ 0.025
0.32 Ϯ 0.005
0.67 Ϯ 0.29 a
50
n.a.
n.a.
50
present work
Berlin and others (1970)
Mrevlishzili and Privalov (1969)
Inoue and Ishikawa (1997)
BSA e
Ovalbumine
Myoglobin e
a
tuna filtrate g/(g solute) from our previous work
e
Lysozyme
Lactate dehydrogenase e 90
e
Elastase
90
42
42
39
29 to 42
50
49
14
16
MetHb f
167
172
162
169
165
170
280
265
236
174 a
177 a
170 a
176 to 178 a
lished data for other binary systems and plotted them in Fig 7.
For example, the lysozyme-water system is included in the
present results. The ⌬Cp* of poly-L-ASN systems (Green and oth-
ers 1994) and the ⌬Cp* of met-hemoglobin and met-myoglobin
(Sartor and others 1992, 1995) are within the variance of the cor-
relation. These included data that do not contradict our hypoth-
esis.
The slope of the line in Figure 7 is 2.43 Ϯ 0.44 J/(g H2O K)
(95%). According to our hypothesis, this could be interpreted as
the specific heat change at the glass-to-rubber transition of the
UFW. We estimated the difference in specific heats of pure water
and ice at 180 and 190 K by subtracting the specific heat value
for ice from the extrapolated specific heat value for pure water at
180 and 190 K. The estimated specific heat differences between
pure water and ice were 2.82 J/(g H2O K) and 2.73 J/(g H2O K) at
180 and 190 K, respectively. Our Cp of UFW, derived from the cor-
relation (Figure 7), was 15% smaller than these values. The UFW
in the present systems would be expected to have a lower degree
of freedom than that of pure liquid water because protein mole-
cules, sodium ions, and chloride ions would restrict its mobility.
MetMb f
Lysozyme
MetHb g
f
Myoglobin h
h
Cytochrome c
Poly-L-ASN h
289 a
273 a
247 a
230 a
21
29
219
i
Lysozyme
10.5
13.5
15.4
0 to 19
68 to 90
300 b
238 b
219 b
253 to 140c
186 to 195
Legumin j
Tuna filtrate
k
200 to 202
a
calculated by the authors from addition of half of the reported width of glass-to-rubber
onset.b
c
detected by positron annihilation lifetime spectroscopy. determined by
transition to T
g
d
e
f
the authors from the depicted DSC curves. present work. Chang and Randall (1992). Sartor
and others (1995). Sartor and others (1992). Green and others (1994). Gregory and Chai
(1993). Sochava (1997). Inoue and Ishikawa (1997).
g
h
I
j
k
The calculated UFW content of the ternary systems was af-
fected by the NaCl:BSA ratio but not by moisture content (Figure
6). These results suggest that there may be 2 states in the BSA-
water-NaCl system, depending on the NaCl:BSA ratio. In the
lower NaCl content region, where NaCl:BSA is less than or equal
to 5:5, UFW (g H2O)/(g BSA) increased with increasing NaCl,
which confirms the effect of NaCl on the formation of UFW in the
BSA-water-NaCl system. In the higher NaCl content region, UFW
did not vary with the NaCl:BSA ratio. BSA may be denatured by
high NaCl concentrations. Inthe case of the 90-wt-% water sam-
ples, the sensitivity of the calculated UFW value to the choice of
the DSC baseline was substantially larger than for other data.
Hatley and others (1991) examined the influence of the defi-
nition of the ice melting peak baseline on the measured peak
area and found that the area of a typical ice melting peak for a
sucrose solution changed by up to 12%, depending on the choice
of the baseline. Our choice of the ice melting peak baseline corre-
sponds to that given as the minimum peak area by their meth-
ods (Hatley and others 1991). We checked to see if the choice of
the baseline would affect the UFW value by multiplying our re-
sults for the peak area by 1.10. When we did this, the calculated
UFW values for some of the 90-wt-% water samples became neg-
ative. We concluded, therefore, that UFW data for 90-wt-% mois-
ture samples are not reliable. In our calculation method for UFW,
it is obvious that samples with higher moisture content would be
more sensitive to any experimental error because the fraction of
UFW in the total amount of water was lower in samples with high-
er moisture contents, and therefore a small error in the ice melt-
ing peak area would result in a significant change in the calculat-
ed amount of UFW per BSA. In the present study, we used the re-
sults for 60-to-80-wt-% water samples in our discussion.
In the present work, the ⌬Cp* in the lower NaCl content region
correlated well with the UFW content (Figure 7). This correlation
is consistent with our hypothesis that the glass transition of pro-
tein solutions is governed by a lower-order structures of a pro-
tein, such as the hydrophilic groups on the surface of a protein.
The data for the tuna filtrate system (Inoue and Ishikawa 1997)
are also included in these results. We calculated ⌬Cp*’s from pub-
Figure 7—The correlation between the ⌬Cp* and UFW of BSA-water
and BSA-water-NaCl systems. The ⌬Cp* (Figure 5) as compared with
UFW (Figure 6) was plotted from 0:10 to 5:5 for NaCl:BSA and from
60 to 80 wt % water. The width of the ⌬Cp* and UFW is the scatter of
three samples, as shown of Figures 5 and 6. The slope of the line is
2.43 ؎ 0.44 (95%). The calculated ⌬Cp* values, using the reported
date (Table 1) for protein-water systems, were also plotted. aGreen
and others 1994. bSartor and others 1995. cSartor and others 1992.
dInoue and Ishiwaka 1997.
Vol. 65, No. 7, 2000—JOURNAL OF FOOD SCIENCE 1191

Contribution of Water to the Specific Heat Change . . .
Therefore, the Cp of UFW at the glass-to-rubber transition would retain their mobility down to temperatures much lower than the
be smaller than that of pure water.
ice-crystallization temperature. This unfrozen water would lose
Several experimental measurements of the ⌬Cp of pure water its mobility only at Tg with an accompanying change in specific
at its glass-to-rubber transition have been reported: 1.94 J/(g heat. (2) Sodium ions could form a hydration shell with bulk wa-
H2O K) at 135 K (Sugisaki and others 1968); 1.05-1.38 J/(g H2O K) ter molecules. Some of these clusters would remain at the surface
at 139 K, which was extrapolated from the Cp of salt solutions of the protein molecule, in an intramolecular cavity of the pro-
(Angell and Tucker 1980); and 0.08 0.005 J/(g H2O K) at 136 1 K tein, when the NaCl:BSA ratio was less than or equal to 5:5. In
(Hallbrucker and Mayer 1989). Our Cp value is similar to those of such a location, the degree of freedom of water molecules would
Sugisaki and others (1968) and Angell and Tucker (1980), if we be restricted by the presence of both protein molecules and sodi-
take into account the difference in the glass-transition tempera- um ions. Therefore, these water molecules could form neither ice
tures of pure water and the present system. However, the report- nor eutectic crystals but would retain their mobility down to
ed values of Cp for pure water are scattered in a wide range.
much lower temperatures than the ice- or eutectic-crystallization
The amount of UFW in the present ternary systems appeared temperatures. This unfrozen water would lose its mobility only at
to correlate with the amount of NaCl in the glass in the lower the Tg, with an accompanying change in specific heat.
NaCl:BSA region (less than or equal to 5:5) (Figure 8). The corre-
Conclusions
lation coefficient was 1.80 Ϯ 0.28 (g H2O)/(g NaCl), which would
corresponds to 6 H2O molecules per NaCl molecule. The reported
hydration number forsodium ions, which depends upon the ex-
perimental method, is from 4 to 6 H2O molecules per sodium ion
(Franks 1985). Our result of 6 H2O molecules per NaCl molecule is
in good agreement with that of Enderby and Nielson (1979), ob-
tained using a neutron scattering method. Our results suggest
that the UFW in the ternary system might be associated with the
sodium ions in the glass.
The glass-to-rubber transition temperature and the specific
heat change for the ternary BSA-water-NaCl system are influ-
enced by the ratio of NaCl:BSA but not by moisture content un-
der our experimental conditions. The specific heat change at the
glass-to-rubber transition of the ternary protein-water-NaCl sys-
tem might be attributable to a change in the degrees of freedom
of the UFW. The amount of UFW is determined by the NaCl:BSA
ratio.
The glass transition in the BSA-water-NaCl system could be
explained by the following processes: (1) In binary protein-water
solutions, water molecules in the vicinity of protein molecules
could interact with hydrophilic groups at the surface of the pro-
tein molecules. Such water molecules might have a lower degree
of freedom than that of water molecules in the bulk state. There-
fore, these water molecules would not form ice crystals but would
List of Symbols
total amount of water in the sample per BSA (g
H2O)/(g BSA)
aw
aw
aw
aw
aN
aN
amount of water contributing to the heat of ice
melting per BSA (g H2O)/(g BSA)
ice
amount of water contributing to the heat of eu-
tectic melting per BSA (g H2O)/(g BSA)
eutectic
recrystallization
amount of water contributing to the recrystalli-
zation exotherm per BSA (g H2O)/(g BSA)
total amount of NaCl in the sample per BSA (g
NaCl)/(g BSA)
amount of NaCl per BSA in the glass (g NaCl)/
(g BSA)
glass
aN
amount of NaCl in the eutectic crystal per BSA
(g NaCl)/(g BSA)
eutectic
aN
amount of NaCl in the recrystallized solid per
BSA (g NaCl)/(g BSA)
recrystallization
bBSA
Cp
weight fraction of BSA in the sample (g NaCl)/
(g BSA)
specific heat change per sample at Tg, J/(g sam-
ple K)
Cp*
specific heat change per BSA at Tg, J/(g BSA K)
specific heat of ice, J/(g H2O K)
ice
Cp
water
Cp
specific heat of water, J/(g H2O K)
0
Hm
integral heat of fusion of water at Tm0, 333.88 J/
(g H2O)
ice
Hm
integral heat of fusion of pure ice at Tmice, J/(g
H2O)
eutectic
Hm
Hm
qice
integral heat of fusion of eutectic mixture at
Tmeutectic , J/(g eutectic mixture)
recrystallization
integral heat of recrystallization of eutectic mix-
ture at Tmrecrystallization , J/(g eutectic mixture)
integral heat of fusion of ice per sample, J/(g
sample)
Figure 8—The correlation between UFW content and NaCl content in
the glass. The calculated UFW (Figure 6) as compared with the calcu-
lated NaCl in the glass was plotted from 0:10 to 5:5 NaCl-BSA, and
from 60 to 80 wt % moisture content. The slope of the line is
1.80 ؎ 0.28 (95%). The y-axis intercept is 0.60 ؎ 0.083 (95%).
qeutectic
integral heat of fusion of eutectic crystal per
sample, J/(g sample)
1192 JOURNAL OF FOOD SCIENCE—Vol. 65, No. 7, 2000

Contribution of Water to the Specific Heat Change . . .
tion as a probe of protein structural dynamics. Nature (Lond) 280:558-563.
Green JL, Fan J, Angell CA. 1994. The protein-glass analogy: some insights from homopep-
tide comparisons. J Phys Chem 98:13780-13790.
qrecrystallization
integral heat of recrystallization per sample, J/
(g sample)
Gregory RB, Chai K-J. 1993. 4th International Workshop on Positron and Positronium Chem-
istry. Ed. I. Billard. J Physique IV 3:305-310.
0
Tm
equilibrium melting temperature of pure ice,
273.15 K
Hallbrucker A, Mayer E. 1989. The heat capacity and glass transition of hyperquenched
glassy water. Philosophical Magazine B 60:179-187.
Hatley RHM, van den Berg C, Franks F. 1991. The unfrozen water content of maximally
freeze-concentrated carbohydrate solutions: validity of the methods used for its deter-
mination. Cryo-Letters 12:113-124.
Inoue C, Ishikawa M. 1997. Glass transition of tuna flesh at low temperature and effects of
salt and moisture. J Food Sci 62:496-499.
Landolt-Börnstein 6 Auflage “Ahlenwerte und Funktionen” Springer Verlag. 1972, p 203.
LeBlanc EL, LeBlanc RJ, Blum IE. 1988. Prediction of quality in frozen cod (Gadus morhua)
fillets. J Food Sci 53:328-340.
ice
Tm
melting peak temperature of ice, K
melting peak temperature of eutectic crystal, K
recrystallization peak temperature, K
glass transition temperature, K
eutectic
Tm
recrystallization
Tm
Tg
onset
Tg
onset point of the glass transition temperature,
K
Miyazaki Y, Matsuo T, Suga H. 1993. Glass transition of myogrobin crystal. Chem Phys Lett
213:303-308.
Mrevlishvili GM, Privalov PL. 1969. Water in biological systems. Kayuskin LP, editor. New
York: Plenum Press. p 63.
Parak F, Knapp EW, Kucheida D. 1982. Protein dynamics. Mössbauer spectroscopy on deox-
ymyoglobin crystal. J Mol Biol 161:177-194.
Perry RH, Chilton CH. 1973. Chemical engineer handbook 3. p 110.
RoosY, Karel M. 1991. Phase transitions of amorphous sucrose and frozen sucrose solutions.
J Food Sci 56:266-267.
Saito Y. 1990. Foundation of thermal analysis (in Japanese). Tokyo: Kyoritushuppan Ltd.
Press. p 126-127.
middle
Tg
mid-point of the glass transition temperature,
K
end
Tg
end point of the glass transition temperature, K
UFW
amount of unfrozen water per BSA (g H2O)/ (g
BSA)
weight fraction of water in the eutectic mixture, 0.7669 (g H2O)/(g
eutectic mixture)
Sartor G, Hallbrucker A, Hofer K, Mayer E. 1992. Calorimetric glass-liquid transition and
crystallization behavior of a vitreous, but freezable, water fraction in hydrated methemo-
globin. J Phys Chem 96:5133-5138.
Sartor G, Hallbrucker A, Mayer E. 1995. Characterizing the secondary hydration shell on
hydrated myoglobin, hemoglobin, and lysozyme powders by its vitrification behavior on
cooling and its calorimetric glass-liquid transition and crystallization behavior on reheat-
ing. Biophys J 69:2679-2694.
Seyler RJ, editor. 1994. Assignment of the glass transition. ASTM STP 1249. The American
Society for Testing and Materials. Philadelphia, PA.
References
Ahn JS, Kitagawa T, Kanematsu Y, Nishikawa Y, Kushida T. 1995. Glass Transition of Zn-
substituted myoglobin probed by adsorption and site-selective fluorescence spec-
troscopies. J Lumin. 61:81-85.
Shenouda SYK. 1980. Theories of protein denaturation during frozen storage of fish. Adv
Food Res 26:275-311.
Angell CA, Tucker JC. 1980. Heat capacity changes in glass-forming aqueous solutions and
the glass transition of vitreous water. J Phys Chem 84:268-272.
Berlin E, Kliman PG, Pallansch MJ. 1970. Changes in state of water in protein aqueous
systems. J Colloid Interface Sci 34:488-494.
Bito M. 1965. Studies on the retention of meat color of frozen tuna-II. Effect of storage
temperature on preventing discoloration of tuna meat during freezing storage. Bull Jap
Soc Sci Fish 31:534-539.
Slade L, Levine H, Finley JW. 1989. Protein-water interactions: water as a plasticizer of gluten
and other protein polymers. In: Protein quality and the effects of processing. Phillips RD,
Finley JW, editors. New York: Marcel Dekker. p 9-124.
Sochava IV. 1997. Heat capacity and thermodynamic characteristics of denaturation and
glass transition of hydrated and anhydrous proteins. Biophys Chem 69:31-41.
Sugisaki M, Suga H, Seki S. 1968. Calorimetric study of the glassy state. IV. Heat capacities of
glassy water and cubic ice. Bull Chem Soc Japan 41:2591-2599.
MS 19991133
Chang BS, Randall CS. 1992. Use of subambient thermal analysis to optimize protein lyo-
philization. Cryobiology 29:632-656.
Enderby JE, Nielson GW. 1979. Water—a comprehensive treatise. Vol. 6. New York: Plenum
Press. p 1-45.
Franks F.1985. Biophysics and biochemistry at low temperatures, Cambridge: Cambridge
University Press, p 195-219.
Authors are with the Dept.of Food Science and Technology,Tokyo University
of Fisheries, 4-5-7 Konan, Minato-Ku, Tokyo 108-8477, Japan. Direct corre-
spondence to Chiharu Inoue (E-mail: chiharui@vbl.hiroshima-u.ac.jp).
Frauenfelder H, Petsko GA, Tsernoglou D. 1979. Temperature dependent X???-ray diffrac-
Vol. 65, No. 7, 2000—JOURNAL OF FOOD SCIENCE 1193