Professional Documents
Culture Documents
Calorimetry in Food Processing by Gonul Kaletunc
Calorimetry in Food Processing by Gonul Kaletunc
Food Processing:
Analysis and
Design of Food Systems
Calorimetry in
Food Processing:
Analysis and
Design of Food Systems
EDITOR
Gnl Kaletun
Dedication
vii
Dedication
viii
Table of Contents
Preface
Contributor List
xiii
xvii
Chapter 2
Chapter 3
Chapter 4
15
51
67
87
Table of Contents
Chapter 6
Chapter 7
Chapter 8
Part 2
Chapter 9
Chapter 10
Chapter 11
Chapter 12
119
147
169
199
201
237
265
289
Table of Contents
xi
Chapter 13
Chapter 14
Chapter 15
Index
311
351
369
Preface
xiv
Preface
Preface
xv
Contributors
xviii
Contributors
Contributors
xix
Calorimetry in
Food Processing:
Analysis and
Design of Food Systems
Part 1
Analysis of Food and Biological
Materials by Calorimetry
Chapter 1
Calorimetric Methods as Applied to Food:
An Overview
Gnl Kaletun
Introduction
Calorimetry
An Overview of the Book
References
5
6
8
13
Introduction
Several thermal and nonthermal methods are applied to process and
preserve food materials and to manufacture value-added products. The
goals of food processing are to inactivate spoilage and pathogenic
microorganisms and to maintain this status in storage during the
intended shelf life of the product. During processing, changes take
place in food components, including vitamins, lipids, carbohydrates,
and proteins. Such changes lead to structural and functional changes
in foods at the micro- and macromolecular levels that affect the physical, organoleptic, and nutritional properties of the food.
Food materials are complex biological systems. Food products may
have a broad range of structures spanning the three states of matter,
including dilute to concentrated liquids, solids, and mixtures of multiliquid, liquid-solid, liquid-gas, and solid-gas structures. The combination of complex structures making up complex biological compounds
makes the characterization of food systems challenging. To address
the wide variety of compositions and structures, many biophysical
techniques are uesd to characterize the structure and properties of
food materials before and after processing to develop a fundamental
5
Calorimetry
Among biophysical techniques, calorimetry presents itself as particularly well suited for analysis of food materials. Among many reasons,
the first is the relevance of the experimental protocols of calorimetry
to the majority of processes employed in food preservation. Specifically, because many food-processing methods involve thermal treatment (heating, cooling, freezing) of the materials, thermal
characterization of food systems and their components leads to data
that can be related directly to the processing protocols. Determination
of thermal properties of food materials, such as specific heat as a function of temperature, is essential for heat transfer and energy balance
calculations (Kaletun 2007). Generation of a reliable database to
develop equations predicting thermal properties of food materials for
optimization of food processes can be accomplished by using calorimetry. Moreover, food materials and their components go through conformational and phase transitions during processing. Calorimetry data
can be analyzed to evaluate the thermal and thermodynamic stability
of various phases for a rational design of food product formulations
and process conditions.
Differential scanning calorimetry, which measures heat capacity
as a function of temperature, is a well-established thermal analysis
technique that detects and monitors thermally induced conformational
transitions and phase transitions as a function of temperature. During
temperature scanning, depending on the complexity of the material,
many peaks or inflection points (one to several) reflecting the thermally
induced transitions can be observed. The direction of the peak
corresponds to the nature of the transition, being heat absorbing (endotherms) or heat releasing (exotherms). While melting of solids and
denaturation of proteins display endotherms, crystallization of carbohydrates and aggregation of proteins manifest themselves as exotherms.
The temperatures for the endothermic and exothermic transitions and
the heat involved in such transitions are measured using a calorimeter. Inflection points are indicative of glass transitions; that is,
transitions from a glassy to rubbery state. The transition temperatures
(Tpeak or Tg) reflect the thermal stability of the phase or state going
through the transition. One can extract from calorimetry data values
for the thermal and thermodynamic changes in free energy (G),
enthalpy (H), entropy (S), and heat capacity (Cp) of the various
transitions in addition to determination of the bulk heat capacity of the
material.
The basis for thermodynamic study of food materials is that the
relevant initial and final states (preprocessing and postprocessing
states) can be defined and the energetic and structural differences
between these states can be measured using calorimetric instrumentation. To this end, calorimetry can be used to evaluate the effect of other
physical and chemical variables by comparing the thermograms of the
materials before and after exposure to the variable outside the
calorimetry.
The basics of application of calorimetry to food materials are discussed in detail in this book. However, it is important to start the
discussion with a summary of the advantages of using calorimetry for
study of biological materials. These advantages can be outlined as
follows:
Direct measurement of the energetics of the transition is obtained
(H and Cp). The experimental results are not model dependent.
Calorimetry can be applied to a range of materials, pure or complex.
Materials do not have to be optically transparent or have chromophores as required by spectroscopic methods.
Materials do not have to be uniform or have to be a homogeneous
mixture. In fact, in addition to pure materials, the technique can be
used to evaluate the interactions among the components in a complex
system and how the interactions are altered by the processing.
Calorimetry does not require elaborate or destructive sample
preparation.
Calorimetry is an established technique which has been around since
the 16th century (Haines 1995). Today, the instruments are highly
developed for accurate measurement of thermal events. The theory
behind the technique is well developed, which facilitates interpretation of the data (Hhne et al. 2003).
While the technique is powerful, the validity and utility of the data
depend strongly on the careful use of the equipment and correct interpretation of data. Some analytical methods provide results specific to
materials; however, calorimetry data depends on the conditions used
during the experiment (Haines 1995). One must be careful in choosing
the calorimetry parameters:
1. Time scale: Especially in dynamic measurement systems, for events
to be detected the experimental time scale should match the time
scale of the observed event.
2. Magnitude of the heat flow: If the energy associated with the transition is small, it can lead to ambiguities in its detection. Increasing
the scanning rate enhances the signal; however, it may cause deviation from equilibrium conditions, which requires models beyond
the standard equilibrium thermodynamics treatment of calorimetry
data.
3. Moisture loss during experiment: Biological samples in general are
high-moisture content materials. If the sample cell is not sealed
well, the moisture content of the sample will change due to evaporation during the course of experiment. This may lead to overestimation of the transition temperature as well as the transition enthalpy
change.
4. Interpretation of overlapping peaks: Biological samples may contain
multiple components that undergo thermally induced transitions at
similar temperatures. As a result, overlapping peaks may be observed
on a differential scanning calorimetry (DSC) thermogram. Even if
the origin of the event is known, because the peak temperatures may
shift due to overlap, individual events may appear to happen at different temperatures. The individual peaks can be resolved experimentally (Barrett et al. 2002, 2005), or the complex thermograms
can be deconvoluted by using special software (Fessas and Schiraldi
2000).
10
11
12
13
Foods show chemical reactivity leading to self-heating and selfignition of hot spots. Especially handling of dry powders in bulk, such
as in milling, drying, and packaging, can be dangerous due to potential
dust explosions. Chapter 15 reviews the evaluation by calorimetry of
the thermal consequences of exothermic decompositions in foods,
describes the methodology for quantifying the risk in terms of its severity and its probability, and discusses methods for collecting the stability data correctly. Specific cases of formation of hot spots in dryers,
storage and hot discharge, and transport safety are discussed. The
importance of establishing safe conditions for handling of materials in
prevention of accidents in the food industry is emphasized.
References
Barrett A., Cardello A., Maguire P., Richardson M., Kaletun G., and Lesher L.
2002. Effects of Sucrose Ester, Dough Conditioner, and Storage Temperature on
Long-Term Textural Stability of Shelf-Stable Bread. Cereal Chem, 79(6):
806811.
Barrett A.H., Marando G., Leung H., and Kaletun G. 2005. Effect of Different
Enzymes on the Textural Stability of Shelf-stable Bread. Cereal Chem, 82(2):
152157.
Fessas D., and Schiraldi A. 2000. Starch Gelatinization Kinetics in Bread Dough,
DSC Investigations on Simulated Baking Processes. J Therm Anal Calorim,
61:411423.
Haines P.J. 1995. Thermal Methods of Analysis, Principles, Applications and
Problems. Glasgow: Blackie.
Hhne G.W.H, Hemminger, W., and Flammersheim, H.J. Differential Scanning
Calorimetry: an Introduction for Practitioners. 2nd Ed. Berlin; New York:
Springer-Verlag, 2003.
Kaletun G. 2007. Prediction of Heat Capacity of Cereal Flours: A Quantitative
Empirical Correlation. J Food Eng, 82(2):589594.
Ollivon M., Keller G., Bourgaux C., Kalnin D., Villeneuve P., and Lesieur P. 2006.
DSC and High Resolution X-Ray Diffraction Coupling. J Therm Anal Calorim,
85:219224.
Randzio S.L. 1996. Scanning Transitiometry. Chemical Society Reviews, 25:383.
Yoshida H., Ichimura Y., Kinoshita R., and Teramoto Y. 1996. Kinetic Analysis of
the Isothermal Crystallization of an N-Alkane and Polyethylene Observed
by Simultaneous DSC/FT-IR/WAXD Measurement. Thermochim Acta, 282/283:
443452.
Yoshida H. 1999. Structure Relaxation of N-Alkanes Observed by the Simultaneous
DSC/FTIR Method. J Therm Anal Calorim, 57(3):679685.
Chapter 2
Methods and Applications of
Microcalorimetry in Food
Pierre Le Parlour and Luc Benoist
Introduction
The Heat Flux Calorimetric Principle
DSC versus Heat Flux Microcalorimetry
Comparison between DSC and Heat Flux Microcalorimetry
The Calvet Principle
Calibration
Description of Different Heat Flux Calorimeters Used for Food
Characterization
High Sensitivity Heat Flux Calorimeter
The Mixing and Reaction Heat Flux Microcalorimeter
Methods of Microcalorimetry in Food
Heat Capacity Determination
Heating Mode
Mixing and Reaction Calorimetry
Pressure Calorimetry
Calorimetry under Controlled Relative Humidity
Conclusion
References
15
17
19
19
22
23
26
26
29
30
30
35
40
43
45
45
46
Introduction
Heat is involved at different steps in the preparation of foods, such as
cooking and processing. During heating, cooling, or freezing, the food
products undergo different types of transformations, including melting,
15
16
17
(2.1)
(2.2)
Equation 2.1 shows that the thermal contribution due to the heat capacity of the sample and container is very large and will provide a major
disturbance at the introduction of the container in the calorimeter.
From Equation 2.2, it is also evident that any temperature perturbation
of the thermostatic block will affect the calorimetric measurement.
To solve these issues, a symmetrical calorimeter is preferred. Two
identical calorimetric chambers, one housing a container with
the sample and an identical reference container an inert material
18
Sample + container
Thermostatic block
R
Tp
Heating elements
Heat sink
(the reference container may also be empty) are placed in the thermostatic block at the same temperature, Tp. The heat flux difference is
measured between the two chambers.
dq dqs dqr
dh
dT
dT
=
=
+ C s s Cr r
dt
dt
dt
dt
dt
dt
(2.3)
(2.4)
or by derivation
R
d 2 q dTr dTs
=
dt 2
dt
dt
(2.5)
(2.6)
19
Table 2.1. Some endothermic and exothermic effects for different food types.
Food Type
Endothermic Effect
Exothermic Effect
Fat, oil
Protein
Enzyme
Denaturation
Starch
Gelatinization, glass
transition
Melting
Melting
Melting, glass transition
Crystallization, oxidation
Aggregation,
crystallization
Aggregation, enzymatic
reaction
Retrogradation, oxidation
Milk
Hydrocolloid, gelatin
Carbohydrates
Yeast
Bacteria
Crystallization, oxidation
Gelation
Crystallization,
decomposition
Fermentation
Growth, metabolism,
fermentation
20
21
50
% Flux
40
30
20
0.01 mm
0.05 mm
0.10 mm
10
0
0
150
300
450
600
Temperature (C)
Figure 2.3. Efficiency ratio of a flat-shaped DSC as a function of the sensor plate
thickness.
thermal modeling software, shows that only around half of the heat
flux is dissipated through the plate (Daudon 1996; Le Parlour and
Mathonat 2005). Figure 2.3 clearly shows that the efficiency rapidly
decreases with the temperature and the thickness of the plate. The
efficiency is also affected by the amount of the sample tested. Therefore,
it is recommended to work with small amount of material (about
510 mg) when using a plate DSC to minimize the heat losses. The
thermal conductivities of the crucible and the gas used in the experimental chambers also are very important parameters to be considered
in the efficiency of the heat exchange. For example, a very heatconductive gas (helium) will favor the heat transfer between the
crucible and the detector, but at the same time increase the heat losses.
Hence, the calibration of a plate-type DSC (heat flux or powercompensated type) is very critical and has to be run with the experimental conditions selected for testing the sample.
The main difference between heat flux calorimetry and DSC (heat
flux or power compensated type) is that in a microcalorimeter, the heat
exchange between the sample and the detector is completely measured.
Such a high efficiency is achieved by applying the technological principle developed by Tian and Calvet.
Calorimeters are also designed on the power-compensating principle using a detector that surrounds the sample in the same way.
MicroCal (www.microcal.com) and CSC (now TA Instruments)
22
(www.tainstruments.com) have developed such ultrasensitive instruments, mostly used for the investigations of dilute liquids. Because
these calorimeters operate with fixed vessels, they are not well-adapted
for the characterization of foods.
The Calvet Principle
The detection is based on a three-dimensional fluxmeter sensor. The
fluxmeter element consists of a ring of several thermocouples in
series (Figure 2.4). The corresponding thermopile of high thermal
conductivity surrounds the experimental space within the calorimetric
block. The radial arrangement of the thermopiles guarantees an almost
complete integration of the heat. This is verified by the calculation of
the efficiency ratio that indicates that an average value of 94% 1%
of heat is transmitted through the sensor on the full range of temperature of the Calvet-type DSC (Figure 2.5). In this setup, the sensitivity
of the DSC is not affected by the type of crucible, the type of purge
gas, or the flow rate. The main advantage of the setup is the increase
23
100
200
300
400
500
T (C)
24
K=S/P
S
(2.7)
25
wi
ei g i
q i
(2.8)
i
wi
i
(2.9)
wi
or
E=
(2.10)
Equation 2.10 shows that the power dissipated in the vessel is directly
correlated with the heat flux. The term / corresponds to the calibration factor of the calorimeter.
26
27
Figure 2.8. Standard and mixing vessels (batch), fluid circulation vessel (flow).
Table 2.2. Applications of the MicroDSC technique versus the vessel and the
heating mode.
Vessel
Heating Mode
Component
Application
Batch
Scanning
Isothermal
Scanning
Scanning
Protein (animal,
cereal)
Protein
Enzyme
Starch
Isothermal
Scanning
Starch
Milk
Scanning
Fat
Isothermal
Scanning
Scanning
Fat
Hydrocolloids
Sugar
Isothermal
Scanning
Aroma
Fat, chocolate
Denaturation, aggregation,
lyophilization
Crystallization
denaturation, stability
Gelatinization,
retrogradation, glass
transition
Crystallization (stalling)
Melting, crystallization,
denaturation, aggregation
Melting, crystallization,
lipidic transition,
polymorphism
Crystallization
Melting, gelation
Melting, crystallization,
glass transition
(amorphism)
Stability
Polymorphism versus
pressure
Glass transition versus
pressure
Enzymatic reaction
Wetting
Yogurt processing
Dough and bread processing
Bacteria growth, food safety
Oxidative stability
Enzymatic reaction
Batch high
pressure
Starch
Batch
mixing
Isothermal
One fluid
vessel
Two-fluid
mixing
vessel
Isothermal
Enzyme
Starch
Dairy bacteria
Yeast
Bacteria
Oil
Isothermal
Enzyme
28
29
30
(metal or PTFE). The vessel is fitted with a metallic rod that is operated
from outside the calorimeter. The mixing of components is obtained
by pushing the rod to break the membrane. The rod is also used as a
stirrer during the test.
The ampoule mixing vessel is designed for a slow dissolution
process and for a wetting operation. The sample is sealed under vacuum
in a breakable ampoule. The vacuum operation allows desorbing the
surface of the solid sample for easier dissolution. The sealed ampoule
and the solution are introduced into the vessel. By breaking the ampoule,
the solid and liquid samples are brought into contact.
The Table 2.3 gives an overview of the major calorimetric applications, either in scanning or isothermal modes.
Methods of Microcalorimetry in Food
Microcalorimetry offers a variety of methods that are applied to the
characterization of foods and their components.
Heat Capacity Determination
Heat capacity plays an important role in thermal process and in refrigeration applications. Heat loads, processing times, and industrial equipment sizes are influenced by the heat capacity of the material. Combined
with thermal conductivity and thermal diffusivity, heat capacity data
are needed for modeling of the thermal processes. Heat capacity varies
with temperature and composition, as well as water content (Kaletun,
2007). Because food material can be in solid or liquid form, different
ways of measuring heat capacity using the calorimetric techniques
have been developed.
Heat capacity is thermodynamically defined as the ratio of a small
amount of heat Q added to the substance to the corresponding small
increase in its temperature dT:
C=
Q
dT
(2.11)
H
Cp =
T p
(2.12)
31
Table 2.3. Applications of the C80 calorimetric technique versus the vessel
and the heating mode.
Vessel
Mode
Component
Application
Batch standard
Scanning
Starch
Salt
Carbohydrate
Batch high
pressure
Scanning
Coffee
Oil
Gelatinization, retrogradation
Solubility
Melting, crystallization,
amorphism, decomposition
Safety (roasting),
supercritical CO2
extraction
Self-ignition, explosion
(powder)
Gelatinization under
pressure, glass transition
versus pressure
Polymorphism versus
pressure
Oxidative stability
Oil
Sugar
Salt
Enzyme
Hydrocolloid
Starch
Yeast
Food powder
Neutralization
Dissolution
Dissolution
Enzymatic reaction
Binding
Wetting, gelatinization
Fermentation
Wetting, dissolution
Cereal
Starch
Fat, chocolate
Gas flow
Mixing
(reversing)
Mixing
(membrane)
Mixing
(ampoule)
Scanning,
isothermal
Isothermal
Isothermal
Isothermal
32
(2.13)
where ms and mcs are, respectively, sample mass and vessel mass
(including the cover) and cp(s) and cp(cs) are, respectively, specific heat
capacity of the sample and its vessel.
For the reference side, an empty vessel is used giving the corresponding signal:
dq = ( m c ) dT
cr p( cr )
dt r
dt
(2.14)
where mcr is reference vessel mass and cp(cr) is specific heat capacity of
reference vessel (equal to cp(cs)).
The resulting differential calorimetric signal dq/dt is given by the
following equation:
dq = ( m c + m c m c ) dT
s p( s)
cs p( cs)
cr p( cr )
dt
dt
(2.15)
To get rid of the thermal effect generated by both vessels, the same
test (called blank test) is run with identical empty containers. The following equation describes the blank test heat flow.
dq = ( m c m c ) dT
cs p( cs)
cr p(cr )
dt b
dt
(2.16)
1 dq dq dT
ms dt dt b dt
(2.17)
33
Ab
As
T
time
rately known to determine the specific heat capacity of the sample cp(s)
(expressed in J.g1.C1) at a given temperature. For DSC technique, a
third test is needed using a standard reference material (sapphire) that
has a known specific heat capacity.
cp determination in the temperature step mode
The technique described in previous section is easy to use, but has a
drawback regarding the accuracy of the cp determination. Using the
temperature scanning mode, the sample is continuously heated and is
never at the thermal equilibrium. However, cp is a thermodynamical
parameter, defined at the thermal equilibrium. The temperature step
mode has been developed to address this limitation. A temperature step
is applied to the sample, and the thermal equilibrium is established
(characterized by return of the baseline) after each step. If Equation
2.15 is integrated from time t0 (beginning of the step) to time tn (return
to the baseline), the corresponding equation is obtained:
(2.18)
where cp corresponds to the mean cp value between the two temperatures defining the step of temperature. Q is obtained by integrating the
corresponding surface defined by the calorimetric signal between t0
and tn. The signal corresponding to the blank test is subtracted when
34
1
(Q Qb ) T
ms
(2.19)
35
(2.20)
Q2 Q0 = V T S 2 c p 2
(2.21)
(2.22)
The determination of the specific heat capacity requires the measurement of the density of the liquid sample. This cp measurement does
not need vapor phase correction.
Heat capacity of foods
The specific heat of foods depends on their composition, specifically
the water content (Kaletun 2007). Table 2.4 gives an overview of the
specific heat of selected foods above and below freezing (www.
engineeringtoolbox.com).
Heating Mode
Microcalorimetry is used under the various heating modes are described
next.
Scanning calorimetry
The scanning mode (heating or cooling) is the usual method that
applies to the standard DSC technique. A microcalorimeter also can
36
Type
Fruit
Apple
Grapefruit
Orange juice
Cabbage
Potato
Pork (bacon)
Pork (ham)
Salmon
Carp
Butter
Cream
Milk (cow)
Milk (coconut)
Ice cream
Vegetable
Meat
Fish
Dairy
Cp before Freezing
(J g1C11)
Cp above Freezing
(J g11C11)
1.76
1.84
1.8
1.88
1.72
1.05
1.42
1.55
1.72
1
1.88
1.97
1.76
1.67
3.64
3.81
3.73
3.94
3.43
1.51
2.6
2.97
3.43
1.26
3.77
3.77
3.98
3.1
be used as a DSC, but with low or very low scanning rates (less than
2 C.min1). Longer time of experimentation may be considered to be
a disadvantage, but it provides a better resolution of different thermal
processes.
Melting and crystallization Physical state transformations (crystallization, melting, polymorphism) in fat samples are associated with
thermal effects that are easily measured by DSC. Microcalorimetry
used in the scanning mode allows improvement of the resolution of
different effects because of low scanning rate, especially for characterization of emulsions (Relkin and Sourdet 2005).
Denaturation and aggregation Proteins are the food components
most studied by the microcalorimetric technique and include studies
of conformation changes of food proteins (animal, vegetable, plant),
food enzymes and enzyme preparations for the food industry, as well
as effects of various additives on their thermal properties.
The denaturation and aggregation processes in thermal gelation of
whey proteins were resolved with the microcalorimetric technique
(Fitzsimons et al. 2007). Numerous previous studies of the thermal
gelation of whey proteins, carried out on conventional (fast-scanning)
37
3.2
3.1
90
100
3.0
Heat flow (mW)
85
2.9
2.8
2.7
2.6
2.5
2.4
40
50
60
70
80
Temperature (C)
90
100
Figure 2.11. DSC heating scans (1.0C/min) of 3.0 wt.% WPI pH 7.0) in the
presence of NaCl at the different concentrations (80, 85, 90, and 100 mM NaCl). From
Fitzsimons et al. (2007).
DSC calorimeters (typical sample mass 1550 mg), have shown only
endothermic transitions. Slow transfer of heat into a large vessel
(850 mg of sample) allows the exothermic heat flow from the slow
aggregation process to keep pace with the endothermic heat flow from
the more rapid denaturation process and give a detectable exotherm
(Figure 2.11). The same resolution effect with a slow scanning rate has
also been noticed on pea storage protein, vicilin (Bacon et al. 1989),
on bovine serum albumin (Barone et al. 1992, 1995), and on ovalbumin
(Hagolle et al. 1997; Relkin 2004).
Gelation Microcalorimetry is applied to investigation of gels formed
by biopolymers, such as carrageenan (Williams et al. 1991a, 1992),
xanthan (Williams et al. 1991a, b), gellan (Miyoshi et al. 1995; Robinson
et al. 1991), agar (Cooke et al. 1996), pectin, and gelatin. Polysaccharides
are widely used for their gelling and thickening properties in the food
industry. In presence of a cation (for example, potassium K+), a solution of kappa-carrageenan gives an aggregate structure during heating.
The temperature of transformation and the reversibility of the reaction
38
39
0.14
0.1
A-PASTEURIZED
WHOLE EGG
T=21.7C
exo
HF / mW g1
0.12
0.08
T=14.7C
0.06
0.04
T=9.6C
0.02
0
0
4
6
Time / days
10
0.12
0.08
T=24.6C
B-PASTEURIZED
WHOLE MILK
exo
HF / mW g1
0.1
0.06
T=14.7C
T=19.6C
0.04
0.02
0
0
Time / days
0.14
0.1
0.08
C-FRESH CARROTS
T=24.6C
exo
HF / mW g1
0.12
T=19.6C
0.06
T=14.6C
0.04
0.02
0
0
0.5
1.5
2.5
Time / days
Figure 2.12. Isothermal traces at different temperatures for pasteurized whole egg,
pasteurized whole milk, and fresh carrots. From Franzetti et al. (1995).
40
0
0.10
10
15
20
25
30
35
40
0.08
0.06
0.04
0.02
0.00
40
30
T / C
20
10
0
10
15
20
25
time / hours
30
35
0
40
41
2. Flow mixing: The two components A and B at a given flow rate are
pumped and mixed in the vessel. The concentration of the mixture
can be adjusted by modifying the flow rate of A or B.
Dissolution, solubility
In food industry, solid-liquid and liquid-liquid interactions are often
encountered, such as dissolution of powder (sugar, salt) and solubility
of proteins, lipids, and fibers. For such studies, the batch-mixing vessel
is ideal because it provides information relevant to the start of the
reaction and the corresponding kinetics.
Neutralization
The batch-mixing vessel is also convenient for the investigation of any
reaction occurring during a food process, such as neutralization of
edible oils by soda. Raw edible oils contain free fatty acids that have
to be neutralized before being used. The amount of soda necessary for
neutralization has to be adjusted based on the acidity of the oil. The
simulation of the operation was performed on a microcalorimeter using
edible oil with variable acidities (Figure 2.14).
Binding
A mixing calorimeter is useful to investigate the impact of the weak
nonspecific physical interactions of the food biopolymers (proteins,
heat flow
4 mW
exo
1
2 peanut
3%
3 peanut
0,85 %
time (mn)
0
10
15
20
25
30
42
50W
0.5
1.0
1.5
2.0
polysaccharides) with each other and with the major low-molecularweight ingredients of the multicomponent food colloids (sugars,
mineral salts, small-molecule surfactants). The structure formation in
the bulk aqueous phase and at the interfaces of colloidal systems, as
well as the functional properties, depend on these weak interactions
(Semenova 2007).
Enzymatic reactions
The example for the enzymatic reaction of transformation of maltose
using glucoamylase illustrates the two modes of mixing. In batch
mode, 30 mg of maltose powder is mixed with 20 l of glucoamylase.
The corresponding exothermic effect indicates the transformation of
maltose for a given concentration of enzyme (Figure 2.15). In flow
mode, maltose (1% in H2O) is circulated in both tubes of the mixing
vessel to establish the baseline of the test. Then, glucoamylase (1% in
H2O) is introduced. A corresponding exothermic deviation measures
the efficiency of the enzyme for maltose transformation at a given
temperature and concentration. When the enzyme flow is replaced by
maltose, the calorimetric signal is returned to the baseline (Figure
2.16). This mixing mode is flexible because various enzyme concentrations can be tested consecutively.
43
HEAT FLOW
20W
maltose
+
maltose
enzyme
+
maltose
0
maltose
+
maltose
10
20
30
Time (min)
44
nomena in food to improve and develop new technologies. High hydrostatic pressure can cause denaturation of proteins, solidification of lipids,
inactivation of microorganisms, and destabilization of biomembranes.
The calorimetric technique is used to investigate the transformations
induced by pressure. Because a commercial high-pressure calorimeter is
not available (see Chapter 3) for high hydrostatic pressure applications,
the sample is processed in a specific pressure device, outside the calorimeter. This application was used to investigate the protein denaturation
in egg white after high-pressure processing (Andrassy et al. 2006) to
detect the structural changes in milk protein beta-lactoglobulin induced
by combined effects of pressure and temperature (Tedford and Schaschke
2000; Kolakowski et al. 2001). High-pressure induced modifications of
soy protein in soy milk, studied using microcalorimetry, showed that
denaturation of b-conglycinin and glycinin occurred at 300 MPa and
400 MPa, respectively, as judged by the absence of endothermic peaks in
thermograms of pressure-treated samples (Zhanga et al. 2005).
Microcalorimetry in the scanning mode also was used to evaluate
the relative high hydrostatic pressure resistance of bacterial strains
from Staphylococcus aureus (Figure 2.17) and Escherichia coli in
A
Heat Flow
0.5 mW
20
40
60
80
Temperature (C)
100
120
45
vivo. The total apparent enthalpy change and thermal stability were
two calorimetric parameters used to compare bacterial strains of
untreated control and pressure-treated bacteria (Alpas et al. 2003).
Pressure can be applied on the sample inside the calorimetric vessel.
A maximum gas pressure of 100 MPa was reported by Le Parlour et
al. (2004) to be used for investigation of polymorphic changes in fatty
compounds, the modification of the glass transition temperature for
food with amorphous phase (sugar, starch, dough, frozen foods), or
oxidative stability of oils. Supercritical extraction process using CO2
as a fluid was simulated in a high-pressure calorimeter with an adapted
vessel (Stassi and Schiraldi 1994). Such a setup allows investigation
either at constant pressure and at variable supercritical CO2 flow rate,
or as batch system with a variable pressure. The latter allows the monitoring of the solubility-pressure relationship.
Calorimetry under Controlled Relative Humidity
Many food processes (extrusion, baking, drying, milling) may generate
amorphous compounds. Their stability upon storage, especially in a
humid atmosphere, has to be controlled. Water induces plasticization and
leads to depression of the glass transition temperature, causing significant changes in the physicochemical and crystallization properties of the
food components containing an amorphous phase. Scanning calorimetry
is recognized as an efficient technique for measurement of the glass transition temperature of hydrated products (Bizot et al. 1997; Borde et al.
2002). Scanning calorimetry was also used to monitor crystallization of
lactose in humidified powders, because lactose crystallisation and
Maillard reaction are two major modifications occurring in milk and
whey powders during processing and storage (Morgan et al. 2005).
Typically, samples are equilibrated outside the calorimeter, and the influence of water content or water activity are investigated using calorimetry.
However, the sample can be prepared with a defined water content, and
thermal properties can be measured inside a special vessel by combining
DSC and humidity generator (Le Parlour and Mathonat 2003).
Conclusion
As in other industrial domains, the food industry is more and more
interested in applying new techniques of investigations to improve the
46
References
Alpas H., Lee J., Bozoglu F., and Kaletunc G. 2003. Evaluation of high hydrostatic
pressure sensitivity of Staphylococcus aureus and Escherichia coli O157:H7 by
DSC. Int J Food Microbiol, 87(3):229237.
Andlid T., Blomberg L., Gustafsson L., and Blomberg A. 1999. Characterization
of Saccharomyces cerevisiae CBS 7764 isolated from rainbow trout intestine.
Syst Appl Microbiol, 22(1):145155.
Andrassy E., Farkas J., Seregely Z., Dalmadi I., Tuboly E., and Lebovics V.
2006. Changes of hen eggs and their components caused by non-pasteurizing
treatments. II. Some non-microbiological effects of gamma irradiation or hydrostatic pressure processing on liquid egg white and egg yolk. Acta Alimentaria,
35(3):305318.
Bacon J.R., Noel T.R., and Wright D.J. 1989. Studies on the thermal behaviour of
pea (Pisum sativum) vicilin. J Sci Food Agri, 49:335345.
Barone G., Capasso S., Del Vecchio P., De Sena C., Giancola C., Graziano G. 1995.
Thermal denaturation of bovine serum albumin and its oligomers and derivatives
PH dependence. J Thermal Anal, 45:12551264.
Barone G., Giancola C., and Verdoliva A. 1992. DSC studies on the denaturation and
aggregation of serum albumins. Thermochim Acta, 199:197205.
Berland S., Relkin P., and Launay B. 2003. Calorimetric and rheological properties
of wheat flour suspensions and doughs. Effects of wheat types and milling procedure. J Thermal Anal Calorim, 71:311320.
Berridge N.J., Cousins C.M., and Cliffe A.J. 1974. Microcalorimetry applied to
certain species of bacteria growing in sterilized separated milk. J Dairy Res,
41:203.
Bizot H., Le Bail P., Leroux B., Davy J., Roger P., and Buleon A. 1997. Calorimetric
evaluation of the glass transition in hydrated, linear and branched polyanhydroglucose compounds. Carbohydrate Polymers, 32:3350.
Borde B., Bizot H., Vigier G., and Buleon A. 2002. Calorimetric analysis of the
structural relaxation in partially hydrated amorphous polysaccharides. II.
Phenomenological study of physical aging. Carbohydrate Polymers, 48(1):8396.
47
48
49
Chapter 3
High-Pressure Differential
Scanning Calorimetry
Gnther W.H. Hhne and Gnl Kaletun
Introduction
Construction of the High-Pressure DSC
Calibration of the High-Pressure DSC
Temperature Calibration Procedure
Heat Calibration Procedure
Applications of the High-Pressure DSC
Conclusion
References
51
53
57
58
61
63
63
64
Introduction
Pressure is an essential variable in physical chemistry. Measurements
at different pressures are therefore of great importance from the thermodynamic perspective. The change of pressure provides increased
insight into the thermodynamic behavior of materials. The wider the
pressure region, the better description of material response to pressure
is obtained, which enables one to develop predictive capability. Higher
pressure is often used during production and processing of materials,
and the change of properties that occurs with pressure is therefore of
great interest for optimization of processing conditions. In particular,
the latent heat of reactions, phase and conformational transitions, and
their pressure dependence are valuable information for quantitative
51
52
53
54
Figure 3.1. High-pressure DSC setup. (a) DSC; (b) spindle pump; (c) autoclave
(Ledru et al. 2006 in accordance with Blankenhorn and Hhne 1991).
Figure 3.2. High-pressure DSC: Autoclave unit (Ledru 2006 in accordance with
Blankenhorn and Hhne 1991).
Figure 3.3. High-pressure DSC: Ceramic housing with the silver furnace inside
(Ledru et al. 2006 in accordance with Blankenhorn and Hhne 1991).
55
56
Figure 3.4. High-pressure DSC: Silver furnace with sample pan inside (Ledru et al.
2006 in accordance with Blankenhorn and Hhne 1991).
were electrically isolated with special ceramic glue. They were provided with platinum wire windings for both the heaters and sensors to
match the resistance of the original DSC cell and with leads through
the autoclave closing caps to the calorimeter control system.
The purpose of locating the two furnaces within the ceramic housings is to isolate them, preventing cross talk between the sample and
reference and minimizing the disturbance of the heat flow signal caused
by convection currents in the oil. These housings contain holes for oil
entry and escape during the filling/emptying and pressure change
process, and they have screw caps (Figure 3.3) to allow their volume
to be adjusted in order to balance the two furnaces to obtain a flat
baseline.
Using a branched silicone oil of approximately 100 mPa/s (Wacker
AS 100, Wacker-Chemie GmbH, Germany), the HP-DSC may be
operated in a temperature range from 20 C to 300 C at pressures from
ambient to 500 MPa and with various heating and cooling rates (from
0.5 K.min1 to 20 K.min1). The actual pressure value within the autoclave is measured using a pressure transducer close to the reference
cell, which can be connected to a data logger for pressure recording.
The sample must, of course, be encapsulated to avoid any contact
with the pressurizing medium. This is not an easy matter, as the encap-
57
sulation must be oil-tight on the one hand and free from air bubbles
and empty space on the other. The latter would lead to large deformation of the sample container when the pressure rises; thus, the container
possibly would not remain oil-tight and the measurement would be
faulty. There are several possible ways to solve the problem. One is to
prepare the sample to fit exactly between two aluminum crucibles that
then are welded together with a proper press. Another possibility is to
put the samples into crucibles of a plastic metal such as indium or lead
and close the crucible hermetically.
In operation, the furnaces are placed within the autoclave with their
axes horizontal. The autoclave has a lid (Figure 3.4) that screws down
onto the crucible, ensuring that it is firmly located within the silver
furnace and that good thermal contact is made.
The HP-DSC constructed this way had the following properties:
pressure and temperature ranges from 0.1 to 500 MPa and from ambient
to 600 K (330 C), respectively, and thermal noise 50100 W peak to
peak (much larger than in normal DSCs because of the oil convection).
The detection limit for transitions (peak area) is about 5 mJ (i.e., 1 J g1).
Compared to common DSCs, the baseline repeatability of the HP-DSC
is poor (23 mW) because of unavoidable small differences in oil
volume between the sample and reference cell when a new sample is
remounted. This makes it impossible to determine heat capacities of a
sample with the usual method, which is to subtract an empty pan run
from the sample run. The change in baseline after remounting a new
sample is often larger than the expected difference of the heat flow rate
between the two runs. This unavoidable effect is a serious disadvantage
of the high-pressure DSC.
58
(3.1)
Normally, the correction Tcorr depends on the heating rate, too. This
influence is, however, of minor importance if the same heating rate is
always used for all measurements.
59
(3.2)
In
= 429.7485 K is the fixed melting point of the ITS-90 for
where Tfus,0
indium at normal pressure (Preston-Thomas 1990) and p is the
pressure. The given standard deviation defines the limits of this best
value estimation and leads to a maximum uncertainty of 1.5 K at
500 MPa.
For tin, to our knowledge two reliable publications exist for the
pressure dependence of the melting point (McDaniel, Babb, and Scott
1962; Sandrock 1982b). The best value approximation for these data
reads:
Sn
Sn
Tfus
K ] + [(0.0324 0.0025) K MPa 1 ][ p MPa ]
( p ) K = [Tfus,0
(3.3)
[(1.45 4.86)106 K MPa 2 ][ p MPa ]2
Sn
where Tfus,0
= 505.078 K , the fixed melting point of the ITS-90 for tin
at normal pressure (Preston-Thomas, 1990) and p the pressure. The
standard deviations were defined from this best value estimation and,
according to the literature data, lead to a maximum uncertainty of 2.5 K
at 500 MPa.
For the pressure dependence of the melting point of lead, to our
knowledge only one publication exists (McDaniel, Babb, and Scott
1962). Any best value estimation, therefore, cannot be performed for
the pressure dependence of its melting point, and the uncertainty is not
known. Consequently, lead cannot be used for calibration of the HPDSC. Because we have two reference substances only, we have to
restrict ourselves to a linear approximation of the temperature dependence of the correction function.
60
15
Tcoor/K
10
0.1 MPa
50 MPa
100 MPa
150 MPa
200 MPa
250 MPa
300 MPa
350 MPa
400 MPa
450 MPa
500 MPa
5
330
380
430
480
530
Measured temperature/K
61
(3.4)
where Tcorr(p,T) is the correction obtained from the calibration procedure and Tsr is an additional correction that corresponds to the
difference between the reference and sample temperature sensor. This
difference is often temperature and pressure independent and can be
taken as a constant value compared with the overall uncertainty of the
corrected temperatures, which are within the range of 14 K (depending on temperature and pressure) for our calorimeter.
(3.5)
62
In
fus H ref
( p)
In
fus H meas ( p )
(3.6)
For the pressure dependence of the latter, the best value estimation
from Preston-Thomas (1990) is used:
In
fus H ref
J g 1 = fus H 0In, ref J g 1 + [(3.3 2 )10 3 g 1 MPa 1 ][ p MPa ]
[(2.6 2)107 J g1 MPa 2 ][ p MPa ]2
(3.7)
where fus H 0Inref = 28.62 0.11J g 1 is taken as the best value for the
heat of fusion of indium at normal pressure and p the pressure. This
best value estimation includes an uncertainty of 0.1 J g1 at ambient
pressure increasing to 1.1 J g1 at 500 MPa for the best value of the heat
of fusion of indium. A best value estimation of the pressure dependence of the melting enthalpy for tin has not been reported; therefore,
we did not use the melting enthalpy of tin for calibration purposes. The
total uncertainty of a heat measurement with the HP-DSC is the sum
of the uncertainties of the reference material and the respective measurement, which is about 0.92 J g1. In Figure 3.6 Rcorr(p) resulting
from such a heat calibration is given as an example.
63
Conclusion
High-pressure differential scanning calorimeters are not available
commercially. They are, however, very valuable instruments for obtaining essential thermodynamic data, including food applications. If highpressure measurements are needed, one must build such an instrument
oneself. This is not an easy task and needs experience. As described
above, the high-pressure power compensated DSC has been built
and works well. The advantages of this type of high-pressure calorimeter are as follows:
64
References
Alpas H., Lee J., Bozoglu F., and Kaletun G. 2003. Differential scanning calorimetry
of pressure-resistant and pressure-sensitive strains of Staphylococcus aureus and
Escherichia coli O157 : H7. International J Food Microbiol, 87:229237.
Arntz H. 1980. New high pressure low temperature differential scanning calorimeter.
Rev Sci Instrum, 51(7):965967.
Blankenhorn K. and Hhne G.W.H. 1991. Design, specifications and application of
a high pressure DSC cell. Thermochim Acta, 187:219224.
Eichler A. and Gey W. 1979. Method for the determination of the specific heat of
metals at low temperatures under high pressures. Rev Sci Instrum,
50(11):14451452.
Hhne G.W.H. 1999. High pressure differential scanning calorimetry on polymers.
Thermochim Acta, 332:115123.
Hhne G.W.H. and Blankenhorn K. 1994. High pressure DSC investigations
on n-alkanes, n-alkane mixtures and polyethylene. Thermochim Acta, 238:
351370.
Hhne G.W.H., Dollhopf W., Blankenhorn K., and Mayr P.U. 1996. On the pressure
dependence of the heat of fusion and melting temperature of indium. Thermochim
Acta, 273:1724.
Hhne G.W.H., Hemminger W., and Flammersheim H.J. 2003. Differential Scanning
Calorimetry, 2nd revised and enlarged ed. Springer-Verlag: Berlin.
65
Hhne G.W.H., Rastogi S., and Wunderlich B. 2000. High pressure differential scanning calorimetry of poly(4-methyl-pentene-1). Polymer, 41:88698878.
Hhne G.W.H., Schawe J.E.K., and Shulgin A.I. 1997. The phase transition behaviour
of linear polyethylenes at high pressure. Thermochim Acta, 296:110.
Ingram M.D., Imrie C.T., Ledru J., and Hutchinson J.M. 2008. Unified approach
to ion transport and structural relaxation in amorphous polymers and glasses.
J Phys Chem, 112:859866.
Kaletun G., Lee J., Alpas H., and Bozoglu F. 2004. Evaluation of structural changes
induced by high hydrostatic pressure in Leuconostoc mesenteroides. Appl Environ
Microbiol, 70:11161122.
Kamphausen M. 1975. New differential scanning high pressure microcalorimeter. Rev
Sci Instrum, 46(6):668669.
Ledru J., Imrie C.T., Hutchinson J.M., and Hhne G.W.H. 2006. High pressure
differential scanning calorimetry: Aspects of calibration. Thermochim Acta,
446:6672.
Lee J. and Kaletun G. 2002a. Evaluation by differential scanning calorimetry of the
heat inactivation of Escherichia coli and Lactobacillus plantarum. Appl Environ
Microbiol, 68:53795386.
Lee J. and Kaletun G. 2002b. Calorimetric determination of inactivation parameters
of microorganisms. J Appl Microbiol, 93:178189.
Lee J. and Kaletun G. 2005. Evaluation by differential scanning calorimetry of
the effect of acid, ethanol, and NaCl on Escherichia coli. J Food Prot, 68:
487493.
Mackey B.M., Miles C.A., Parsons S.E., and Seymour D.A. 1991. Thermal denaturation of whole cells and cell components of Escherichia coli examined by differential scanning calorimetry. J Gen Microbiol, 137(10):23612374.
McDaniel M.L., Babb Jr. S.E., and Scott G.J. 1962. Melting curves of five metals
under high pressure. J Chem Phys, 37(4):822828.
Mellander B.E., Baranowski B., Lund A. 1981. Transition enthalpies of silver iodide
in the high-pressure region determined by DSC. Phys Rev, 23(8):37703773.
Miles C.A., Mackey B.M., and Parsons S.E. 1986. Differential scanning calorimetry
of bacteria. J Gen Microbiol, 132(4):939952.
Nakafuku C. and Sugiuchi T. 1996. Effect of pressure on the phase diagram of binary
mixtures of n-alkanes. Polymer, 34(23):49454952.
Niven G.W., Miles C.A., and Mackey B.M. 1999. The effects of hydrostatic pressure
on ribosome conformation in Escherichia coli: An in vivo study using differential
scanning calorimetry. Microbiol, 145:419425.
Preston-Thomas H. 1990. The international temperature scale of 1990 (ITS90).
Metrologia, 27:310.
Randzio S.L. 1983. A pressure-scanning calorimeter. J Physics E Sci Instrum,
16:691694.
Rastogi S., Hhne G.W.H., and Keller A. 1999. Unusual pressure-induced phase
behavior in crystalline poly(4-methylpentene-1): Calorimetric and spectroscopic
results and further implications. Macromolecules, 32:88978909.
66
Sandrock R. 1982a. High-pressure high-temperature differential scanning calorimeter. Rev Sci Instrum, 53(7):10791081.
Sandrock R. 1982b. Differentialkalorimetrie (DSC) bei hohen Drcken; Phasenverhalten
und Umwandlungsenthalpien sowie daraus abgeleitete thermodynamische Gren
von Polyethylen und Diamantan bis 6000 bar und 600 K. Dissertation, RuhrUniversitt: Bochum.
Schmidt C., Rittmeier-Kettner M., Becker H., Ellert J., Krombach R., and Schneider
G.M. 1994. Differential thermal analysis (DTA) and differential scanning calorimetry (DSC) at high pressures. Experimental techniques and selected results.
Thermochim Acta, 238:321336.
Schneider G.M. 1985. Recent developments of microcalorimetry at high pressures.
Thermochim Acta, 88:159168.
Shulgin A.I. and Godovsky Y. 1992. DTA measurements on polymers under high
pressurepolyethylene and poly(diethylsiloxane). J Thermal Anal, 38:
12431250.
Szab J., Luft G., and Steiner R. 1969. Anwendung der Differentialthermoanalyse
zu reaktioiiskinetischen Untersuchungen von Hochdruckreaktionen. Chemie Ing
Technik, 41:1007101.
Zhu S., Bulut S., Le Bail A., and Ramaswamy H.S. 2004a. High pressure differential
scanning calorimetry (DSC): Equipment and technique validation using water-ice
phase-transition data. J Food Process Eng, 27:359376.
Zhu S., Ramaswamy H.S., and Le Bail A. 2004b. High-pressure differential scanning
calorimetry: Evaluation of phase transitions in pork muscle at high pressures.
J Food Proc Eng, 27: 377.
Chapter 4
Calorimetry of Proteins in Dilute Solution
G. Eric Plum
Introduction
Differential Scanning Calorimetry
Isothermal Titration Calorimetry (ITC)
Conclusion
References
67
68
77
83
84
Introduction
As the food industry begins to take advantage of recent developments
in protein chemistry by introducing enzymes and structural proteins
into modern food materials and their processing, detailed understanding of protein chemical and physical properties becomes increasingly
important. Development of a predictive understanding of the energetics-structure-function relationships will be required to fully exploit the
possibilities presented to engineer proteins with novel substrate specificity or enhanced physical properties, including thermal stability, pH,
and ionic strength optima.
Calorimetry provides several valuable tools for the characterization
of the thermal properties of proteins and their interactions with
other macromolecules and small-molecule affectors. The objective
of this chapter is to introduce and summarize the methods of modern
67
68
69
70
The sample must be prepared so that the only the difference between
the sample and reference solutions is the presence of the macromolecule of interest. This is generally best accomplished by exhaustive
equilibrium dialysis of the sample solution (dialysand) against the
buffer solution (dialysate) and subsequent use of the dialysate as the
reference solution. Even with the greatest care, the solutions in the
sample and reference cells cannot be matched exactly. The macromolecule under study displaces solution that contains solvent, buffers,
salts, etc., which have temperature-dependent heat capacities. These
effects, coupled with imperfect matching of the calorimeter cells, mean
that a small but nontrivial baseline correction must be made to separate
the contribution of the macromolecule order-disorder transition from
the solution displacement and instrumental effects. Small errors in
sample preparation and baseline correction can lead to large errors in
the calculated thermodynamic parameters. Some attempts have been
made to standardize DSC experimental and analysis procedures to
reduce interlaboratory variability (Hinz and Schwarz 2001), but
care should be exercised when comparing data from different
laboratories.
Basic equations
With modern instrumentation, complex multidomain DSC thermograms can be resolved into individual transitions (Vlker et al. 1999).
However, for this discussion it is assumed that the thermogram is of a
single transition, which is typically observed for single domain globular proteins in solution. Thus, only monomolecular unfolding processes, those involving only a single polypeptide chain, are considered.
Methods to address multisubunit proteins exist but are beyond the
scope of this discussion.
Figure 4.1 shows a simulated DSC thermogram of the temperatureinduced unfolding of a small globular protein. Differential scanning
calorimetry data are analyzed using a set of standard thermodynamic
relations (Privalov and Potekhin 1986). All thermodynamic relations
herein are at constant pressure. The DSC curve is analyzed in terms of
the relationships between the measured heat capacity and the thermodynamic parameters of interest.
The excess heat capacity, C pex, is the difference in heat capacity
between the sample solution containing the macromolecule of interest
relative to the reference solution. Tm is the temperature at the mid-
71
12,000
Tm
Cp cal mol1 K1
10,000
8,000
H(Tm)
6,000
4,000
2,000
Cp
0
300
325
350
375
400
Temperature (K)
Figure 4.1. Simulated DSC thermogram of a small globular protein. The excess heat
capacity versus temperature curve is calculated using Tm = 350, H(Tm) = 100 kcal/
mol, and Cp = 1.5 kcal/mol K.
point of the transition; that is, the temperature at which the concentrations of the folded and unfolded forms of the protein are equal. The
maximum of the C pex versus temperature curve will correspond to Tm
only when the unfolding unit is monomolecular and the difference in
heat capacity between the folded and unfolded forms of the protein is
negligible. The enthalpy change at Tm is determined from integration
of the DSC curve
H (Tm ) = C pex dT
(4.1)
where the integration covers the entire temperature range of the denaturation transition. The free energy change at temperature T, which is
a measure of the proteins stability at that temperature, depends on the
enthalpy and entropy changes.
G (T ) = H (T ) T S (T )
(4.2)
72
H (T ) = H (Tm ) + T C p dT
(4.3)
C p
dT
T
(4.4)
S (T ) = S (Tm ) + T
H (Tm )
Tm
(4.5)
(4.6)
T
S (T ) S (Tm ) C p ln m
T
(4.7)
Combining Equations 4.2, 4.6, and 4.7 permits calculation of the free
energy change at any temperature from the three parameters, H(Tm),
Tm, and Cp, obtainable from the DSC curve.
G (T )
Tm T
T
H (Tm ) (Tm T ) C p + T C p ln m
T
Tm
(4.8)
Using these expressions one can predict the stability of the protein and
its thermodynamic origins at any temperature.
The vant Hoff enthalpy change
While the enthalpy change measured by DSC does not depend on a
model, comparison with models can provide insight into the nature of
the order-disorder transition. Consider an equilibrium constant, Keq.
K eq =
[ unfolded ]
[folded ]
(4.9)
73
Because the shape of the DSC curve reflects the change in the equilibrium constant as a function of temperature, the vant Hoff model
can be applied to DSC data as an alternate means of enthalpy change
determination. The vant Hoff model is based on the temperature
dependence of the dimensionless equilibrium constant.
ln K eq
ln K eq
H vH = RT 2
= R
T
(1 / T )
(4.10)
Note that the units of the vant Hoff enthalpy, HvH, are defined by the
units of the constant R.
Comparison of the model independent calorimetrically determined
H(Tm) value with the HvH value assesses the validity of the assumptions employed in the derivation of the vant Hoff relation. Specifically,
it is assumed that the transition from the ordered, low temperature form
to the disordered, high temperature form passes through no thermodynamically significant intermediate states (two-state assumption); that
is, there is no partial unfolding of the protein in the denaturation
pathway. The HvH reports the enthalpy change associated with disruption of a single cooperative unit, the fraction of the protein that acts as
a single thermodynamic unit.
There are several methods used to extract HvH from the shape of
equilibrium denaturation curves, which are frequently based on indirect observations such as temperature-dependent spectroscopic measurements (Marky and Breslauer 1987). The equation below comes
most directly from the DSC curve. Because the equation, as written
here, does not account for changes in heat capacity, a DSC curve from
which the contribution of Cp has been subtracted should be used.
2
H vH = 4 RTmax
C pmax
H (Tmax )
(4.11)
74
(4.12)
where cp and cnp are empirical coefficients, Ap and Anp are the differences between the folded and unfolded forms of the protein in polar and
nonpolar surface area in contact with solvent. Although the structure of
the folded form of most proteins is stable and the relevant surface areas
readily calculated, the fluctuating structure of the unfolded form is
poorly defined, and thus different methods of calculating the surface
areas of the unfolded form lead to different empirical equations.
75
Cold denaturation
Due to the sign and magnitude of the heat capacity change relative to
the changes in enthalpy and entropy observed for globular proteins, at
some temperature a maximum in the free energy change is observed
(Privalov 1990). This further implies that, in addition to the high temperature (Tm) at which G = 0, there is a low temperature that satisfies
the same condition. In most cases, this implied cold denaturation temperature is below the freezing point of water; however, there are
examples of cold denaturation observable within the aqueous liquid
temperature range accessible experimentally with fixed cell instrumentation, wherein the cells must be filled completely (Privalov 1990). The
cold denaturation phenomenon may be particularly important in freezing and lyophilization processes.
DSC data can quantify high-affinity binding
Binding equilibria between a protein and a small molecule effector,
such as a cofactor or drug, or a second protein subunit, can alter
the proteins denaturation temperature. If the second molecule binds
more tightly to the folded form of the protein than to the unfolded
form, the denaturation will shift to higher temperature. Conversely,
preferential binding to the unfolded form shifts the denaturation to
lower temperature. DSC thermograms are particularly well suited
to measure very tight binding based on the observed bindinginduced changes in the heat capacity versus temperature profiles.
Brandts and Lin (1990) present methods and models for analysis of
the shapes of DSC thermograms to quantify binding affinities for
protein-small molecule and protein-protein interactions up to 1040 M1,
whereas most methods cannot quantify binding constants greater than
1010 M1.
Assumptions
The above described analysis of DSC thermograms in terms of equilibrium thermodynamic parameters depends on a number of assumptions about the system and the design of the experiment. It is
important to evaluate the validity of the assumptions to assure the
quality of the derived thermodynamic data and because deviations
from the assumed behavior can provide insight into the protein unfolding process.
76
U
mD
(4.13)
77
aggregation resulting in precipitation is often apparent by wild fluctuation in the high temperature baseline.
Concentration and purity
The enthalpy and heat capacity measurements conducted using DSC
are based on the amount of material thought to be in the sample.
Whether the objective is molar or specific values, the measured thermal
effects are divided by the quantity of the analyte protein. Therefore,
accurate DSC results depend on accurate determination of the amount
of analyte protein present, as well as its purity. The best means of
determining concentration of the analyte will vary with the properties
of the particular protein under study.
It is generally assumed that the concentrations of all components in
the solution are sufficiently low that all activity coefficients may be
satisfactorily approximated by unity. While this assumption may not
be correct, in most cases there is little alternative.
Selection of hydrogen ion buffer
The selection of hydrogen ion buffer is critically important in DSC
experiments. Buffers with high heats of ionization lead to temperature
dependent changes in the pH of the solution. Thus, any pH dependent
changes in the protein will be superimposed onto the temperature
dependent changes, which the DSC experiment is designed to measure.
Unfortunately, some of the most widely used buffers for general biological macromolecule studies exhibit high ionization heats. A particularly egregious example is tris buffer, although most buffers carrying
amine groups are problematic.
A second important issue in buffer selection for DSC studies is the
buffers propensity for metal ion chelation. Because proteins frequently
carry anionic functionalities on their surfaces or require metal ion
cofactors, interactions with metal ions are often important in stabilizing
their structures. Competition for metal ion binding between the protein
and the buffer can compromise the DSC experiment.
Isothermal Titration Calorimetry (ITC)
Information content
Most biological processes involve one or more binding events. The
types of binding reactions are varied and include, but are not limited
78
79
(4.14)
Figure 4.2. Isothermal titration calorimetry of binding of two inhibitors to glucosidase (Zechel et al. 2003). (a) The raw data (upper panel) and the integrated
data with fitted curve (lower panel) for an inhibitor with K = 1.25 105 M1. (b) An
inhibitor with K = 1.5 107 M1. Reprinted with permission from J. Am. Chem. Soc.
2003, 125, 1431323. Copyright 2003 American Chemical Society.
80
81
(4.15)
1 1 +
2
[ M ] K [ M ]
d [ LM ] 1
(4.16)
= +
1
2
2
2 [L] 2
d [L]
1
1
[L]
[ M ] 2 [ M ] 1 [ M ] K + 1 + [ M ] K
The ITC curve can then be fitted by standard nonlinear least
squares techniques for H and K. The free energy and entropy changes
are determined by the standard relations G(T) = RTlnK and
H (T ) G (T )
. A series of titrations as a function of
T
temperature provide the heat capacity change associated with the
binding reaction via the temperature dependence of the enthalpy
dH
.
change, C p =
dT
Only for the most simple models can a closed form expression be
written relating the change in heat accompanying introduction of an
aliquot of titrant to the parameters of the binding reaction. More elaborate models can be evaluated numerically. Software provided with the
instruments can analyze ITC data in terms of several different models,
including the single binding site case considered here, multiple identical binding sites that may or may not interact, and multiple classes of
distinct binding sites.
Many, if not most, binding events involving macromolecules and
small molecules or other macromolecules are coupled to changes in
ionization state of one or more charged groups. Thus, the measured
S (T ) =
82
heat associated with the binding event includes the net ionization heats
of the groups for which protonation or deprotonation occurs as well as
compensating effects from the included hydrogen ion buffer. It is
therefore advisable to measure the association heat in two or more
hydrogen ion buffers, which differ in ionization heat, at identical pH
values (Baker and Murphy 1996). The plot of the measured association
heat as a function of the buffer ionization heat permits determination
of the intrinsic heat of binding and the net number of protons taken up
or released upon binding. The intercept at zero buffer ionization heat
corresponds to the intrinsic enthalpy of binding; the slope corresponds
to the net change in protonation.
Due to the strong model dependence of the ITC method, challenges
arise that are not part of the analysis of DSC data. Most ITC-binding
isotherms comprise a small number of features: specifically, the intercept on the enthalpy axis and one or two inflection points. Thus, the
number of fitting parameters that the data can support is limited. Many
processes of interest involve competing equilibria with multiple binding
species and binding sites that require elaborate multiparameter models
to describe. Caution, and appropriate statistical tests, should be applied
to ensure that the data can support the applied model. It is relatively
easy to devise a model that cannot be adequately addressed by ITC
isotherms alone; however, if complementary data are available to fix
some of the fitting parameters, such models may become tenable.
Range of applicability
The classical ITC experiment described here is limited to binding
affinities and concentrations that produce a titration curve with a shape
that can be fit accurately by nonlinear least squares methods. A rough
estimate of this range defined by the macromolecule concentration and
the equilibrium association constant is 1 < [M]K < 1000 (Wiseman et
al. 1989), with 10 < [M]K < 100 being preferred. Note that while the
macromolecule concentration could always be adjusted to place [M]K
in range, the heat produced by the binding reaction must be detectable.
This places an effective limit on the range of binding affinities accessible to direct ITC measurement.
Figure 4.3 shows how the shape of the ITC titration curve varies
with [M]K. As [M]K , the ITC curve becomes a step function with
essentially all the L injected binding to M, until all of the available
binding sites are exhausted. So dQ/d[L] = H for [L]/[M] < n and
83
Figure 4.3. Dependence of the shape of the ITC titration curve on [M]K for the reaction nL + M MLn.
Conclusion
Modern ultrasensitive calorimetry provides powerful tools for understanding the stability of proteins in solution, the forces that maintain
their folded structures, and their interactions with other macromolecules and small molecules.
Differential scanning calorimetry quantifies the thermal (Tm) and
thermodynamic (G) stabilities of the protein. The thermodynamic
origins of the stability (H, S, and Cp) can be interpreted to dissect
the forces maintaining the folded structure and how they depend on
84
the structure of the protein and the solution conditions. The modelindependent values derived directly from the DSC thermogram can
be compared to model-dependent values derived from the shape of
the DSC curves to gain insight into the unfolding mechanism and
aggregation.
Isothermal titration calorimetry provides a means to directly measure
the heat of interaction (H) of a protein with another macromolecule
or with small molecules. Application of models for the association
reaction provides detailed thermodynamic characterization (K, G, S,
and Cp) of the binding process.
Taken together, the techniques of modern solution calorimetry
provide a predictive understanding of the stability of a protein and its
interactions with other molecules as a function of temperature and
solution conditions.
References
Baker B.M. and Murphy K.P. 1996. Evaluation of linked protonation effects in
protein binding reactions using isothermal titration calorimtery. Biophys J,
71:204955.
Brandts J.F. and Lin L. 1990. Study of strong to ultra tight protein interactions using
differential scanning calorimetry. Biochemistry, 29:692740.
Doyle M.L., Louie G., Dal Monte P.R., and Sokoloski T.D. 1995. Tight binding
affinities determined from thermodynamic linkage to protons by titration calorimetry. Methods Enzymol, 259:18394.
Hinz H.J. and Schwarz F.P. 2001. Measurement and analysis of results obtained on
biological substances with differential scanning calorimetry. Pure Appl Chem,
73(4):74559.
Marky L.A. and Breslauer K.J. 1987. Calculating thermodynamic data for transitions
of any molecularity from equilibrium melting curves. Biopolymers, 26:160120.
Mittag T. and Forman-Kay J.D. 2007. Atomic-level characterization of disordered
protein ensembles. Curr Opin Struct Biol, 17:314.
Plotnikov V.V., Brandts J.M., Lin L., and Brandts J.F. 1997. A new ultrasensitive
scanning calorimeter. Anal Biochem, 250:237244.
Prabhu N.V. and Sharp K.A. 2005. Heat capacity in proteins. Annu Rev Phys Chem,
56:52148.
Privalov G., Kavina V., Freire E., and Privalov P.L. 1995. Precise scanning calorimeter for studying thermal properties of biological macromolecules in dilute solution. Anal Biochem, 232:7985.
Privalov P.L. 1990. Cold denaturation of proteins. Crit Rev Biochem Mol Biol,
25(4):281305.
Privalov P.L. and Potekhin S.A. 1986. Scanning microcalorimetry in studying tem-
85
Chapter 5
Thermal Analysis of Denaturation and
Aggregation of Proteins and Protein
Interactions in a Real Food System
Valerji Y. Grinberg, Tatiana V. Burova, and Vladimir B. Tolstoguzov
Introduction
Effects of pH on Thermal Denaturation of Food Proteins
Effects of Salts on Thermal Denaturation of Food Proteins
Effects of Alcohols on Thermal Denaturation of Food Proteins
Effects of Odorants on Thermal Denaturation of Food Proteins
Effects of Polysaccharides on Thermal Denaturation of Food
Proteins
Postdenaturation Aggregation of Food Proteins
Conclusion
References
87
89
95
99
102
105
110
112
113
Introduction
Normally, food contains a heterogeneous, heterophase mixture of
high- and low-molecular-weight components and their aggregates and
complexes. Among food macromolecules, proteins and polysaccharides are largely responsible for the structural changes accompanying
food processing and for mechanical and other physical properties of
foods.
Proteins are both most-multifunctional biopolymers and mostversatile food macromolecules. Normally, a functional protein has a
unique ordered (crystal-like) molecular structure, which appears to be
87
88
89
90
91
92
d G (T0 , pH ) = d H ( pH ) 1
+ d C p [T0 Td ( pH )0 ]
Td ( pH )
T0
T0 d C p ln
Td ( pH )
(5.1)
Here, the standard temperature T0 = 352 K is the denaturation temperature at a reference value of pH (pH0 3.5).
The excess denaturation free energy, that is, a general criterion of
pH effects on protein denaturation, can then be determined:
d G E ( pH ) = d G (T0 , pH ) d G (T0 , pH 0 )
(5.2)
= 2.303 d
RT pH T
(5.3)
In the case of 11S globulin, dH+ 3 per each constituent polypeptide chain. This value is close to the similar estimates for small globular
proteins (Tanford 1968, 1970; Nicoli and Benedek 1976).
The origin of the denaturation proton adsorption by proteins at acid
pH values is well-known (Tanford 1968, 1970). In the native protein
93
molecule, there are often abnormal carboxyl groups with pKa 1.5
(Nicoli and Benedek 1976) localized in the nonpolar interior of the
molecule. Their ionized state is stabilized by hydrogen bonds with
neighboring residues of tyrosine, lysine, or histidine. Several of these
bonds decrease with decreasing pH, which leads to a decrease in the
conformational stability of the protein globule. In the course of the
denaturation, the hydrogen bonds tyrosyl (histidyl)-carboxylate are
broken, and the abnormal carboxyl groups become normal groups with
pKa 3 to 4. They turn to the nonionized state by the adsorption of
protons. Typically, several of the abnormal carboxyl groups in the
protein molecule are rather small; for example, they do not exceed 2
to 3 for some small globular proteins (Nicoli and Benedek 1976).
The considered approach was used to analyze the pH effects on the
stability of a number of other oligomeric and small food proteins, such
as ribulose 1,5 biphosphate carboxylase (RBPC) of alfalfa (Burova et
al. 1989B), tobacco (Burova et al. 1991), and other green leaves; 7S
globulin of French beans (Burova et al. 1989b, 1992); soybean trypsin
Kunitz inhibitor (Varfolomeeva et al. 1989; Burova et al. 1990;
Grinberg et al. 2000); and porcine -lactoglobulin (Burova et al.
2002a).
The 7S globulin of French beans, phaseolin, is an oligomeric storage
protein with the molecular weight of about 150 kDa. It contains three
subunits (Paaren et al. 1987). Each subunit involves two domains
(Lawrence et al. 1990). The thermal denaturation of 7S globulin was
studied in the range of pH 2.0 to 10.9 (Burova et al. 1992). The quaternary structure of the protein is stable within this pH range; however
the denaturation thermogram of phaseolin has a complex profile. In
addition to the main heat capacity peak, there is a lower temperature
shoulder. The thermogram recalculated per polypeptide chain can be
deconvoluted into two independent two-state transitions that represent
unfolding of the domains of phaseolin. Dependences of the denaturation temperature and enthalpy of both domains on pH pass through a
maximum at pH 5.4. The latter corresponds to the isoelectric point of
phaseolin. For both domains, the temperature dependences of the denaturation enthalpies strictly follow the Kirchhoff s law. The excess
denaturation free energies of the domains are maximal in the vicinity
of the isoelectric point of phaseolin. The analysis of the excess denaturation free energies of the domains as a function of pH showed
that there are at least two types of the side-chain hydrogen bonds:
94
95
96
(5.4)
97
98
and by the lyotropic effect of the salt on the structure of water. The
excess denaturation free energy can be determined as a sum of contributions of these two effects:
d G E (Cs ) = d GeE (Cs ) + d GsE (Cs )
(5.5)
where d GeE (Cs ) and d GsE (Cs ) are the contributions of the electrostatic and lyotropic effects, respectively. The contribution of the electrostatic effect can be expressed in the Debye-Hckel approximation
(Tanford 1965):
d GeE (Cs ) = RT Be [ F ( ) F ( 0 )]
(5.6)
where
F ( ) =
2 1 + RN
(5.7)
(5.8)
99
Table 5.1. Salt-protein interaction parameters Be and dBs for 11S globulin
from broad beans.
dBs, L mol1
Salt
NaCl
KCl
(NH4)2SO4
106Be, cm
Experimental
Group contribution
method
MelanderHorvath theory
1.2
1.2
2.3
8.2
8.2
10.6
8.4
8.8
7.8
11.0
According to the obtained dBs values, the salts can form a series
(NH4)2SO4 >> NaCl = KCl in agreement with general features of the
lyotropic effect (von Hippel and Schleich 1969).
It is possible to calculate the parameter dBs by the group contribution method (Nandi and Robinson 1972) and the Melander-Horvath
theory (Melander and Horvath 1977). In both cases, the theoretical
estimates of the parameter dBs of 11S globulin are very close to its
experimental values (Table 5.1).
100
(5.9)
Figure 5.3. Thermal denaturation of 11S broad bean globulin in the presence
of ethanol at pH 7.6. (a) Thermograms at the different alcohol concentrations, CA.
(b) Denaturation temperature and enthalpy versus the alcohol concentration. (c)
Correlation between values of the denaturation enthalpy and the denaturation temperature obtained at the different alcohol concentrations according to Kirchhoff s law.
The solid line corresponds to the average value of the experimental denaturation heat
capacity increment, dCp = 0.31 0.02 J g1 K1. The dashed lines represent a prediction range of the correlation. (d) Excess denaturation free energy per constituent chain
of the protein versus ethanol concentration at temperature T0 = 352 K. The line is
calculated by a two-state model using the linear approximation of alcohol-protein
interactions in terms of the denaturation increment of the second virial coefficient,
dBA. The insert gives a correlation between values of dBA determined experimentally and those calculated by an additive scheme based on independent group contributions of amino acids for lysozyme, ribonuclease, and 11S broad bean globulin. The
correlation coefficient is more than 0.999.
101
102
103
104
(5.10)
105
106
Table 5.2. Denaturation temperature of 11S globulin from broad beans (C) in
the presence of polysaccharides.*
pH 7.6; q = 20
Polysaccharide
Dextran
Sodium alginate
Pectin
Carboxymethylcellulose
Methyl cellulose
Arabic gum
Dextran sulfate
-Carrageenan
-Carrageenan
Free protein
0.01 M NaCl
0.4 M NaCl
Peak 1
Peak 2
74.9
75.1
74.3
75.4
74
74.3
76.3
77.9
76.4
76.0
63.0
61.0
64.0
90.3
90.8
93.1
93.0
pH 4.2; q = 1
77.0
59.0
62.0
50.5
Table 5.2 shows that the polysaccharides incompatible with 11S globulin (pH 7.6; 0.01 M NaCl), such as dextran, alginate, pectin, methyland carboxymethylcellulose, and Arabic gum, do not affect the thermal
denaturation of the protein. Under the same conditions, the sulfatecontaining polysaccharides can form complexes with 11S globulin.
The result is a destabilized form of 11S globulin coexisting with the
free protein. When the complexation is inhibited by 0.4 M NaCl, the
peak 2 disappears, and only the peak 1, corresponding to the denaturation of the free 11S globulin, remains in the thermogram. These data
imply that the 11S globulin-polysaccharide incompatibility does not
significantly affect the protein unfolding. On the contrary, the 11S
globulin-polysaccharide complexation results in the destabilization of
the protein.
At pH 4.2 11S globulin is able to form cooperative electrostatic
complexes with anionic polysaccharides, both sulfate- and carboxylcontaining. Under these conditions in the presence of alginate, pectin,
and dextran sulfate, the thermograms of 11S globulin have a single
denaturation peak (peak 2) at temperatures well below the denaturation
temperature of the free 11S globulin (Table 5.2). No free protein (peak
107
108
109
sulfate was lower than that of the free protein in the whole range of q
values and decreased slightly with increasing q values (Figure 5.5c).
Let us analyze possible mechanisms responsible for changes in
conformational stability of a protein bound to a polymer matrix. The
protein stability to denaturation is determined by the free energy of
denaturation, dG, which is the difference between free energies of the
protein in unfolded, GD, and native, GN, form: dG = GD GN. When
N or D forms of the protein are bound to a matrix, the free energy of
each form decreases by a value of the free energy of binding, bGN or
bGD, respectively. Reasonably, the free energy of binding depends on
the number of contacts formed by the protein upon its fixation on the
matrix. As a rule, the D form of a protein possesses a higher number
of accessible binding sites than the N form, consequently bGD << bGN
(the case of a preferential binding of the D form). This situation is the
most probable in complexes with a low protein occupancy on the
matrix (Figure 5.5d). A gradual saturation of the complex with protein
results in a decrease in the number of free binding sites on the polysaccharide matrix (Figure 5.5e). This leads to a decreasing probability of
preferential binding of the D form. In such a densely occupied complex,
the stability of bound protein approaches that of the free protein, but
does not exceed it.
One could expect that thermodynamic incompatibility of proteins
and polysaccharides may result in an increase in the denaturation temperature of protein due to the excluded volume effect (Grinberg and
Tolstoguzov 1997). The minor manifestation of this effect in the case
Figure 5.5. Thermal denaturation of soybean trypsin (Kunitz) inhibitor in the presence of dextran sulfate under the protein-polysaccharide complexation conditions (pH
3.0, ionic strength 0.005). (a) Thermograms at the different protein-polysaccharide
weight ratios, q. (b) Sedimentation coefficients of the components of the proteinpolysaccharide mixtures at the different protein-polysaccharide weight ratios; 1:
heavy protein-polysaccharide complex; 2: light protein-polysaccharide complex.
The dashed line corresponds to the free protein (sw,200 2S). (c) Denaturation
temperature and enthalpy of the protein in the complexes versus the proteinpolysaccharide weight ratio. The dashed lines represent the corresponding denaturation parameters of the free protein. (d and e) Schematic presentation of the
denaturation of a protein (P) bound to a polymer matrix (M) at the loose (d) and dense
(e) protein occupancy. bGN and bGD are free energies of binding of the protein in
the native and denatured states to the matrix. Td is the change in the denaturation
temperature of the protein due to binding.
110
111
When the rate of protein denaturation is much less than that of the
aggregation of the unfolded protein molecules, the aggregation rate is
determined by the denaturation rate. In this case, the aggregation
process could be considered as a first-order reaction. Under these conditions, information about the aggregation kinetics can be directly
derived from the DSC denaturation thermograms obtained at different
heating rates.
Such an approach was used to study the postdenaturation aggregation of ovalbumin (Weijers et al. 2003). From the denaturation temperature-heating rate dependence (Sanchez-Ruiz et al. 1988) the
activation energy and the aggregation frequency factor were determined. As a result, the aggregation rates were calculated for the temperature range (6787 C), which covers the denaturation heat capacity
peak of the protein in the DSC thermogram. The aggregation constant
of ovalbumin increases by more than 4 orders of magnitude over the
temperature range under investigation.
An attempt was made to extract kinetic parameters of the postdenaturation aggregation of a protein directly from its denaturation thermograms (Remmele et al. 2005). It was suggested that the unfolded form
of protein, U, participates in the aggregation. Its concentration is generally determined by the conformational equilibrium. During the beginning stage of the aggregation, dimers of unfolded protein molecules
(the form D) are mainly formed by two paths of dimerization according
to the mono- and bimolecular mechanisms. An essence of the model
is illustrated by the scheme:
k
k1
D1
k2
D2
U
N
]= D
(5.11)
112
(5.12)
d Happ (t )
d H
(5.13)
113
114
115
Danilenko A.N., Bikbov T.M., Grinberg V.Y., Leontiev A.L., Burova T.V., Surikov
V.V., Borisov Y.A., and Tolstoguzov V.B. 1987. Effect of pH on conformational
stability of 11S globulin from Glycine max seeds according to differential scanning
microcalorimetry. Biofizika, 32(3):402406.
Delben F. and Stefancich S. 1998. Interaction of food polysaccharides with ovalbumin. Food Hydrocolloids, 12(3):291299.
Grinberg V.Y. and Tolstoguzov V.B. 1997. Thermodynamic incompatibility of proteins and polysaccharides in solutions. Food Hydrocolloids, 11(2):145158.
Grinberg V.Y., Burova T.V., Grinberg N.V., and Mashkevich A.Y. 1993. On the
effect of the denaturation degree of food proteins on their functional properties.
In: Food Proteins: Structure and Functionality, edited by K.D. Schwenke and R.
Mothes, pp. 4047. Wiley-VCH Publishing: Weinheim, Germany.
Grinberg V.Y., Burova T.V., Haertle T., and Tolstoguzov V.B. 2000. Interpretation
of DSC data on protein denaturation complicated by kinetic and irreversible
effects. J Biotechnol, 79(3):269280.
Grinberg V.Y., Danilenko A.N., Burova T.V., and Tolstoguzov V.B. 1988. On physical mechanism of thermal transitions in 11S globulins from some seeds. Biofizika,
33(4):559561.
Grinberg V.Y., Danilenko A.N., Burova T.V., and Tolstoguzov V.B. 1989.
Conformational stability of 11S globulins from seeds. J Sci Food Agric, 49(2):
235248.
Grinberg V.Y., Grinberg N.V., Burova T.V., Dalgalarrondo M., and Haertle T. 1998.
Ethanol induced conformational transitions in holo--lactalbumin: Spectral and
calorimetric studies. Biopolymers, 46(4):253265.
Grinberg V.Y., Grinberg N.V., Mashkevich A.Y., Burova T.V., and Tolstoguzov V.B.
2002. Calorimetric study of interaction of ovalbumin with vanillin. Food
Hydrocolloids, 16(4):333343.
Grozav E.K., Danilenko A.N., Bikbov T.M., Grinberg V.Y., and Tolstoguzov V.B.
1985. Studies on the effect of ethanol on thermal denaturation of soybean globulins
by differential scanning microcalorimetry. J Food Sci, 50(5):12661270.
Hambling S.G., McAlpine A.S., and Sawyer L. 1992. -Lactoglobulin. In: Advanced
Dairy Chemistry: Proteins, P.F. Fox, editor, pp. 141190. Elsevier Applied
Science: London.
Hoedemaeker F.J., Visschers R.W., Alting A.C., de Kruif C.G., Kuil M.E., and
Abrahams J.P. 2002. A novel pH-dependent dimerization motif in -lactoglobulin
from pig (Sus Scrofa). Acta Crystallogr D, 58(3):480486.
Holt C. 2000. Molecular basis of whey protein food functionalities. Aust J Dairy
Technol, 55(2):5355.
Ibanoglu E. 2005. Effect of hydrocolloids on the thermal denaturation of proteins.
Food Chem, 90(4):621626.
Ibanoglu E. and Ercelebi E.A. 2007. Thermal denaturation and functional properties
of egg proteins in the presence of hydrocolloid gums. Food Chem, 101(2):
626633.
116
117
118
Tolstoguzov V. 1991. Functional properties of food proteins and role of proteinpolysaccharide interaction. Food Hydrocolloids, 4(6):429468.
Tolstoguzov V.B. 1998. Functional properties of protein-polysaccharide mixtures. In:
Functional Properties of Food Macromolecules, J.R. Mitchell, D.A. Ledward, and
S. Hill, editors, pp. 252277. Blackie Academic & Professional: London.
Tolstoguzov V.B. 2000. Foods as dispersed systems. Thermodynamic aspects of
composition-property relationships in formulated food. J Thermal Anal Calorim,
61(2):397409.
Tolstoguzov V.B. 2002. Thermodynamic aspects of biopolymer functionality in
biological systems, foods, and beverages. Crit Rev Biotechnol, 22(2):89174.
Tolstoguzov V.B., Grinberg V.Y., and Gurov A.N. 1985. Some physicochemical
approaches to the problem of protein texturization. J Agri Food Chem, 33(2):
151159.
Tsereteli G.I. 1982. Thermal denaturation of collagen in solutions and fibrils. Biofizika,
27(5):780785.
van Koningsveld G.A., Gruppen, H., de Jongh H.H.J., Wijngaards G., van Boekel
M.A.J.S., Walstra P., and Voragen A.G.J. 2002. Effects of ethanol on structure and
solubility of potato proteins and the effects of its presence during the preparation of
a protein isolate. J Agric Food Chem, 50(10):29472956.
Varfolomeeva E.P., Burova T.V., Grinberg V.Y., and Tolstoguzov V.B. 1989.
Thermodynamic and kinetic study of thermal denaturation of the Kunitz trypsin
inhibitor from soybean by differential scanning microcalorimetry. Mol Biol
(Moscow), 23(5):10001008.
von Hippel P. and Schleich T. 1969. The effects of neutral salts on the structure and
conformational stability of macromolecules in solution. In: Structure and Stability
of Biological Macromolecules, C. Timasheff and G. Fasman, editors, pp. 417574.
Marcel Dekker: New York.
Wang Q., Tolkach A., and Kulozik U. 2006. Quantitative assessment of thermal
denaturation of bovine -lactalbumin via low-intensity ultrasound, HPLC, and
DSC. J Agric Food Chem, 54(18):65016506.
Weijers M., Barneveld P.A., Cohen-Stuart M.A., and Visschers R.W. 2003. Heatinduced denaturation and aggregation of ovalbumin at neutral pH described by
irreversible first-order kinetics. Protein Sci, 12(12):26932703.
Wu Y.V. and Scheraga H.A. 1962. Studies of soybean trypsin inhibitor. I.
Physicochemical properties. Biochemistry, 1(4):698705.
Yamasaki M., Yano H., and Aoki K. 1991. Differential scanning calorimetric studies
on bovine serum albumin. 2. Effects of neutral salts and urea. International J Biol
Macromolecules, 13(6):322328.
Zhang G.Y., Foegeding E.A., and Hardin C.C. 2004. Effect of sulfated polysaccharides on heat induced structural changes in -lactoglobulin. J Agric Food Chem,
52(12):39753981.
Chapter 6
Heat-Induced Phase Transformations of
Protein Solutions and Fat Droplets in Oilin-Water Emulsions: A Thermodynamic
and Kinetic Study
Perla Relkin
Introduction
Heat-Induced Transformations in Protein Solutions
Protein Structures
Thermodynamics of Protein Heat-Induced Transformations
Denaturation-Aggregation of Globular Proteins in Bulk Phase
System
Thermodynamics and Kinetics of Heat-Induced Transformations
Heat-Induced Transformations in Oil-in-Water Emulsions
Crystallization and Melting of Fat Droplets
Kinetics of Fat Droplet Crystallization in Oil-in-Water
Emulsions
Conclusion
References
119
121
121
123
124
129
132
132
136
141
141
Introduction
The thermomechanical treatments applied for food manufacturing
involve batch or continuous heating and cooling steps for mixing, aging,
pasteurization, cooking, or storage. Monitoring the effects of heatinduced transformations in raw ingredients and additives can help to
optimize such food processing or storage conditions in terms of macroscopic properties, quality attributes, and shelf life of the final products.
119
120
121
122
structures (characterized by a hydrophobic core from which H2O molecules are squeezed and surface-exposed charged amino acids), and
eventually quaternary structures resulting from non-covalent bonding
between monomers. Under the effect of heat treatment, the initial
structure of globular proteins is altered without hydrolysis of primary
covalent bonds. The thermal transition between an initial low-temperature state to a high-temperature state is called denaturation (Privalov
1979; Brandts and Lin 1990; Privalov and Potekin 1996). Compared
with the protein initial state (native protein), the newly created conformational state called denatured is characterized by a lower proportion
of high-order structures and a higher exposure to the solvent environment of hydrophobic groups initially buried in the protein core (Mills
1976; Cooper 1999). In addition, globular proteins possessing disulfide, thiol groups prone to SH/S-S interchange, and intermolecular
disulfide bonds have S-S reactions at neutral or basic pH values, especially in denaturing conditions (Liu, Relkin, and Launay 1994).
Multimeric proteins may dissociate into monomers, before or during
denaturation. In such conditions, thermodynamic laws are not relevant
to the study of heat-induced transformations, which may occur successively or simultaneously with interaction mechanisms between
unfolded proteins themselves or between other solute or surfaces,
contributing to aggregation or gelation, ligand binding, and interfacial
properties (Lefbvre and Relkin 1996).
Gelatin molecules derive from chemical transformation of collagen.
They display extended overall shapes, but likely as for the other polypeptides they present and secondary structures sharing basically
similar conformational change properties under heating and cooling.
But contrary to globular proteins, and likely for linear polysaccharide
chains, gelatin molecules undergo heat-induced reversible-ordered
helix-to-disordered random coil transitions in most of the conditions
used in a variety of food applications (low-fat yogurt or creams,
mousses). Due to their thickening and gelling properties, they are
added to milk proteins to improve the texture and firmness of dairy
products (Fiszman, Lluch, and Salvador 1999).
Caseins, the major milk protein component, is not susceptible to
heat-induced denaturation. When considered as individual molecules,
they are much less compact and organized than other proteins, such as
gelatin (helical secondary structure), or globular proteins (high-order
tertiary structure). However, the degree to which globular proteins
123
(e.g., whey, blood plasma, egg white, soya proteins) or extended proteins (gelatin) can interact with casein is also related to their structural
properties (Haque, Kristjansson, and Kinsella 1987; Kinsella and
Whitehead 1989).
Thermodynamics of Protein Heat-Induced Transformations
The denaturation mechanism of small-molecular-weight globular proteins may occur after a reversible two-state model (Privalov 1979;
Brandts and Lin 1990; Relkin 1996; Cooper 1999):
K eq
[U ]
[N ]
GNU (T ) = H NU (T ) T SNU (T ) = RT ln K (T )
(6.1)
(6.2)
(6.3)
124
k1
k2
(6.4)
125
126
2%
4%
6%
10%
20
30
40
50
60
70
80
Figure 6.1. Heating curves obtained from solutions of whey proteins at different
concentrations (pH 6.6, 0.1 C.min1).
127
1
20 mWg1
endo
0.75
0.5
0.25
0.1
40
50
60
70
Temperature, C
80
90
Figure 6.2. Heating curves obtained at different scanning rates from a solution of
whey proteins at 4.15% protein concentration (pH 6.6).
128
74
Peak temperature, C
72
20
18
70
16
68
14
12
66
10
15
20
% Protein concentration
25
22
10
30
Figure 6.3. Variations of peak temperatures (Tmax in C) and heat of reaction (Qcal in
J.g1) obtained from solutions of whey proteins at the indicated concentrations (pH
6.6, 0.1 C.min1).
129
ln
Tmax
E A RTmax
(6.5)
Activation enthalpy (H#) and entropy (S#) were deduced from the
following relations:
H # = E A RT
(6.6)
S #
= ln( Zh ) ln(ekbT )
R
(6.7)
130
ln (/Tmax)
8
2.86
2.88
2.9
1000/Tmax
2.92
2.94
131
Table 6.1. Examples of calorimetric (Tmax, Qcal) and thermodynamic (EA, H#)
parameters, deduced from DSC heating curves obtained from a protein solution
(4.15% concentration, pH 6.6) using two weight samples and two ranges of
scanning rates. Lumry-Eyring theory was applied to DSC curves obtained at
the indicated scan rates, and activation energy values, EA, were calculated from
the linear part of plots in Fig. 5.4 (see text).
Sample
weight
(mg)
750
45
dt/dT
(C.min1)
Tmax
(C)
Qcal
(kJ.mol11)
EA
(kJ.mol11)
H#
(kJ.mol11)
0.1
5
68.2
76.3
283
232
342
345
340
342
these differences could indicate differences in heat-induced conformation transitions and reaction mechanisms, depending on the heating
rate and constant rate of reactions, as suggested previously (Relkin and
Launay 1990; Sanchez-Ruiz 1992; Relkin 2004).
Milk proteins are composed by approximately 80% caseins in
micelle form and 20% whey proteins, of which half are -lactoglobulin.
The DSC curves (1 C.min1) in Figure 6.5 were obtained from solutions in simulated milk ultrafiltrate (SMUF, pH 6.6) of a whey protein
isolate, alone or in mixture with 20% casein or 40% casein at 5.3%
total protein concentration. These curves showed the presence of an
exothermic reaction occurring at T > 80 C. Partial replacement of
whey proteins by 20% or 40% casein micelles gave DSC curves
composed of a major endothermic peak at Tmax, accompanied by an
exothermic effect at T > Tmax.
The intensity of the exothermic event seems to increase, whereas
Tmax seems to decrease with increasing casein-to-whey protein weight
ratio. The lowering of Tmax with increased proportion of casein to whey
proteins could be explained by an increase in the rate of irreversible
aggregation mechanism between casein and unfolded whey proteins.
Upon heating, some of the hydrophilic interactions (hydrogen bonds,
van der Waals interactions, electrostatic interactions between charged
groups, specific binding) are weakened, whereas some of the hydrophobic amino acids (initially buried in the interior core of whey proteins) become more exposed at the surface. In the example of Figure
6.1 (2% protein in water; > 0.5 C.min1), the exothermic signal is
shown at T < Tmax, whereas in the example of Figure 6.5 (5.3% protein
132
3.5
3
(c)
2.5
(b)
(a)
1.5
1
40
60
80
Temperature (C)
Figure 6.5. Heating curves obtained from solutions of whey proteins, alone (a), or
in mixtures with either 20% casein (b), or 40% casein (c). 5.3% total protein, pH 6.6
in simulated milk ultrafiltrate, 1 C.min1.
133
water food emulsions, proteins are used in combination with smallmolecular-weight emulsifiers (surfactants) and polysaccharides
(Dickinson 1998). Proteins compete with surfactant molecules for the
adsorption to the oil-water interface, giving appropriate interfacial
properties, whereas polysaccharides are used as thickeners of the
aqueous continuous phase. In addition to these parameters that have
effects on colliding properties of fat droplets and their resistance to
coalescence, crystallization behavior of fat droplets was shown to play
a major role in stability and instability of food emulsions. Several
techniques may be used to study thermal behavior of fat in bulk and
emulsified phases (Dickinson and McClements 1995; Hindle, Povey,
and Smith 2000; Garti and Sato 2001). Because crystallization of fat
releases a large amount of heat, numerous studies were performed
using DSC in nonisothermal and isothermal modes. It was shown that
crystallization temperature of dispersed fat droplets is lowered, compared to bulk fat (Skoda and van den Temple 1963; Walstra and van
Beresteyn 1975; Walstra, Kloeck, Vliet Ton van 2001). In addition to
droplet curvature, supercooling needed to initiate crystallization of fat
globules depends on several other factors, including the origin and
composition of fat, adsorbed materials, and, particularly, added lipophilic or hydrophilic emulsifiers (Dickinson and McClements 1995;
Garti and Sato 2001; Relkin and Sourdet 2005; Relkin et al. 2008).
Besides oil-in-water food emulsions, such as sauce or mayonnaise
where polyunsaturated lipids are used, there are other types of emulsions prepared from anhydrous milk fat (AMF) that are used for fabrication of dairy whipped cream or ice creams. AMF is constituted by
a wide diversity of saturated and unsaturated triacylglycerols (TG),
each characterized by its own melting temperature (Hartel and
Kaylegian 2001). The physical properties of AMF, resulting mainly
from its extraction processing and TG composition, have different
temperature dependency. AMF has broad melting and crystallization
temperature ranges from approximately 40 C to 45 C, and it may
contain more than 50% crystalline fat when stored at 5 C (refrigerator
temperature). Therefore, the manufacturing process of dairy emulsions
consists of successive steps. In the first step, AMF is heated above its
melting temperature (50 C), and lipophilic emulsifiers are dispersed
in the lipid melt. In the second step, this lipid-melt phase is mixed by
stirring with the aqueous phase, which contains water-soluble ingredients (proteins and polysaccharides). In the third step, the premix is
134
passed through a high-pressure homogenizer, and the resulting emulsion is cooled down to a storage or ageing temperature (Relkin, Sourdet,
and Fosseux 2003). Emulsions used to prepare dairy whipped cream
or ice creams are aged at 4 C for a time ranging from 10 to 18 h.
During the aging step, in addition to complete hydration of polysaccharides and partitioning properties of surfactants and proteins between
the oil-in-water interface and the aqueous continuous phase, the behavior of fat droplets and their heat-induced transformations are considered as key factors for the development of desired structures in the
final product (Abd El Rahman et al. 1997; Goff 2002).
DSC is used to study thermal behavior of ingredients (fat, surfactant,
polysaccharide) as a function of processing parameters, and especially
for evaluation of supercooling and kinetics of fat crystallization from
a liquid emulsified phase. Supercooling, needed to initiate fat crystallization from melted systems can be evaluated from the cooling and
re-heating DSC signals registered in a scanning mode. Examples of
DSC curves obtained from AMF in bulk phase (20 mg) or fat droplets
in protein-stabilized emulsions (80 mg) are shown in Figures 6.6 and
6.7. All the samples were heated to 50 C (crystal melting) before
cooling. The curves in Figures 6.6 and 6.7 were obtained from cooling
and reheating cycles at 0.5 C.min1. In the Figure 6.7, besides the DSC
cooling (a) and reheating (b) curves (0.5 C.min1), we present also the
melting curve (c) obtained after cooling to 4 C and holding the emulsion at this temperature for 10 h.
These curves present distinguishable exothermic and endothermic
events corresponding to crystal formation and melting of crystals (or
polymorphs), respectively. Comparison of the temperatures of the
initial scan and scans after cooling and reheating indicated a higher
supercooling in emulsified samples than in bulk fat samples (Table
6.2). The shape of the melting curves obtained by reheating (0.5 C.
min1) just after cooling at the same scan rate was very similar for all
the fat samples (Figures 6.6 and 6.7). However, the shape of the crystallization curves differed depending on the fat sample (bulk- or emulsified-fat sample) and ingredient composition. Melting curves (0.5 C.
min1) obtained after a holding step at 4 C for 10 h applied to bulk
(AMF-0 and AMF-S) or emulsified (E-0 and E-S) fat samples in the
absence or presence of surfactant, respectively, show a broad endothermic curve (Figure 6.7, curve c), with Tmax (maximum peak temperature) located at around 20 C and Qcal (apparent heat of reaction)
Heat flow, mW
AMF
endo
AMF-S
10
20
Temperature, C
30
40
Figure 6.6. Cooling and heating curves (0.5 C.min1) obtained from anhydrous milk
fat alone (AMF) or with 1.75 wt% added surfactant (AMF-S).
endo
(c)
(b)
(a)
10
20
Temperature, C
30
40
Figure 6.7. Cooling (a) and first reheating (b) curves, and second reheating curve (c)
observed at 0.5 C.min1 from protein-stabilized AMF emulsion containing added
surfactant (E-S). The second reheating curve (c) was registered after 10 h holding of
the emulsion at 4 C.
135
136
Tcris
(C)
Tend
(C)
Tmax (10 h)
(C)
AMF
AMF-S
E-0
E-S
22.4
19.7
19.9
19.1
38.0
38.0
38.0
35.5
20.6
20.6
21.1
20.2
close to 65 J.g1. This indicates that for both bulk and emulsified fat
samples there was formation of a similar amount of crystalline fat
during the 10-h aging at 4 C.
Application of DSC in isothermal mode (4 C) to the same bulk fat
and emulsions led to observation of a single exothermic heat flow
signal as seen in Figures 6.8 and 6.9. The maximum heat flow deviation of this exothermic peak occurs after different holding times at
4 C. Compared with the bulk AMF-0 sample, this event seemed to
occur after a similar holding period (27 min) in E-S (emulsion with
surfactant). However, it seems to be anticipated for AMF-S (bulk fat
in presence of surfactant) and more delayed (5 min) in the proteinstabilized emulsion without added surfactant.
Kinetics of Fat Droplet Crystallization in Oil-in-Water Emulsions
Analysis of the heat flow pattern involved during the isothermal step
could be used for evaluation of crystal growth characteristics, such as
the induction time and growth rate values. From the determination of
the partial apparent heat of reaction at time t (calculated from the
partial area under the exothermic heat flow), it is possible to obtain the
fractional fat crystallization, X(t), from the following relation:
X (t ) =
A(t )
Qcal
(6.8)
AMF-0
AMF-S
10
40
60
Holding time at 4C, min
80
Figure 6.8. Isothermal curves registered at 4 C from anhydrous milk fat alone
(AMF-0) or with 1.75 wt% added surfactant (AMF-S).
E-0
2
E-S
20
40
60
Holding time at 4C, min
80
137
138
where Qcal is the total calorimetric heat of reaction calculated from the
area under the exothermic peak registered during the holding time.
The mechanism of fat crystal growth can be described by the Avrami
equation (Avrami 1939):
n
X ( t ) = 1 exp ( t )
(6.9)
or can be linearized:
ln [ ln(1 X ( t ))] = ln( ) + n ln( t )
(6.10)
e
X( t ) = X max exp exp max ( t ind t ) + 1
X max
(6.11)
139
ln (ln (1-TS))
8
1.5
2.5
3
ln (t), min
3.5
4.5
Figure 6.10. Avrami plots obtained from anhydrous milk fat alone (square symbols),
from protein-stabilized emulsion without added surfactant (diamonds), or with added
surfactant (circles).
AMF-0
E-0
E-S
Avrami index
tmax
min
tinduction
min
max
min1
R2
27.2 2.1
32.7 1.2
27.0 3.0
14.2 0.12
22.4 0.04
13.5 0.02
3.48
4.00
3.28
3.461
4.038
2.654
0.999
0.992
0.999
140
50
80
60
X (t)
100
40
20
20
40
60
Holding time at 4C min1
80
0
100
Figure 6.11 Examples of time evolution of X(t) and partial completion of crystallization reaction at 4 C, determined from anhydrous milk fat sample (experimental
values, circles) and its Gompertz representation (sigmoidal dashed curve), as deduced
from the endothermic heat flow signal (see text).
Table 6.3. According to the Gompertz method, AMF in the bulk phase
and in the emulsion containing milk proteins with surfactant (E473)
have very close tmax (27 min) and tind values (14 min), an n value of 3.5
and 2.7, respectively, and an max value of 3.5 min1 and 3.3 min1,
respectively. On the other hand, the protein-stabilized emulsion (E-0)
without added surfactant exhibits a higher value of n 4 (indication
of a spherulic crystal growth) and higher values of tmax, tind (delay in
nucleation), and max (increase in the crystal growth rate).
Emulsification procedure and ingredient complexity have a dominant role in characteristics of fat droplets, such as particle average
diameters and size distributions, composition and physical properties
of surrounding surface layers, and crystalline fat content and polymorphism (Skoda and van den Tempel 1963; Walstra 1975; McClements
et al. 1993; Dickinson and McClements 1995; Kaneko et al.1999;
Hindle, Povey, and Smith 2000; Relkin, Sourdet, and Fosseux 2003;
Relkin and Sourdet 2005). The supercooling effect (temperature needed
to initiate fat crystallization in globules) has been shown to differ
depending on fat composition and mean droplet size of fat droplets and
on the concentration and lipophilic or hydrophilic nature of emulsifiers
141
(Garti and Jano 2001). D50 values (average median diameters of fat
droplets) determined from laser light scattering measurements (Relkin
and Sourdet 2005) were close to 3.1 m for E-0 and E-S emulsions.
However, whereas a similar supercooling was determined by DSC in
scanning mode (Table 6.2), kinetic parameters (tmax, tind, max) and n
(Avrami index) deduced from the DSC isothermal method were different (Table 6.3). These results indicate, as expected from numerous
studies, that hydrophobic surfactant acts as a catalyzer for crystal
nucleation in fat globules, where crystallization mechanism is considered as homogeneous.
Conclusion
DSC has been used for several years to investigate heat-induced conformational or structural changes of a broad range of food ingredients
(biopolymers, proteins, fats, sugars, emulsifiers) in various physicochemical conditions. Most of the previous DSC studies showed the
ability of food ingredients in heat-induced structural or physical state
changes, which are of great importance for manufacturing of food
products with controlled structures. The examples described in this
chapter indicate that combining DSC in nonisothermal and isothermal
methods can provide thermodynamic and kinetic data to contribute to
better understanding and control of structure-forming mechanisms in
food systems.
References
Abd El Rahman A.M., Madkor S.A., Ibrahim F.S., and Kilara A. 1997. Physical
characteristics of frozen desserts made with cream, anhydrous milk fat or milk fat
fraction. J Dairy Sci, 80:19261935.
Avrami M. 1939. Kinetics of phase change. I. General theory. J Chemical Physic,
7:11031112.
Barfod N.M., Krog N., Larsen G., and Buchheim W. 1991. Effects of emulsifiers on
protein-fat mixtures in ice-cream mix during aging. I. Quantitative analyses. Fat
Sci Technol, 93:2429.
Bazmi A., Duquenoy A., and Relkin P. 2007. Aeration of low fat dairy emulsions:
Effects of saturated-unsaturated triglycerides. Int Dairy J, 17:10211027.
142
Bolliger S., Goff H.D., and Tharp B.W. 2000. Correlation between colloidal properties of ice cream mix and ice cream. Int Dairy J, 10:303309.
Boode K., Walstra P., and de Groot-Mostert A.E. 1993. Partial coalescence in oil-inwater emulsions.2. Influence of the properties of the fat. Colloids Surfaces A,
81:139151.
Brandts J.F. and Lin L.N. 1990. Study of strong to ultratight protein interactions using
differential scanning calorimetry. Biochemistry, 29:69276940.
Clark A.H. and Lee-Tuffnell C.D. 1986. Gelation of globular proteins. In: Functional
Properties of Food Macromolecules, J.R. Mitchell and D.A. Ledward, editors, pp.
203272. Elsevier: London.
Cooper A. 1999. Thermodynamics of protein folding and stability. In: Protein: A
Comprehensive Treatise, Vol. 2, A. Geoffrey, editor, pp. 217270. JAI Press:
Stamford, CT.
Dalgleish D.G. 1996. Conformations and structures of milk proteins adsorbed to oilwater interfaces. Food Res Int, 29:541547.
Dalgleish D.G., Van Mourik L., and Corredig M. 1997. Heat-induced interactions of
whey proteins and casein micelles with different concentrations of -lactalbumin
and -lactoglobulin. J Agric Food Chem, 45:48064813.
De Wit J.N. and Klarenbeek G. 1984. Effects of various treatments on structure and
solubility of whey proteins. J Dairy Sci, 67:27012710.
Dickinson, E. 1992. Structure and composition of adsorbed protein layers and
the relationship to emulsion stability. J Chem Soc Faraday Trans, 88:
29732983.
Dickinson E. 1997. Properties of emulsions stabilized with milk proteins: Overview
of some recent developments. J Dairy Sci, 80:26072619.
Dickinson, E. 1998 Stability and rheological implications of electrostatic milkproteinpolysaccharide interactions. Trends Food Sci Technol, 9:347354.
Dickinson E. and McClements D.J. 1995. Fat crystallization in oil-in-water emulsions. In: Advances in Food Colloids, E. Dickinson and D.J. McClements, editors,
pp. 211246. Blackie Academic & Professional: London.
Donovan M. and Mulvihill D.M. 1987. Thermal denaturation and aggregation of
whey proteins. Irish J Food Sci Technol, 11:87100.
Fiszman S.M., Lluch M.A., and Salvador A. 1999. Effect of addition of gelatin on
microstructure of acidic milk gels and yoghurt and on their rheological properties.
Int Dairy J, 9:895901.
Galani D. and Apenten R.K.O. 1999. Heat-induced denaturation and aggregation of
-Lactoglobulin: Kinetics of formation of hydrophobic and disulphide-linked
aggregates. Int J Food Sci Technol, 34:467476.
Garti N. and Sato, J. 2001. The roles of emulsifiers in fat crystallization. In:
Crystallization Processes in Fats and Lipid Systems, N. Garti and K. Sato, editors,
pp. 211250. Marcel Dekker: New York.
Goff H.D. 2002. Formation and stabilization of structure in ice cream and related
products. Cur Opin Colloid Interface Sci, 7:432437.
143
Hagolle N., Launay B., and Relkin P. 1998. Impact of structural changes and aggregation on adsorption kinetics of ovalbumin at acid and neutral pH. Colloids Surfaces
B Biointerfaces, 10:191198.
Haque Z., Kristjansson M.M., and Kinsella, J.E. 1987. Interaction between -casein
and -lactoglobulin: Possible mechanisms. J Agric Food Chem, 35:644649.
Harwalkar V.R. and Ma C.Y. 1990. Thermal Analysis of Foods. Elsevier Applied
Science Publications: England.
Hartel R.W. and Kaylegian K.E. 2001. Advances in milk fat crystallisation, technology, and applications. In: Crystallization Processes in Fats and Lipid Systems, N.
Garti and K. Sato, editors, p. 381. Marcel Dekker: New York.
Hindle S., Povey M.J.I., and Smith K. 2000. Kinetics of crystallization in n-hexadecane and cocoa butter oil-in-water emulsions accounting for droplet collisionmediated nucleation. J Colloid Interface Sci, 232:370380.
Kinsella J.E. and Whitehead D.M. 1989. Proteins in whey: Chemical, physical, and
functional properties. Adv Food Nutr Res, 33:343438.
Lefbvre J. and Relkin P. 1996. Denaturation of globular proteins in relation to their
functional properties. In: Surface Activity of Proteins, S. Magdassi, editor. Marcel
Dekker: New York.
Liu T., Relkin P., and Launay B. 1994. Thermal denaturation and heat-induced gelation of -lactoglobulin: Effects of some chemical parameters. Thermochim Acta,
246:387403.
Lrinczi D. 2004. The Nature of Biological Systems as Revealed by Thermal Analysis.
Kluwer Academic Publishers: London.
Lumry R., and Eyring H. 1954. Conformation changes of proteins. J Phys Chem,
58:110120.
Mills O.E. 1976. Effect of temperature on tryptophan fluorescence of -lactoglobulin.
Biochem Biophys Acta, 434:324332.
McClements D.J., Duncan S.R., German J.B., Simoneau C., and Kinsella J.E. 1993.
Droplet size and emulsifier type affect crystallization and melting of hydrocarbonin-water emulsions. J Food Sci, 58:11481151.
Morr C. and Ha E.Y.W. 1993. Whey protein concentrates and isolates: Processing
and functional properties. Food Sci and Nutr, 33:431476.
Morr C.V. and Ha E.Y.W. 1993. Whey protein concentrates and isolates: Processing
and functional properties. CRC Crit Rev Food Sci Nutr, 33:431476.
Mulvihill D.M. and Ennis M.P. 2003. Functional milk proteins: Production and utilization. In: Advanced Dairy Chemistry, Part B, Vol. 1, pp. 11751228, P. F. Fox and
P.L.H. McSweeney, editors. Kluwer Academic Publishers: New York.
Park K.H. and Lund D.B. 1984. Calorimetric study of thermal denaturation of lactoglobulin. J Dairy Sci, 67:16991706.
Pelan B.M.C., Watts K.M., Campbell I.J., and Lips A. 1997. The stability of aerated
milk protein emulsions in the presence of small molecule surfactants. J Dairy Sci,
80:2631.
Privalov P.L. and Potekin S.A. 1996. Scanning calorimetry in studying temperatureinduced changes in proteins. Methods Enzymol, 131:451.
144
Privalov P.L. 1979. Stability of proteins. Small globular proteins. Adv Protein Chem,
33:167241.
Relkin P. and Launay B. 1990. Concentration effects on the kinetics of -lactoglobulin
heat denaturation: A differential scanning calorimetric study. Food Hydrocolloids,
4:1932.
Relkin P. 1996. Thermal unfolding of -lactoglobulin, -lactalbumin and bovine
serum albumin. A thermodynamical approach. Crit Rev Food Sci Nutr,
36:565601.
Relkin P., Meylheuc T., Launay B., and Raynal K. 1998. Heat-induced gelation of
globular protein mixtures. A DSC and a SEM study. J Thermal Anal, 51:
747755.
Relkin P., Hagolle N., Dalgleish D.G., and Launay B. 1999. Foam formation and
stabilisation by pre-denatured ovalbumin. Colloids Surfaces B Biointerfaces, 12:
409416.
Relkin P., Sourdet S., and Fosseux P.Y. 2003. Fat crystallization in complex food
emulsions. Effects of adsorbed milk proteins and of a whipping process. J Thermal
Anal Cal, 71:187195.
Relkin P. 2004. Using DSC for monitoring protein conformation stability and effect
of fat droplets crystallinity in complex food emulsions. In: The Nature of Biological
Systems as Revealed by Thermal Analysis, D. Lrincz, editor, pp 99126. Kluwer
Academic Publishers: Londres.
Relkin P. and Sourdet S. 2005. Factors affecting fat droplets aggregation in whipped
frozen protein-stabilized emulsions. Food Hydrocolloids, 19:503511.
Relkin P., Bernard C., Meylheuc T., Vasseur J., and Courtois F. 2007. Production of
whey protein aggregates with controlled end-use properties. Le Lait,
87:337348.
Relkin P., Yung J.M., Kalnin D., and Ollivon M. 2008. Structural behaviour of lipid
droplets in protein-stabilized nano-emulsions and stability of -tocopherol, Food
Biophys, 3:163168.
Roos Y. H. 1995. Phase Transitions in Foods. Academic Press: London.
Ruegg M.P., Morr U., and Blanc B. 1987. A calorimetry study of the thermal denaturation of whey proteins in simulated milk ultrafiltrate. J Dairy Res,
44:509520.
Sanchez-Ruiz J.M. 1992. Theoretical analysis of Lumry-Eyring models in differential
scanning calorimetry. Biophys J, 61:921935.
Singh H. and Hevea P. 2003. Thermal denaturation, aggregation, and gelation of whey
proteins. In: Advanced Dairy Chemistry, Part B, Vol. 1, P.F. Fox and P.L.H.
McSweeney editors, pp. 12611287. Kluwer Academic Publishers: New York.
Skoda W. and van den Tempel M. 1963. Crystallization of emulsified triglycerides.
J Colloid Sci 18:568584.
Sourdet S., Relkin P., Aubry V., and Fosseux, P.Y. 2002. Composition of fat protein
layer in complex food emulsions at various weight ratios of casein-to-whey proteins, Le Lait, 82:567578.
Sourdet S., Relkin P., and Cesar B. 2003. Effects of milk protein type and pre-heating
on physical stability of whipped and frozen emulsions. Colloids Surfaces B,
31:5564.
145
Chapter 7
Analysis of Foodborne Bacteria by
Differential Scanning Calorimetry
Michael H. Tunick, John S. Novak, Darrell O. Bayles,
Jaesung Lee, and Gnl Kaletun
Introduction
C. perfringens and L. monocytogenes Analysis by DSC
Sample Preparations
C. perfringens Results
L. monocytogenes Results
Effect of Antibiotics on Bacteria
E. coli and Lactobacillus plantarum Analysis by DSC
Sample Preparations
E. coli and L. plantarum Results
Application of DSC for Evaluation of Food-Processing Treatments
Determination of Heat Inactivation Parameters of Bacteria from
Calorimetric Data
Determination of Efficacy of Nonthermal Treatments from
Calorimetric Data
Determination of Impact of Antimicrobials on Bacteria from
Calorimetric Data
Conclusions
References
147
149
149
150
152
153
155
156
156
158
158
161
163
164
164
Introduction
The World Health Organization estimates that 325,000 hospitalizations
and 5000 deaths result from foodborne illness in the United States each
147
148
149
erties of lipids in cell membranes of Mycoplasma laidlawii were investigated by Steim et al. (1969) with DSC, by using whole cells, cell
membranes, and extracted lipids. DSC thermograms of both isolated
cell membranes and extracted membrane lipids showed an endothermic transition around 40 C, suggesting extraction of lipids did not
change the stability. However, whole-cell thermogram did not exhibit
any distinguishable peaks (Bach and Chapman 1980).
The first successful DSC on whole cells was the study on heat inactivation and spontaneous germination of bacterial spores. Maeda and
colleagues (1974) observed that germinated Bacillus megaterium
spores had endothermic peaks at about 100 C and 130 C. For vegetative cells, Verrips and Kwast (1977) reported eight endothermic peaks
on the whole-cell thermogram of Citrobacter freundii.
It is necessary to obtain distinguishable and reproducible transitions
to identify the origin of the transitions and to examine their stability.
The resolution of peaks can be enhanced by increasing viable cell
density in the sample, by increasing sample size, and by improving the
sensitivity of DSC instrument. Recent studies showed that larger and
more distinguishable peaks can be obtained by using cell pellets instead
of cell suspensions and by using the cells at a late logarithmic growth
stage (Mackey et al. 1991; Lee and Kaletun 2002a,b).
This chapter focuses on the characterization of bacterial inactivation
by using DSC-relevant conditions on precooking, refrigerating, or
high-pressure processing of food to ensure its safety.
150
151
152
supermarket or by a consumer, caused the peaks of both the heatshocked and control samples to flatten and shift to lower temperatures
(Figure 7.1, curves D and E). The increased heat resistance of the heatshocked cells was lost, supporting the theory that this resistance is
transient (Heredia, Labb, and Garca-Alvarado 1998). Heat is uniformly distributed in a cell, resulting in damage to the most sensitive
molecules within it. The results suggest that conformational changes
in ribosomal proteins in response to temperature differences alter
protein synthesis in C. perfringens and that refrigeration will destroy
this organism in food. These conformational changes, which may
involve changing the shape and structure of the protein, are readily
discerned by evaluation of DSC scans.
L. monocytogenes Results
The DSC curve of L. monocytogenes cells (Figure 7.2A) exhibited
melting transitions at 67.5 0.4 C, corresponding to thermal denaturation of the 30S subunit, and at 73.4 0.1 C, corresponding to the
combined 50S subunit and 70S particle (Bayles et al. 2000). Cold
shocking the cells at 0 C for 3 h caused a shift in the 50S/70S peak
denaturation temperature to 72.1 0.5 C (Figure 7.2B). The position
153
of the 30S peak did not shift significantly. Similar results were observed
with a cold shock to 5 C. Peak shoulders observed around 81 C were
due to bacterial DNA (Miles, Mackey, and Parsons 1986), as with the
Clostridium samples. The results indicate that intracellular changes in
the ribosomes, such as an alteration in the association status of the 70S
particles, are correlated with changes in the thermal properties of L.
monocytogenes. The 30S and 50S subunits are more thermally labile
than the associated 70S particle, so any change that causes dissociation
of 70S would make the ribosome more sensitive to heat (Stephens and
Jones 1993).
Effect of Antibiotics on Bacteria
Certain antibiotics inhibit protein synthesis by selectively targeting
bacterial 70S ribosomes while leaving eukaryotic ribosomes unaffected (Weisblum and Davies 1968). The effects of seven antibiotics,
six active against the ribosome and one (rifampin) active against RNA
polymerase, were tested on the cells to determine whether the antibiotic treatment produced alterations in peak denaturation temperatures
corresponding to ribosomes or their subunits. Figure 7.3, curve A, is
Figure 7.3. DSC of L. monocytogenes cells treated with antibiotic. Curve A, control;
curve B, kanamycin-treated cells; curve C, tetracycline-treated cells.
154
the DSC curve of a control similar to that in Figure 7.2, curve A, and
Figure 7.3, curve B, is the curve of cells treated with kanamycin. The
50S/70S peak shifted from 73.3 0.1 C to 72.1 0.7 C, a shift
that was similar to the temperature reduction in cells that had been cold
shocked (Figure 7.2). Treatment with tetracycline removed the 30S
transition that had been observed around 67 C (Figure 7.3, curve C).
Thus, DSC analysis showed evidence of structural changes in the
ribosomal protein. Treatment with chloramphenicol, erythromycin,
puromycin, rifampin, or streptomycin produced results that were
similar to those of the control.
Cells were also cold shocked from 37 to 0 C for 3 h and then
thermally challenged at 60 C to determine thermal tolerance. Previous
research revealed that the D60 value of L. monocytogenes is 75.6 s
(Miller, Bayles, and Eblen 2000). Kanamycin and tetracycline, which
measurably altered the DSC curves of L. monocytogenes cells, were
the antibiotics that caused reductions in thermal tolerance; chloramphenicol, erythromycin, puromycin, rifampin, and streptomycin did
not alter the D60 values. Compared with the controls, kanamycin and
tetracycline each reduced the D60 value by 20 s. These 26% reductions
were approximately the same as those observed following cold shocks
of 37 0 C (Miller, Bayles, and Eblen 2000) and 37 5 C. The antibiotic treatment data indicate that ribosomal changes have a significant
impact on the thermal resistance of L. monocytogenes. Cold shock and
certain antibiotics alter the state and modify the structure of ribosomes,
as reflected by changes in the DSC curves. The results are probably
due to disassociation of the 30S subunits, which are more thermally
labile and more effectively denatured by heat.
Similar results were observed in Dr. Kaletuns laboratory when
erythromycin-treated E. coli cells were analyzed by DSC (Figure 7.4).
E. coli cells suspended in HEPES buffer were treated with erythromycin for 40 min. With increasing concentration of erythromycin, it
appears that major ribosomal transition shifts to a higher temperature
in comparison with the thermogram of untreated cells. Furthermore,
the shape of the peak changes and becomes less broad. Erythromycin
is known to bind the 50S of bacterial ribosome, blocking the exit of
the growing peptide chain, thus inhibiting the translocation of peptide.
It can be speculated that treatment with erythromycin removes the 50S
transition from the 50S/70S peak observed in the control cell thermo-
155
Figure 7.4. Thermograms of whole cells of E. coli treated with erythromycin, control
(thick dashes), 10 /ml erythromycin (thin dashes), 50 g/ml erythromycin (dots).
156
Sample Preparations
E. coli cells were grown in trypticase soy broth, and Lactobacillus
plantarum cells were grown in MRS broth at 37 C to late exponential
growth phase. The final concentration of cells in the medium was
1.3 0.1 109 cfu ml1 for E. coli and 9.0 0.1 108 cfu ml1 for L.
plantarum. The cells were harvested by centrifugation at 10,000 g for
10 min at 4 C. The supernatant was discarded and the pellets were
washed with sterile distilled water and centrifuged for a second time
before transferring into DSC crucibles.
A differential scanning calorimeter (DSC 111, Setaram, Lyon,
France) was used to record thermograms of microorganisms heated at
a 3 C min1. All DSC measurements were conducted using fluid-tight,
stainless steel crucibles. For each DSC run, the reference crucible was
filled with distilled water equivalent to the water content of the sample.
After heating in the DSC, samples were cooled rapidly by liquid nitrogen and rescanned to evaluate the reversibility of transitions. DSC
thermograms were corrected for differences in the empty crucibles by
subtracting an empty crucible baseline.
157
Figure 7.5. Thermograms of whole cells of E. coli (dashes) and L. plantarum (dots)
obtained by DSC (1 to 150 C with 3 C min1 heating rate). From Lee and Kaletun
(2002a).
158
159
Te
N0
z
Der
z
(7.1)
160
20
30
40
50
60
70
80
Figure 7.6. DSC thermogram for whole cells of E. coli K12 displaying curve baseline
used to determine the apparent enthalpy value. From Alpas et al. (2003).
N 0 H 0 H f
(7.2)
H H f 2.303
ln ln
=
T + ln
Te
H 0 H f
z
Der
z
(7.3)
161
obtained after heat treatment in the DSC and after isothermal treatment
displayed close agreement. This approach provides reproducible and
accurate results in a short time compared with the plate count technique
because the DSC approach eliminates the incubation time normally
used for plating, which might take 2 days or more.
Determination of Efficacy of Nonthermal Treatments from
Calorimetric Data
There is a growing interest in using techniques alternative to thermal
processing for food preservation to enhance safety and shelf life of
perishable foods (Hoover et al. 1989; Knorr 1993). Among nonthermal
treatment processes, high hydrostatic pressure (HHP) appears to be the
most promising technology. HHP processing has the advantage over
conventional heat treatments in that, while this technique is effective
in inactivation of nonspore-forming microorganisms, substantial food
quality retention can be retained by avoiding the destruction of small
molecular compounds such as vitamins.
It is reported that cell death increases as the level of the pressure
applied increases, implying that critical cellular activities or processes
have been irreversibly damaged (Hoover et al. 1989; Cheftel 1995).
However, the pressure tolerance varies among the species of bacteria
and even among the various strains of the same species (Styles et al.
1991; Patterson et al. 1995; Hauben et al. 1997; Alpas et al. 1999;
Benito et al. 1999).
Although DSC is a thermal analysis technique, it has been applied
to evaluate the impact of HHP processing on inactivation of bacteria
by comparing the pre- and postprocess thermograms (Niven, Miles,
and Mackey 1999; Alpas et al. 2003; Kaletun et al. 2004). The comparison of various final states as a function of various physical and
chemical factors, starting from the same initial state, makes it possible
to use DSC to predict the effectiveness of methods to inactivate
microorganisms.
Niven and colleagues (1999) demonstrated by DSC studies that cell
death due to high-pressure treatment may also be related to irreversible
ribosomal damage. Alpas et al. (2003) confirmed quantitatively that
cell viability decreases as the extent of ribosomal denaturation assessed
by calorimetry increases. The ribosomal denaturation was evaluated
by comparing the total apparent enthalpy of the control and pressure-
162
Bacteria
S. aureus 485
Control
S. aureus 485
345 MPa
S. aureus 765
Control
S. aureus 765
345 MPa
E. coli
O157:H7
933
Control
E. coli
O157:H7
933
275 MPa
E. coli
O157:H7
931
Control
E. coli
O157:H7
931
275 MPa
Apparent
enthalpy
(J/g wet
weight)
Fractional
reduction in
apparent
enthalpy
(H0
H)/H0
4.0
2.7
0.32
3.8
2.4
0.37
3.7
2.8
0.24
3.7
2.7
0.27
Viable cells
(cfu/ml)
Log reduction
in viability
log10(N/N0)
1.6 109
5.0 106
2.5
2.0 109
1.6 106
3.1
2.0 109
2.0 107
2.0
1.3 109
6.3 106
2.3
163
164
than in the cells treated with ethanol and salt. Differences also were
observed in the DSC profiles of bacterial cells treated with organic or
inorganic acid, suggesting that the mechanism of reduced thermal
tolerance of bacterial cells by these acids may be different. For design
of hurdle technology application in food processing, DSC studies in
vivo provide valuable information relevant to the effectiveness of
hurdles.
Conclusions
DSC is a valuable tool when investigating the effect of physical or
chemical treatments applied during food preservation on inactivation
of bacteria. Among the cellular components in a bacterial cell, the
damage to ribosomal proteins due to thermal, nonthermal, chemical,
or antibiotic treatments appears to be related to loss of cell viability.
DSC scans show that protein synthesis in C. perfringens and L. monocytogenes ribosomes is more efficiently destroyed during heating when
conformational changes and disassociation of the 30S subunits are
induced by temperature shocks. DSC thermograms display information
about the cellular components affected by various preservation treatments, thereby providing insight into the mechanism of bacterial inactivation. Furthermore, the calorimetric data can be analyzed to obtain
quantitative information about bacterial inactivation, including thermal
stability, thermal energy required for bacterial inactivation, and the
kinetic parameters of inactivation. Calorimetric data can be used to
optimize the processing conditions of food preservation in a rational
manner.
References
Adams, M. R., OBrien, P. J. and Taylor, G. T., 1989. Effect of ethanol content of
beer on the heat resistance of a spoilage Lactobacillus. J Appl Bacteriol,
66:491495.
Alpas, H., Kalchayanand, N., Bozoglu, F., Sikes, A., Dunne, C.P. and Ray, B., 1999.
Variation in resistance to hydrostatic pressure among strains of food-borne pathogens. Appl Environ Microbiol, 65(9):42484251.
Alpas, H., Lee, J., Bozoglu, F. and Kaletun, G. 2003. Differential scanning calorimetry of pressure-resistant and pressure-sensitive strains of Staphylococcus aureus
and Escherichia coli O157:H7. Int J Food Microbiol, 87:229237.
165
Anderson, W.A., Hedges, N.D., Jones, M.V. and Cole, M.B. 1991. Thermal inactivation of Listeria monocytogenes studied in differential scanning calorimetry. J Gen
Microbiol, 137:14191424.
Bach, D. and Chapman, D. 1980. Calorimetric studies of biomembranes and their
molecular components. In Biological microcalorimetry ed. Beezer, A.E. pp. 275
309. Academic Press: London.
Bayles, Darrell O., Tunick, Michael H., Foglia, Thomas A. and Miller, Arthur J. 2000.
Cold shock and its effect on ribosomes and thermal tolerance in Listeria monocytogenes. Appl Environ Microbiol, 66(10):43514355.
Benito, A., Ventoura, G., Casadei, M., Robinson, T. and Mackey, B. 1999. Variation
in resistance of natural isolates of Escherichia coli O157 to high hydrostatic pressure, mild heat, and other stresses. Appl Environ Microbiol, 65(4):15641569.
Borman, Stu 2007. Protein factory reveals its secrets. Chem Eng News, 85(8):
1316.
Cameron, M. S., Leonard, S. J. and Barret, E.L. 1980. Effect of moderately acidic
pH on heat resistance of Clostridium sporogenes spores in phosphate buffer and
in buffered pea puree. Appl Environ Microbiol, 39:943949.
Casadei, M. A., Ingram, I., Hitchings, E., Archer, J. and Gaze, J. E. 2001. Heat resistance of Bacillus cereus, Salmonella typhimurium and Lactobacillus delbrueckii
in relation to pH and ethanol. Int J Food Microbiol, 63:125134.
Cheftel, J.-C., 1995. High pressure, microbial inactivation and food preservation.
Food Sci Technol, 1:7590.
Hauben, K.J.A., Bartlett, D.H., Soontjens, C.C.F., Cornelis, K., Wuytack, E.Y. and
Michiels, C.W., 1997. Escherichia coli mutants resistant to inactivation by high
hydrostatic pressure. Appl Environ Microbiol, 63(3):945950.
Heredia, Norma L., Labb, Ronald G. and Garca-Alvarado, Jos Santos. 1998.
Alteration in sporulation, enterotoxin production, and protein synthesis by
Clostridium perfringens type A following heat shock. J Food Prot, 61(9):
11431147.
Hoover, D.G., Metrick, C., Papineau, A.M., Farkas, D.F. and Knorr, D., 1989.
Biological effects of high hydrostatic pressure on food microorganisms. Food
Technol, 43(3):99107.
Kaletun, G., Lee, J., Alpas, H. and Bozoglu, F. 2004. Evaluation of structural
changes induced by high hydrostatic pressure in Leuconostoc mesenteroides. Appl
Environ Microbiol, 70:11161122.
Knorr, D., 1993. Effect of high hydrostatic pressure processes on food safety and
quality. Food Technol, 47(6):156161.
Lee, J. and Kaletun, G. 2002a. Evaluation by differential scanning calorimetry of
the heat inactivation of Escherichia coli and Lactobacillus plantarum. Appl
Environ Microbiol, 68:53795386.
Lee, J. and Kaletun, G. 2002b. Calorimetric determination of inactivation parameters
of microorganisms. J Appl Microbiol, 93:178189.
Lee, J. and Kaletun, G. 2005. Evaluation by differential scanning calorimetry of
the effect of acid, ethanol, and NaCl on Escherichia coli. J Food Prot,
68:487493.
166
Leistner, L. 2000. Basic aspects of food preservation by hurdle technology. Int J Food
Microbiol, 55:181186.
Mackey, B.M., Miles, C.A., Parsons, S.E. and Seymour, D.A. 1991. Thermal denaturation of whole cells and cell components of Escherichia coli examined by
differential scanning calorimetry. J Gen Microbiol, 137 (10):23612374.
Mackey, B.M., Miles, C.A., Seymour, D.A. and Parsons, S.E. 1993. Thermal denaturation and loss of viability in Escherichia coli and Bacillus stearothermophilus.
Lett Appl Microbiol, 16:5658.
Mackey, B.M., Parsons, S.E., Miles, C.A. and Owen, R.J. 1988. The relationship
between base composition of bacterial DNA and its intracellular melting temperature as determined by differential scanning calorimetry. J Gen Microbiol, 134:
11851195.
Maeda, Y., Noguchi, S. and Koga, S. 1974. Differential scanning calorimetric study
of spontaneous germination of Bacillus megaterium spore by water vapor. J Gen
Microbiol, 20:1119.
Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin,
P.M. and Tauxe, R.V. 1999. Food-related illness and death in the United States.
Emerg Infect Dis, 5:607625.
Miles, C.A. and Mackey, B.M. 1994. A mathematical analysis of microbial inactivation at linearly rising temperatures: calculation of the temperature rise needed to
kill Listeria monocytogenes in different foods and methods for dynamic measurements of D and z values. J Appl Bacteriol, 77:1420.
Miles, C.A., Mackey, B.M. and Parsons, S.E. 1986. Differential scanning calorimetry
of Bacteria. J Gen Microbiol, 132(4):939952.
Miller, Arthur J., Bayles, Darrell O. and Eblen, B. Shawn. 2000. Cold shock inactivation of thermal sensitivity in Listeria monocytogenes. Appl Environ Microbiol,
66(10):43454350.
Mohacsi-Farkas, Cs., Farkas, J., Meszaros, L., Reichart, O. and Andrassy, E. 1999.
Thermal denaturation of bacterial cells examined by differential scanning calorimetry. J Therm Anal Calorim, 57:409414.
Niven, G.W., Miles, C.A., Mackey, B.M., 1999. The effects of hydrostatic pressure
on ribosome conformation in Escherichia coli: an in vivo study using differential
scanning calorimetry. Microbiol, 145:419425.
Novak, John S., Tunick, Michael H. and Juneja, Vijay K. 2001. Heat treatment adaptations in Clostridium perfringens vegetative cells. J Food Prot, 64(10):
15271534.
Patterson, M.F., Quinn, M., Simpson, R., Gilmore, A. 1995. Sensitivity of vegetative
pathogens to high hydrostatic pressure treatment in phosphate-buffered saline and
foods. J Food Prot, 58:524529.
Peleg, M. 1999. On calculating sterility in thermal and non-thermal preservation
methods. Food Res Int, 32:271278.
Reichart, O. 1979 A new experimental method for the determination of the heat
destruction parameters of microorganisms. Acta Alimentaria, 8:131155.
Steim, J.M., Tourtellotte, M.E., Reinert, J.C., McElhaney, R.N., Rader, R.L. 1969.
Calorimetric evidence for the liquid-crystalline state of lipids in a biomembrane.
Proc Nut Acad Sci, 63:104109.
167
Stephens, Peter J. and Jones, Martin V. 1993. Reduced ribosomal thermal denaturation in Listeria monocytogenes following osmotic and heat shocks. FEMS
Microbiol Lett, 106(2):177182.
Styles, M.F., Hoover, D.G., and Farkas, D.F., 1991. Response of Listeria monocytogenes and Vibrio parahaemolyticus to high hydrostatic pressure. J Food Sci, 56:
14041407.
Thompson, D.R., Willardsen, R.R., Busta, F.F and Allen, C.E. (1979a) Clostridium
perfringens population dynamics during constant and rising temperatures in beef.
J Food Sci, 44:646651.
Thompson, W.S., Busta, F.F., Thompson, D.R. and Allen, C.E. (1979b) Inactivation
of salmonellae in autoclaved ground beef exposed to constantly rising temperatures. J Food Prot, 42:410415.
Van Impe, J.F., Nicolai, B.M., Martens, T., De Baerdemaeker, J. and Vandewalle, J.
(1992) Dynamic mathematical model to predict microbial growth and inactivation
during food processing. Appl Environ Microbiol, 58:29012909.
Verrips, C.T. and Kwast, R.H. 1977. Heat resistance of Citrobacter freundii in media
with various water activities. Eur J Appl Microbiol, 4:225231.
Weisblum, Bernard and Davies, Julian 1968. Antibiotic inhibitors of the bacterial
ribosome. Bacteriol Rev, 32(4):493528.
World Health Organization (WHO). 2007. Food safety and foodborne illness. Fact
Sheet No. 237. WHO, Geneva, Switzerland.
Chapter 8
Coupling of Differential Scanning
Calorimetry and X-Ray Diffraction to
Study the Crystallization Properties and
Polymorphism of Triacylglycerols
Christelle Lopez, Daniel J.E. Kalnin, and Michel R. Ollivon*
Introduction
Thermal and Crystallographic Properties of Triacylglycerols
Polymorphism of Triacylglycerols
Differential Scanning Calorimetry
X-Ray Diffraction
Coupling of XRD and DSC: MICROCALIX
Applications and Results
Cocoa Butter and Its Components
Milk Fat
Lard
Conclusion
References
169
170
170
173
175
176
179
179
184
190
193
194
Introduction
Lipids from vegetable or animal origins are widely consumed in food
products, for example, in chocolate, shortenings, margarine, and butter.
The composition of triacylglycerols (TG), which are the main constituents of natural fats and oils or hydrogenated and interesterified fats
(i.e., such as in margarine), their suprastructure, and the physical prop*This chapter is dedicated to Michel Ollivon who passed away on June 16th, 2007.
169
170
171
Table 8.1. Melting point of the main fatty acids found in natural fats and oils
and of triacylglycerols.
Fatty Acids
Formula
C4:0
C6:0
C8:0
C10:0
C12:0
C14:0
C16:0
C18:0
C18:1 9c
C18:2 9c, 12c
C18:3 9c,
12c, 15c
Triacylglycerols
Name
(Abbreviation)
Melting
Point (C)
TG
Abbreviation
Melting
Point (C)
Butyric acid
(B)
Caproic acid
Caprylic acid
Capric acid
Lauric acid
(L)
Myristic acid
(M)
Palmitic acid
(P)
Stearic acid
(St)
Oleic acid (O)
Linoleic acid
Linolenic acid
BBB
75
4
16
31
44
OOO
StOO
StStO
StOSt
5
24
38
44
54
PPO
34
63
POP
36
70
LLL
47
16
5
14
MMM
PPP
StStSt
58
66
73
172
Table 8.2. Some crystallographic and energetic properties of the three main
polymorphic forms of a selection of TGs.
Polymorphic Form
Property of TG*
Main short
spacings ()
Melting point
(C) of StStSt
Enthalpy of
fusion (J g1) of
StStSt
Melting point
(C) of OOO
Melting point
(C) of LLL
Melting point
(C) of POP
Hexagonal
4.15
Orthorhombic
Perpendicular
Triclinic Parallel
4.6
55
64
72
163
180
230
32
12
15
35
46
21
30
36
*TG, triacylglycerols; St, stearic acid; O, oleic acid; L, lauric acid; P, palmitic acid.
Adapted from Mulder and Walstra 1974.
have been described in detail (Small 1986; Ollivon and Perron 1992):
hexagonal ( form), orthorhombic perpendicular ( form), and triclinic parallel ( form) (Figure 8.1B). The density, enthalpy of fusion,
melting point, and stability increase in the order , , and , according
to the monotropic character of the polymorphism. In the form, the
lateral packing of the fatty acid chains is not very tight and the chains
have considerable rotational freedom, whereas in the form, the chains
are very densely packed. TG crystals are made by the stacking of TG
molecules layers, the thickness of which depends on the length and
unsaturation of the fatty acid chains and their angle of tilt with respect
to the basal planes formed by the methyl end groups of the TG (Figure
8.1C). The longitudinal organization of TG in lamellar structures is
primarily related to the number of chains stacked in the crystalline cell.
For TG in natural fats, the number of fatty acid chains frequently
observed is two or three and corresponds to the stacking of double
(2L)- or triple (3L)-chain length lamellar structures (Small 1986;
Hagemann 1988). Roughly, 3L forms are usually related to lowmelting, long-chain monounsaturated and mixed long- and short-chain
173
Figure 8.1. Main types of triacylglycerol (TG) packings. (A) Lamellar structure
formed by TG molecules in the solid state: for example, form of trilaurin. (B) Left:
The stable conformation of the hydrocarbon chains of saturated fatty acid (FA) is a
planar zigzag shown here as a 3D view along its main axis. Right: Three main types
of lateral chain packings (only carbon atoms are drawn): hexagonal, orthorhombic
perpendicular (O + upside down T), and triclinic parallel (T//) subcell in the order of
their stability, which are named , , and for TG. (C) Two main types of TG
longitudinal chain stacking (fatty acids are drawn as straight lines), 2L and 3L.
174
bles, which may be opened or hermetically sealed, of 10100 l capacity. Because sample size is small, accurate weighing is essential for
quantification of the thermal properties of fats. The reference is often
an empty pan, so as to not exhibit thermally induced transitions within
the temperature range of interest. Transitions that occur in the sample
during the applied temperature program appear as peaks or troughs,
depending on whether they are exothermic or endothermic, on the plot
of differential heat flow versus temperature or time (the thermogram)
that is the output of the instrument. Conventionally, the temperature
program used in DSC is a linear change in temperature with time, with
various cooling and heating rates. After heating a lipid sample to 20 C
over its final melting point to erase its thermal history/memory (Ollivon
and Perron 1992), different rates of cooling permit the investigation of
the crystallization properties of TG, polymorph formation, and transformation. Cooling and heating can also be performed from a temperature at which the fat is partially crystallized (e.g., 4 C for milk fat).
Isothermal DSC, in which the temperature is kept constant at a value at
which a transition of interest is known to occur, is especially useful for
studying the polymorphic evolutions as a function of time (Lopez et al.
2002a). The crystallization properties of fats depend on their thermal
history and on the rates of cooling and heating applied. A combination
of cooling, heating, and isothermal DSC scans can be used to study
polymorphism of TG and fats in complex products.
DSC is a useful tool (1) to record the crystallization and melting
profiles recorded on cooling and heating, respectively; (2) to determine
the characteristic temperatures, such as the temperatures of initial
crystallization (Tonset) and final melting (Toffset); (3) to monitor polymorphic evolutions and measure the heat of transitions; and (4) to quantify
the solid fat content that is proportional to the enthalpy of melting (H)
of fat.
The effects of different rates of cooling and heating on polymorph
formation and polymorphic transitions recorded as a function of temperature or time are easily studied by DSC. DSC studies have given
an insight into the thermodynamics of fat phase transition in bulk and
in emulsions. However, because the complex DSC recordings are often
difficult to interpret and it is not possible to identify unequivocally
polymorphs using DSC, the experiments must be coupled with other
techniques, such as XRD or other techniques yielding structural information. Moreover, DSC allows the characterization of the physical
175
state changes if a change of energy is involved. However, this technique does not provide information on the structure that exists before
and after the phase transition.
X-Ray Diffraction
XRD is a powerful technique to use to provide structural information.
As explained above, characteristic lengths of structures formed by TG
range from atomic distances up to hundreds of ngstroms. The whole
size range can be investigated by means of X-ray scattering. Generally,
X-ray scattering reflects periodical differences of the electron density
within a sample. Thus, X-rays are the ideal direct probe for determining the internal structure of crystalline material because they provide
information on the repetitive patterns of the electron density of the
array of atoms. In their solid state, TG molecules are arranged periodically in planes at repetitive distance d, which can be identified using
XRD (Figure 8.1A).
Wide-angle X-ray scattering (WAXS) monitors the structure at
atomic scale (from about 1 to 10 ). It provides information on intraand intermolecular distances called short spacings. For crystallized
TG, the hydrocarbon chains are arranged in regularly spaced planes.
Crystallographic planes give rise to a reflection line at a distinct angle
, satisfying the well-known Bragg relation: 2d sin = n, where is
the X-ray wavelength, d is the repetitive distance between planes, n is
an integer, and is half the angle between the incident and diffracted
beam (Guinier 1964; Small 1986).
The actual data registered during an XRD experiment is the scattered intensity as a function of 2. It is often convenient to use the
scattering vector q instead of the scattering angle 2 because the former
is independent of the wavelength of the incident beam. They are related
as follows: q = (4/) sin . Thus, the distance d between planes can
be deduced from the position q of the diffraction peak by d = 2/q.
The different packing of aliphatic chains in the three crystalline TG
subcells leads to characteristic wide-angle X-ray reflections enabling
their identification. Short spacings are widely used for identifying the
various crystalline subcells characterizing the polymorphic forms
(Figure 8.1B). A single line around 4.15 characterizes the form
(hexagonal subcell). A strong line at 4.6 and two small lines at about
3.85 and 3.7 identifies the form (triclinic parallel subcell), whereas
176
177
Figure 8.2. Experimental setup of the microcalorimeter MICROCALIX in the timeresolved synchrotron XRD environment. (A) Schematic representation: The cell is
positioned with the capillary containing the sample perpendicular to the beam in such
a way that the diffraction patterns are recorded in the vertical plane by one or two
one-dimensional proportional detectors (LD) at small and wide angles. Counting
electronic (Counting Elect.), nanovoltmeter (nVmeter), and temperature controller
(T Ctrl) are monitored by a single computer. The temperature-controlled cryostat
(TCC) is kept at constant temperature (e.g., 6 C). (B) Setup on the D22 bench
of synchrotron (LURE, Orsay, France). (C) Setup on the D24 bench of synchrotron
(LURE, Orsay, France).
thermal scans in the temperature range 30 C to +230 C, with scanning rates between 0.01 and 10 C/min and with sensitivity comparable with that of a modern commercial apparatus (>100 V/mW).
Scanning temperature is controlled with a resolution of 0.01 C, and
the microcalorimeter is calibrated with lauric acid.
MICROCALIX was used on synchrotron radiation X-ray benches.
The lastest version of the instrument has been adapted for laboratory
bench and conventional source, but it is preferably used with rotating
anode or multilayered mirrors.
MICROCALIX inserted in a laboratory conventional X-ray source
The microcalorimeter can be installed on a laboratory source designed
for simultaneous SAXS and WAXS measurements, such as that of
CNRS group at UMR 8612 in Chtenay-Malabry (France; Lopez et al.
178
179
lected. Note that the time resolution is also limited by the heat conductivity of the sample. If the high flux of synchrotron radiation source
is used to follow very fast transitions, very good heat transfer to a small
sample must be ensured.
Experiments using MICROCALIX have been carried out at three
synchrotron radiation sources, Laboratoire pour lUtilisation du
Rayonnement Electromagntique (Orsay, France), European
Synchrotron Radiation Facility (Grenoble, France), and Elettra (Trieste,
Italy); further experiments are planned at SOLEIL (Saclay, France).
180
Figure 8.3. Summary of the possible arrangements of cocoa butter. Upper left shows
a table resuming the notation, melting point, lateral and longitudinal packing of TGs.
An overlay of two DSC recordings shows the closeness of two thermal events, notably
the occurrence for V and form VI. Underneath is the XRD pattern of what is believed
to be the pure crystalline species of forms V to VI.
181
182
unstable
form
a (3L) 75
intermediate
form
b2 (2L) 3941 30C
183
liquid
35C
50
3
2
10
40
30
20
10
0
0
1000
0
3000
2000
Temperature (C)
21C
15
Long Spacings ()
stable form
b1 (3L ?) 6466
42
3
1
32
0
10
time (sec)
3
2
1
0.2
q (1)
30
40
50
endo
50
45
40
35
30
25
20
15
10
5
0
60
45
40
35
30
25
20
15
10
5
0
0.1
20
temperature (C)
0.3
50
Temperature (C)
37
1.4
1.6
q (1)
1.8
40
3
2
30
20
10
0
100 50
50
100
184
185
186
187
DSC and XRD investigations showed that the fat phase of dairy
products displays a complex polymorphism. Depending on the cooling
rate, six different types of crystals were identified, several of them in
coexistence, and their time- and temperature-dependent evolutions
were quantitatively monitored. They correspond to lamellar structures
with 2L (40.548 ) and 3L (5472 ) organizations of TG. At least
five crystalline subcell species were observed at wide angles: and
sub-, two , and one . All these crystalline structures coexist with
a liquid phase even at low temperature (T < 4 C). Thermal events
recorded by DSC were related to the structural information on the
organization of TG obtained by XRD. These experiments focusing on
the crystallization properties of whole milk fat and fat fractions characterized by different composition and thermal properties contributed
to the development of spreadable butters with defined solid fat content
as a function of temperature.
188
2L1
46.5
3L2
65
0.15C/min
3L2(002)
3L1(003)
10
20
Temperature (C)
30
3L1(005)
16
27
T (C)
22C
Liquid
16
37
26
47
0.1
0.2
0.3
q (1)
0.4
0.5
T (C)
3L1(001) 2L2
71.3 40
Wide-angle XRD
3L1(002)
12C 20C
DSC
189
36
1.1
1.3
1.5
1.7
q (1)
Figure 8.7. Structural evolution, expressed in scattering vector q (1), of TGs dispersed in milk fat globules during cooling from 60 C to 8 C at 0.15 C.min1.
Three-dimensional plots of the XRD patterns recorded at small angles (left) and at
wide angles (right). DSC curve recorded simultaneously.
190
e4 e3
5.0
e2
0.2 ( 12)
4.0
0.5 ( 5)
3.0
1 ( 2.5)
2.0
2 ( 2.5)
endo
cooling
1.0
c2
c3
Tonset (c3)
0.0
5 ( 2)
c1
Tonset (c1)
Tonset (c2)
10
cooling rate (rc) in K/min
1.0
40
20
0
20
Temperature (C)
40
60
Figure 8.8. Characteristic DSC curves of the crystallization of lard. The influence of
the cooling rate (rc) is evidenced in the range of 0.2 up to 10 K/min as indicated.
The normalized heat flow is scaled to the cooling rate of 10 K/min. DSC curves are
shifted relatively to each other and multiplied for clarity as indicated in brackets.
When not associated to Tonset, arrows indicate minor exothermic events.
Table 8.3. Main crystallographic parameters of the fat crystals in lard and
their attribution to major DSC peaks.
Crystalline Form/
Subcell
(hexagonal)
1
(orthorhombic)
2
(pseudoorthorhombic
triclinic)
SAXS
Peaks d ()
WAXS
Peaks d ()
DSC
Exotherms
48.2
35.1
4.18
4.14, 3.83
c1
c2
h3
h4, h5
43.8
3.95, 4.4,
and 4.3
c3
h7, h6
43.8
4.6
c2 (only
upon slow
cooling
rates)
h1, h2
191
DSC
Endotherms
40.0C
500
34.6C
40.0C
34.4C
29.2C
18.9C
29.2C
400
18.9C
8.4C
l (cps)
300
Heating
at rH = 5 K/min
c2
U-shaped
200
43.6
c2
35.1
100
c1
48.2
0
0.05
0.10
0.15
4.14
c3
20.5C
Crystallization
at rC = 5 K/min 15.9C
12.7C
4.8C
8.9C
13.9C
19.2C
0.20
0.25
0.30
c1
4.18
1.10
1.20
1.30
1.40
1.50
8.4C
WAXS
c2
q (1)
c2
= 3.83
1.60
20.5C
12.7C
4.8C
8.9C
13.9C
19.2C
1.70
1.80
q (1)
60
DSC
SAXS
WAXS
h 2 + h1
40
h3
h4
h5
20
Temperature (C)
c3
3.95 = 3.79
c3
rH = 5 K/min
h6
h7
20
0
c3
rc = 5 K/min
c1
c2
20
48.2
43.8
35.1
endo
40
4.15
3.8
3.95.1 + 4.3 + 4.4
60
6
0
Heat Flow (mW)
1.0
0.5
0.0 1.0
0.5
Relative Peak Intensity (%)
0.0
Figure 8.9. A selection of small and wide angle X-ray scattering (SAXS) patterns at
the average temperature indicated is redrawn on top to illustrate crystallographic
properties of lard. Pure tristearin (SSS) WAXS pattern is shown for comparison
(dashed line). All X-ray patterns were recorded for 60 s and shifted relatively to each
other for clarity. The corresponding DSC curves are drawn for comparison next to
SAXS and WAXS relative peak intensity plots vs. T. Evolution of three SAXS and
five WAXS lines were followed and plotted with normalizations on the peak maximum
intensity (figure redrawn from Kalnin et al. 2005). Main crystallographic parameter
of the crystals in lard and their attribution to major DSC peaks (right).
192
193
Conclusion
Elucidating the polymorphism of lipids is important to better understand the properties of fat-rich food products and to improve food
technology.
The presence of one or several long fatty acid chains and their different packings lead to a variety of polymorphic forms that cause the
thermal and structural behaviors of lipids to be rather complex to study.
Both types of properties are currently measured by XRD and DSC
194
References
Adenier H., Ollivon M., Perron R., and Chaveron H. 1975. Le blanchiment gras. I.
Observations et commentaries. Chocolaterie Confiserie France, 315:714.
Ali M.A.R. and Dimick P.S. 1994. Thermal analysis of palm mid-fraction, cocoa
butter, and milk fat blends by differential scanning calorimetry. J Am Oil Chem
Soc, 71:299302.
Arishima T., Sagi N., Mori H., and Sato K. 1991. Polymorphism of POS. I. Occurence
and polymorphic transformation. J Am Oil Chem Soc, 68:710715.
Blanton T.N., Barnes C.L., and Lelental M. 2000. Preparation of silver behenate
coatings to provide low- to mid-angle diffraction calibration. J Appl Crystogr,
33:172173.
Chaiseri S. and Dimick P.S. 1995a. Dynamic crystallization of cocoa butter. I.
Characterization of simple lipids in rapid- and slow-nucleation cocoa butters and
their seed crystals. J Am Oil Chem Soc, 72:14911496.
Chaiseri S. and Dimick P.S. 1995b. Dynamic crystallization of cocoa butter. II.
Morphological, thermal and chemical characteristics during crystal growth. J Am
Oil Chem Soc, 72:14971504.
Chapman G.M., Akehurst E.E., and Wright W.B. 1971. Cocoa butter and confectionery fats. Studies using programmed temperature x-ray diffraction and differential
scanning calorimetry. J Am Oil Chem Soc, 48:824830.
Davis T.R. and Dimick P.S. 1986. Solidification of cocoa butter. Proc PMCA Prod
Conf, 40:104108.
Davis T.R. and Dimick P.S. 1989a. Crystals formed during cocoa butter solidification.
J Am Oil Chem Soc, 66:14881493.
Davis T.R. and Dimick P.S. 1989b. Lipid composition of high-melting seed
crystals formed during cocoa butter Solidification. J Am Oil Chem Soc, 66:
14941498.
195
Famelart M.H., Le Graet Y., Michel F., Richoux R., and Riaublanc A. 2002. Evaluation
des mthodes dapprciation des proprits fonctionnelles des fromages
dEmmental de louest de la France. Lait, 82:225245.
Guinier A. 1964. Thorie et Technique de la Cristallographie, 3rd edition. Dunod:
Paris.
Hagemann J.W. 1988. Thermal behaviour and polymorphism of acylglycerides. In:
Crystallisation and Polymorphism of Fats and Fatty Acids, Garti N. and Sato K.,
editors, pp. 99. Marcel Dekker: New-York.
Huyghebaert A. and Hendrickx H. 1971. Polymorphism of cocoa butter, shown by
differential scanning calorimetry. Lebensm-Wiss U Technol, 4:5963.
Kalnin D., Garnaud G., Amenitsch H. and Ollivon M. 2002. Monitoring fat crystallization in aerated food emulsions by combined DSC and time-resolved synchrotron x-ray diffraction. Food Res Int, 35:927934.
Kalnin D., Lesieur P., Artzner F., Keller G., and Ollivon M. 2005. Systematic investigation of lard polymorphism using combined DSC and time-resolved synchrotron x-ray diffraction. Eur J Lipid Sci Technol, 107:594606.
Kunutsor S.K. and Ollivon M. 1983. Ternary phase diagram of stable forms of
major triglycerides of cocoa butter (POP, POS, SOS). 16th ISF research, World
Congress: Budapest.
Lavigne F. 1995. Polymorphisme et transitions de phases des triglycerides.
Applications aux proprits thermiques et structurales de la matire grasse laitire
anhydre et ses fractions. PhD thesis, Univ Paris VII, Paris XI and ENSIA, France.
Loisel C., Keller G., Lecq G., Launay B., and Ollivon M. 1997a. Tempering of
chocolate in a scraped surface heat exchanger. J Food Sci, 62:773780.
Loisel C., Lecq G., Ponchel G., Keller G., and Ollivon M. 1997b. Fat bloom and
chocolate structure studied by mercury porosimetry. J Food Sci, 62:781788.
Loisel C., Keller G., Lecq G., Bourgaux C., and Ollivon M. 1998a. Phase transitions
and polymorphism of cocoa butter. J Am Oil Chem Soc, 75:425439.
Loisel C., Lecq G., Keller G., and Ollivon M. 1998b. Dynamic crystallization of
dark chocolate as affected by temperature and lipid additives. J Food Sci,
63:7379.
Lopez C., Lesieur P., Keller G., and Ollivon M. 2000. Thermal and structural behavior
of milk fat: 1. Unstable species of cream. J Colloid Interface Sci, 229:6271.
Lopez C., Lavigne F., Lesieur P., Keller G., and Ollivon M. 2001a. Thermal and
structural behavior of milk fat: 1. Unstable species of anhydrous milk fat. J Dairy
Sci, 84:756766.
Lopez C., Lavigne F., Lesieur P., Keller G., and Ollivon M. 2001b. Thermal and
structural behavior of anhydrous milk fat: 2. Crystalline forms obtained by slow
cooling. J Dairy Sci, 84:24022412.
Lopez C., Lesieur P., Bourgaux C., Keller G., and Ollivon M. 2001c. Thermal and
structural behavior of milk fat: 2. Crystalline forms obtained by slow cooling of
cream. J Colloid Interface Sci, 240:150161.
Lopez C., Bourgaux C., Lesieur P., Bernadou S., Keller G., and Ollivon M. 2002b.
Thermal and structural behavior of milk fat: 3. Influence of cooling rate and droplet
size on cream crystallisation. J Colloid Interface Sci, 254:6478.
196
Lopez C., Bourgaux C., Lesieur P., and Ollivon M. 2002a. Crystalline structures
formed in cream and anhydrous milk fat at 4 C. Lait, 82:317335.
Lopez C., Lesieur P., Bourgaux C., and Ollivon M. 2005. Thermal and structural
behavior of anhydrous milk fat. 3. Influence of cooling rate. J Dairy Sci,
88:511526.
Lopez C., Briard-Bion V., Camier B., and Gassi J.Y. 2006a. Milk fat thermal properties and solid fat content in Emmental cheese: A differential scanning calorimetry
study. J Dairy Sci, 89:28942910.
Lopez C., Bourgaux C., Lesieur P., Riaublanc A., and Ollivon M. 2006b. Milk fat
and primary fractions obtained by dry fractionation 1. Chemical composition and
crystallisation properties. Chem Phys Lipids, 144:1733.
Lopez C., Bourgaux C., Lesieur P., and Ollivon M. 2007. Coupling of time-resolved
synchrotron x-ray diffraction and DSC to elucidate the crystallisation properties
and polymorphism of triglycerides in milk fat globules. Lait, 87:459480.
Lopez C., Briard-Bion V., Beaucher E., and Ollivon M. 2008. Multiscale characterization of the organization of triglycerides and phospholipids in Emmental cheese:
From the microscopic to the molecular level. J Agric Food Chem,
56:24062414.
Lovegren N.V., Gline M.S., and Feuge R.O. 1976. Polymorphic changes in mixtures
of confectionery fats. J Am Oil Chem Soc, 53:8388.
Marangoni A.G. and McGauley S.E. 2003. Relationship between crystallization
behavior and structure in cocoa butter. Crystal Growth & Design, 3:95108.
Merken G.V. and Vaeck S.V. 1980. Etude du polymorphisme du beurre de cacao par
calorimtrie DSC. Lebensm-Wiss U Technol, 13:314317.
Michalski M.C., Ollivon M., Briard V., Leconte N., and Lopez C. 2004. Native fat
globules of different sizes selected from raw milk: Thermal and structural behaviour. Chem Phys Lipids, 132:247261.
Mulder H. and Walstra P. 1974. The milk fat globule. In: Emulsion Science as Applied
to Milk Products and Comparable Foods. Commonwealth Agricultural Bureauxz:
Farnham Royal, Bucks, UK.
Ollivon M. and Perron R. 1992. Proprits physiques des corps gras. In: Manuel des
Corps Gras, Karleskind A., Wolff J.P., and Guttman J.F., editors, pp. 433442.
Lavoisier: Paris.
Ollivon M., Keller G., Bourgaux C., Kalnin D., Villeneuve P., and Lesieur P. 2006.
DSC and high resolution x-ray diffraction coupling. J Therm Anal Calorim,
85:219224.
Rowney M., Roupas P., Hickey M., and Everett D.W. 1998. Milkfat structure and
free oil in Mozzarella cheese. Aust J Dairy Technol, 53:110.
Sato K. 1987. Physical and molecular properties of lipid polymorphs: A review. Food
Microstruct, 6:151159.
Sato K., Arishima T., Wang Z.H., Ojima K., Sagi N., and Mori H. 1989. Polymorphism
of POP and SOS. I. Occurence and polymorphic transformation. J Am Oil Chem
Soc, 66:664674.
Small D.M. 1986. In: Handbook of lipid research. The physical chemistry of lipids.
From alkanes to phospholipids. Plenum Press: New York.
197
ten Grotenhuis E., van Aken G.A., van Malssen K.F., and Schenck H. 1999.
Polymorphism of milk fat studied by differential scanning calorimetry and realtime x-ray powder diffraction. J Am Oil Chem Soc, 76:10311039.
Timms R.E. 1980. The phase behavior and polymorphism of milk fat, milk fat fractions, and fully hardened milk fat. Aust J Dairy Technol, 35:4753.
Timms R.E. 1984. Phase behavior of fats and their mixtures. Prog Lipid Res,
23:138.
van den Tempel M., 1979. Crystallization in dispersed systems. In: Physico-chimie
des composes amphiphiles, R. Perron R., P. Bothorel P., editors, pp. 261264.
Colloques nationaux du C.N.R.S. n 938.
van Malssen K.F. 1994. Real-time x-ray powder diffraction applied to cocoa butter
and graphite intercalates. Ph.D., Amsterdam University, The Netherlands.
Wille R.L. and Lutton E.S. 1966. Polymorphism of cocoa butter. J Am Oil Chem Soc,
43:4914496.
Part 2
Calorimetry as a Tool for Process Design
Chapter 9
Overview of Calorimetry as a Tool for
Efficient and Safe Food-Processing Design
Alois Raemy, Corinne Appolonia Nouzille, Pierre Lambelet,
and Alejandro Marabi
Introduction
Generalities About Thermal Analysis and Calorimetry
Techniques
Methods
Samples
Thermal Behavior of Food Constituents
Carbohydrates (sugars)
Lipids
Proteins
Water
Thermal behavior of Raw and Reconstituted Food
Safety Aspects
Other Thermodynamic Parameters
Heat of Solution
Specific Heat
Heat of Combustion
Related Techniques
Interest of Calorimetry for the Food Industry
Conclusion
References
201
202
203
203
205
206
206
206
208
214
216
217
217
218
218
224
225
225
226
226
227
202
Introduction
In the investigation of foods using thermal analysis and calorimetric
techniques, many physicochemical effects can be observed in the temperature range between 50 C and 300 C. These thermal phenomena
may be either endothermic, such as melting, denaturation, and vaporization, or exothermic, such as crystallization, oxidation, and fermentation. Some exothermic reactions present a hazard in industrial
operations or during storage. They can lead either to self-ignition,
causing fires or dust explosions in open systems such as spray-dryers,
or to pressure increase and bursting in closed vessels such as autoclaves or extraction cells. Glass transitions are observed as a shift in
the baseline; this information, associated with humidity and water
activity determination, is of particular interest in relation to storage of
food powders, but also for gas retention in powders predicted to foam
when dissolved. This is the safe-processing aspect of thermal analysis
and calorimetry.
The thermal behavior of food strongly depends on its composition.
We therefore consider primarily thermal characteristics of the major
food constituents: carbohydrates, lipids, proteins, and water. Specific
thermal phenomena of minor constituents (e.g., caffeine), as well as
those of additives such as emulsifiers, are mentioned. Raw and reconstituted foods and finally interactions between food constituents will
be considered. Some of these aspects will only be mentioned and not
fully discussed. The reader should refer to previous work by the same
authors (Raemy and Lambelet 1991; Raemy et al. 2000, 2004).
Emulsifiers (endogenous or exogenous) are often used in the food
industry to stabilize interfaces in emulsions and foams. When added
to an aqueous phase, emulsifiers often spontaneously form self-assembly structures. Such structured fluids can be used as active ingredients
for encapsulation or as microreactors for flavor formation. In this
context, differential scanning calorimetry (DSC) instruments, especially micro-DSC, can help detect liquid crystal phase transitions and
establish phase diagrams. This is the material science aspect of thermal
analysis and calorimetry.
Other thermodynamic parameters, such as heat of solution, specific
heat, and heat of combustion can be determined; they are also important for efficient food-processing design. Here, we focus on heat of a
solution that is of great interest for food powder dissolution.
203
204
a
Mixing rod
Liquid reservoir
Aluminum
membrane
Recipient for solid
sample
205
206
207
55
50
45
dQ/dt [mW]
40
35
30
25
20
15
Aw 0.078
Aw 0.112
Aw 0.176
Aw 0.225
10
5
20
40
60
Temperature [C]
80
100
208
1C/min
Exo
Galactose
f
Sucrose
f
d
Cellobiose
f
50
100
150
200
250
Temperature, C
Figure 9.3. Calorimetric curves of crystalline galactose, sucrose, and cellobiose, all
three heated in sealed cells up to 260 C. Calorimeter Setaram C80, 1 C/min. From
Raemy and Schweizer (1983), with permission.
209
210
Exo
VI
dH/dt (W/g)
2.5
2
1.5
II
III
IV
0.5
0
15
20
25
30
T (C)
35
40
45
Figure 9.4. DSC heating curves of five polymorphs of cocoa butter. Mettler FP900,
5 C/min. From Rousset (1997).
1972; Lambelet and Raemy 1983; Ali and Dimick 1994; Culot 1994;
Elisabettini et al. 1998; Koyano et al. 1992; Rousset et al 1998; Timms
1994).
Kinetics of crystallization Crystallization kinetics of bulk lipids and
the formation and stability of their various polymorphs as a function
of time and temperature are another domain in which DSC is very
useful. For these measurements, the lipid sample has first to be heated
to at least 20 C above the melting temperature of its stable polymorph
to erase all memory effects. Kinetic information has been obtained by
measuring either isothermally after quenching at the desired temperature (Rousset and Rappaz 1997; Metin and Hartel 1998; Toro-Vazquez
et al. 2005) or at constant cooling rate under various cooling conditions
(Kawamura 1980; Cebula and Smith 1991). Complex thermal paths
such as tempering stages also were studied by calorimetry to understand precisely the mechanisms that induce the appearance of stable
crystalline forms (Rousset and Rappaz 1997).
Precise kinetic parameters can be determined from isothermal
experiments. The variation of SFC as a function of time can be obtained
by sequential integration of the crystallization peak. This SFC function
is then used to estimate crystallization parameters with the help of the
Avrami or more complex models (Kloek et al. 2000; Foubert et al.
2002; Rousset 2002). Nucleation induction times, which are periods
211
of time before nucleation appears, also can be determined from isothermal crystallization experiments (Rousset and Rappaz 2001). These
kinetic parameters are useful indicators of the nucleation rate and may
serve as a basis for predicting how to crystallize lipids in the desired
form.
Kinetic studies also help to understand the effects of compositional
changes (TAGs or other minor components) on crystallization (Garti
et al. 1988; Wahnelt et al. 1991; Tan et al. 2000; Vanhoutte et al.
2002a,b). However, as samples are not mixed, results from DSC crystallization studies are often difficult to interpret directly in terms of
process operating conditions (Rousset and Rappaz 2001; Hartel 2001).
DSC curves are often complex because various modes of crystallization and solid state transformations can contribute to a single peak. To
define unambiguously the key signal characteristics such as minima
and maxima, as well as end and start points, a method based on the
determination of the first and the second derivative of the DSC raw
signal has been proposed (Bouzidi et al. 2005).
Quality control DSC crystallization and melting profiles of lipids
have been used to assess the quality of oils, in particular of heated oils
(Gloria and Aguilera 1998; Tan and Man 1999, 2000, 2002). Similarly,
contamination (adulteration) of fats and fat-based products can be
detected by calorimetry (Lambelet and Ganguli 1983; Bringer et al.
1991; Marikkar et al. 2002). Adulteration was demonstrated by the
appearance or a modification (shift in the peak position and peak area)
of a thermal transition occuring in the heating or cooling DSC curves
of lipid mixtures.
Oxidative stability of oils and fats Lipid oxidation is an exothermic
phenomenon that can be observed by continuous monitoring of total
thermal effect either under isothermal or nonisothermal conditions of
measurements (Raemy et al. 1987; Kowalski 1989; Tan and Man 2002;
Ulkowski et al. 2005). Measurements can be performed under a static
air atmosphere or, preferably, under oxygen flow or oxygen pressure
(Litwinienko and Kasprzycka-Guttman 1998).
In isothermal experiments, oxidation induction times correspond to
the time at which a rapid exothermic reaction between the lipid and
oxygen occurred. Tables of oxidation induction times measured by
isothermal heat flux calorimetry around 100 C are reported in the
212
literature (Raemy et al. 1987). For edible oils, induction times obtained
by calorimetry were shown to correlate well with corresponding values
determined by traditional methods (Tan et al. 2002). DSC can therefore
be used to assess the oxidative stability of lipids (Raemy et al. 1987;
Kowalski 1989; Tan and Man 2002).
In the nonisothermal mode, Arrhenius kinetic parameters can be
deduced from the shape of the oxidation curves, and these parameters
can in turn be applied for calculation of the overall first-order rate
constant of oxidation at various temperatures (Litwinienko and
Kasprzycka-Guttman 1998; Ulkowski et al. 2005).
Antioxidant efficacy Food antioxidant activity can be measured by
calorimetry. The efficiency of an antioxidant to protect an oil is measured by the increase of induction time after incorporation of the test
antioxidant into the oil. (Raemy et al. 1987; Irwandi et al. 2000; Tan
et al. 2002; Giuffrida et al. 2006). A good correlation between DSC
oxidative induction time and oxidative stability index determined by
other analytical techniques was found (Tan et al. 2002; Gouveia et al.
2006; Giuffrida et al. 2006).
Also, the radical scavenging activities of antioxidants can be investigated by DSC monitoring of the polymerization of substrates initiated
by radical reactions (Fujisawa and Kadoma 2006).
Dispersed systems
Emulsifier-water systems Lipid-water systems that can be regarded
as models of the lipid matrix of cell membranes include emulsifiers
that may exhibit highly ordered self-assembly structures, which are
liquid crystalline phases. DSC has been applied to the study of endothermic phase transitions occurring in lipid-water systems (Blume 1991;
Figure 9.5. (a) Presentation of the calorimetric curve of a saturated MAG with 20%
water showing melting of different crystalline forms up to 70 C and weak liquid
crystal transitions at 85 C and 110 C. (b) Presentation of the following cooling curve,
which shows that there is practically no hysteresis between the temperatures of the
phenomena; however, the crystalline form melting at 45 C has disappeared.
(c) Presentation of the second heating curve, which confirms the reversible character
of most transitions. (ac) Setaram Micro-DSC III, 0.2 C/min. From Raemy et al.
(2005), with permission.
Heat flow/mW
Exo
(a)
0.5
0.5
1.5
2.5
3.5
4.5
5.5
6.5
7.5
liquid crystal
phase transitions
melting
10 20 30 40 50 60 70 80 90100
Temperature (C)
(b)
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
crystallization
liquid crystal
phase transitions
10 20 30 40 50 60 70 80 90 100
Temperature (C)
Exo
(c)
0.5
melting
liquid crystal
phase transitions
Heat flow/mW
1.5
2.5
3.5
4.5
5.5
6.5
7.5
10 20 30 40 50 60 70 80 90 100
Temperature (C)
213
214
Brandenburg et al. 2006) when they transform from the gel to the liquid
crystal phase (Chapman et al. 1974; Tlgyesi et al. 1985) and to determine thermotropic and lyotropic behavior of these systems (Briggs et
al. 1996; Qiu and Caffrey 1999). Both gel to liquid crystalline
acyl chain melting transitions (Mellier 1988) and structural transitions
between different three-dimensional structures (Willumeit et al. 2005)
were associated with transitions observed in DSC curves. The most
pronounced enthalpy changes were observed for acyl chain melting,
whereas structural transitions, which were not accompanied by acyl
chain melting, exhibited much lower enthalpy changes.
Micro-DSC can be used at low heating and cooling rates to detect
liquid crystalline transitions of exogenous emulsifier-water systems
(Figure 9.5a, b, and c).
Similarly, DSC has been applied to investigate the thermal behavior
of several emulsifier-water systems modified by changing the pH
value, the ionic composition of the environment, or by chemical agents
(Tlgyesi et al. 1985; Forte et al. 1998; Fournier et al. 1998).
Emulsions DSC is useful as it is sufficiently sensitive to measure
transformations in dispersed phases, in particular when used simultaneously with synchrotron X-ray diffraction (XRD; Kalnin et al. 2002).
Thermal behavior of lipids in a dispersion or emulsified form has been
shown to be quite different from that of the same fat in bulk, for
example, crystallization of milk fat (Lopez et al. 2002). Lipid polymorphism in dispersed systems can also be investigated by calorimetry. As shown in Figure 9.6, polymorphism of colloidal suspensions
of TAG has been determined based on DSC melting transitions (Bunjes
et al. 2007).
Proteins
The main phenomena observed by DSC, and especially micro-DSC,
during heating of protein solutions are endothermic phenomena associated with protein denaturation. These phenomena were first described
by Privalov (Privalov and Khechinashvili 1974). These phenomena
were then also observed for products containing large amounts of
proteins and enough water to allow protein mobility; for example, for
proteins of dairy products where small exotherms associated to protein
aggregation also can be detected (Unterhaslberger 2006), fish and meat
Heat flow
2 mW
Heating
Cooling
S100 / SGC
10
20
30
40
50
60
70
80
Heat flow
Temperature (C)
2 mW
Heating
Cooling
S100-3 / SGC
10
20
30
40
50
60
70
80
Temperature (C)
Figure 9.6. DSC heating (10 C/min) and cooling (5 C/min) curves of tristearin
nanoparticles stabilized with phospholipid/bile salt blends containing unmodified
(top) or hydrogenated soybean lecithin (bottom) shortly after preparation. The dashed
arrow in the bottom panel indicates the thermal range of the exothermic event prior
to crystallization. Pyris 1 from Perkin Elmer. Endothermic is upward. From Bunjes
et al. (2007), with permission.
215
216
91.9C
54.9C
0.5
Endo
0.5
Denaturation Enthalpy: 1.82 J/g
1
1.5
2
2.5
3
10
20
30
40
50
60
Temperature [C]
70
80
90
100
Figure 9.7. DSC curve of the ovalbumin fraction of egg (first minus second run).
Setaram Micro-DSC III, 1 C/min. From Ferreira et al. (1997), with permission.
(Harwalkar and Ma 1990), egg (as shown in Figure 9.7; Ferreira et al.
1997; Grinberg et al. 2002), and cereals (Ellepola and Ma 2006).
Denaturation can be quantified by the determination of the denaturation enthalpy. As the native protein sample is considered to be not
denatured at all, its enthalpy in a first scan corresponds to 100% of
denaturation. Thermally processed products can then be considered as
partially or totally denatured, and a percentage of denaturation can thus
be calculated from the determined enthalpies.
The pH is important because it influences denaturation temperature
and thus allows finding protective conditions for industrially treated
proteins. Protein denaturation can also be partially or even totally
reversible according to the results of second runs.
For dry proteins, glass transition and oxidation also can be observed
with thermal analysis techniques.
Water
Crystallization (undercooling) of water, melting of ice, and vaporization can be observed with thermal analysis and calorimetry. Since the
enthalpies corresponding to these phenomena are quite high (333 J/g
for ice melting and 2255 J/g for water vaporization), they can be
217
Safety Aspects
Carbohydrate decomposition, which sometimes immediately follows
melting, lipid oxidation (especially if oil is present as a layer or at the
surface of the product) as well as protein oxidation, and even Maillard
reactions may present a hazard in industrial operations (e.g., roasting,
high-temperature drying).
The role of thermal analysis and calorimetry for determining safe
conditions of industrial processes has already been explained elsewhere (Raemy and Lliger 1985; Raemy et al. 1985; Raemy and
Gardiol 1987; Raemy 1988, 2001). The application of adiabatic
calorimetry to the study of cellulose decomposition, cellulose being
218
219
220
221
0.16
Fat content
0.7%
14.3%
29.3%
35.7%
45.0%
0.14
0.12
0.10
0.08
0.06
0.04
EXO
0.02
0.00
0.02
500
1000
1500
2000
2500
Time [s]
Figure 9.8. Typical dissolution calorimetry curves of a model food powder with
increasing amounts of fat. From Marabi et al. (2008), with permission.
the contact angle between the particulate solid and the liquid (Spagnolo
et al. 1996; Adamson and Gast 1997). The application of immersion
calorimetry could therefore circumvent the difficulties associated with
assessing contact angles of food powders. Several reviews on the use
of isothermal calorimetry in different (e.g., pharmaceutical, microbiological) applications are available (Buckton 1995; Wads 1997), indicating the wide applicability and useful information that can be obtained
from this technique, which surprisingly, has not been exploited in the
field of food science.
The effects of fat content on the dissolution enthalpy and kinetics
of a model food powder were reported previously (Marabi et al. 2008).
Typical dissolution calorimetry curves for five freeze-dried samples
with increasing fat contents are shown in Figure 9.8.
Dissolution of all the powders resulted in an exothermic response.
Increasing the amount of fat in the samples is clearly related to a
decrease in the amount of heat released during the dissolution process.
The decrease in the enthalpy of dissolution ranged from 61.9 to 32.4 J/g
for samples with 0.7% and 45.0% of fat, respectively (Figure 9.9).
This, in turn, is related to the dissolution kinetics of the powders, which
also was shown to be significantly affected by increasing amounts of
fat in the samples (Marabi et al. 2008). However, when the enthalpy
of dissolution is normalized by the amount of fat material, it can be
222
14.3
FAT (%)
29.3
35.7
45.0
0
10
20
Hdiss (J/g)
30
40
50
60
70
Enthalpy of dissolution normalized by fat content
80
90
Figure 9.9. Enthalpy of dissolution of a model food powder with increasing amounts
of fat expressed as J/g of sample and as J/g of nonfat material.
seen that very similar values are obtained for all the samples tested
(Figure 9.9). This fact indicates that the fat has a minimal contribution
to the overall enthalpy measured, which is in agreement with the slight
endothermic response measured when pure fat is mixed with water
(Marabi et al. 2008).
The effects of different moisture contents and the physical state of
maltodextrin (MD) and skim milk powder (SMP) also were studied
(Marabi et al. 2007b). Namely, three different conditions were studied:
after samples were freeze-dried (FD), after equilibration at 54.4% relative humidity, and after FD (water activity again less than 0.01) of the
equilibrated sample. The calorimetric curves are shown in Figure 9.10a
and b. The dissolution was found to be exothermic for all the tested
samples, and the curves showed similar shapes. For both samples, a
clear decrease in the response was observed when the FD and the
equilibrated samples are compared. When the latter were FD again,
the MD resulted in a curve almost identical to that obtained with the
original FD samples. In contrast, the SMP sample showed only an
intermediate response between those of the FD and the equilibrated
samples. The enthalpy of dissolution decreased (became less exothermic) about 12- and 20-fold from the FD state compared with the
equilibrated state for the MD and SMP samples, respectively.
0.16
223
a
MD DE21 - FD
MD DE21 - aW 0.54
MD DE21 - aW 0.54 & FD
0.14
0.12
0.10
0.08
0.06
0.04
Exo
0.02
0.00
500
1000
1500
2000
2500
3000
Time [s]
0.14
b
SMP - FD
SMP - aW 0.54
SMP - aW 0.54 & FD
0.12
0.10
0.08
0.06
0.04
0.02
Exo
0.00
0
500
1000
1500
2000
2500
3000
Time [s]
Figure 9.10. Typical dissolution calorimetry curves of the maltodextrin DE21 (a) and
skim milk powder (b) at different conditions. All the samples showed an exothermic
response; however, increased moisture content or recrystallized lactose led to a less
exothermic dissolution process. From Marabi et al. (2007b), used with permission.
224
225
Heat of Combustion
During burning of a food product, a large amount of energy is liberated. Values determined with calorimetric bombs are about 39 kJ g1
for fat, 23 kJ g1 for protein, and 17 kJ g1 for carbohydrate.
Related Techniques
Some related thermal analysis techniques, such as dynamical mechanical analysis (DMA) or dynamical mechanical thermal analysis (DMTA)
give rheological rather than thermal information. Also, thermogravimetry, sometimes coupled with gas analysis instruments, can provide a
better interpretation of calorimetric curves (e.g., when studying
hydrated carbohydrates) by indicating weight losses associated with
the observed thermal phenomena.
Microscopy techniques, sometimes performed at different temperatures with a hot stage microscope, are often also very helpful in obtaining a clear interpretation of calorimetric curves.
Concerning lipids, as assignments of DSC signals may be ambiguous due to the high number of thermal events, calorimetry often needs
to be combined with XRD (Chung and Caffrey 1992; Keller et al.
1996) or even synchrotron XRD when transformations are rapid.
Recent experiments combining DSC and synchrotron XRD have revolutionized the study of lipid crystallization (Ollivon et al. 2001; Kalnin
et al. 2002; Lopez et al. 2002). DSC can also be used simultaneously
with microscopy to identify morphologies associated with polymorphs
(Rousset et al. 1998).
Fat crystallization in oil-in-water emulsions has been followed by
DSC in combination with either nuclear magnetic resonance spectroscopy (zilgen et al. 1993) or XRD (Awad et al. 2001; Ollivon et al.
2001).
226
227
Even so, new products, new conditions, and new processes push
performance of more calorimetric measurements and testing of new
instrumentations, such as isothermal titration calorimetry (ITC), which
is of interest for studying such interactions as that between proteins
and active substances, or dielectric spectroscopy, which is of interest
in the study of proteins or the amorphous state of carbohydrates.
A new trend is to use coupled systems for performing evolved gas
analysis in relation to the thermal signal; for example, thermogravimetry coupled with mass spectrometry is commonly used today.
Thermal analysis and the study of foods have been mutually beneficial: thermal analysis and calorimetry in general have brought much
information to food science and processing technology, whereas the
study of foods, as the samples are easily available, has allowed the
development and the promotion of many calorimetric techniques.
References
Adamson A.W. and Gast A.P. 1997. Physical Chemistry of Surfaces, 6th edition. John
Wiley & Sons: New York.
Ali A.R.M. and Dimick P.S. 1994. Thermal analysis of palm mid-fraction, cocoa
butter, and milk fat blends by DSC. J Am Oil Chem Soc, 71:299302.
Arishima T., Sagi N., Mori H., and Sato K. 1991. Polymorphism of POS. I. Occurrence
and polymorphic transformation. J Am Oil Chem Soc, 68:710715.
Awad T., Hamada Y., and Sato K. 2001. Effects of addition of diacylglycerols on fat
crystallization in oil-in-water emulsion. Eur J Lipid Sci Technol, 103:735741.
Bartsch A., Schuff P., and Bning-Pfaue H. 1990. Investigations on the compatibility
of fatsintroduced by the example of blends of a lauric fat and milk fat fractions.
Fett Wiss Technol, 92:(6)213221.
Bastos M., Milheiras S., and Bai G.Y. 2004. Enthalpy of solution of alphacyclodextrin in water and in formamide at 298.15K. Thermochim Acta,
420:111117.
Bhaskar A.R., Rizvi S.S.H., Bertoli C., Fay L.B., and Hug B. 1998. A comparison of
physical and chemical properties of milk fat fractions obtained by two processing
technologies. J Am Oil Chem Soc, 75:12491264.
Blanchard J.M.V. and Lillford P.J. 1993. The Glassy State in Foods. Nottingham
University Press: Nottingham, U.K.
Blume A. 1991. Biological calorimetry: Membranes. Thermochim Acta,
193:299347.
Bouzidi L., Boodhoo M., Humphrey Kerry L., and Narine S.S. 2005. Use of first and
second derivatives to accurately determine key parameters of DSC thermographs
in lipid crystallization studies. Thermochim Acta, 439:12, 94102.
228
Brandenburg K., Garidel P., Howe J., Andrae J., Hawkins L., Koch M.H.J., and
Seydel U. 2006. What can calorimetry tell us about changes of three-dimensional
aggregate structures of phospholipids and glycolipids? Thermochim Acta,
445:(2)133143.
Briggs J., Chung H., and Caffrey M. 1996. The temperaturecomposition phase
diagram and mesophase structure characterization of the monoolein/water system,
J Phys II, France, 6:723751.
Bringer R., Rudzik L., Weber T., and Wst E. 1991. Detection of foreign fat in milk
fat by means of differential calorimetry. Milchwissenschaft, 46:304307.
Buckton G. 1995. Applications of isothermal microcalorimetry in the pharmaceutical
sciences. Thermochim Acta, 248:117129.
Bunjes H., Steiniger F., and Richter W. 2007. Triglyceride nanoparticles in different
crystal modifications. Langmuir 23:(7)40054011.
Cebula D.J. and Smith K.W. 1991. Differential Scanning Calorimetry of confectionery fats. Pure triglycerides: Effects of cooling and heating rate variation. J Am Oil
Chem Soc, 68:591595.
Chapman D., Urbina J., and Keough K.M. 1974. Biomembranes phase transitions. J
Biol Chem, 249:25122521.
Chung H. and Caffrey M. 1992. Direct correlation of structure changes and thermal
events in hydrated lipid established by simultaneous calorimetry and time-resolved
x-ray diffraction. Biophysical J, 63:438447.
Claudy P. 2005. Analyse calorimtrique diffrentielle. Lavoisier: Paris.
Conti S., Gaisford S., Buckton G., and Cooke U. 2006. Solution calorimetry to
monitor swelling and dissolution of polymers and polymer blends. Thermochim
Acta, 450:5660.
Culot C. 1994. Modlisation du Comportement Polymorphique des Triglycrides.
Thesis: Universit Notre Dame de la Paix, Namur.
Daniels R.L., Kim H.J., and Min D.B. 2006. Hydrogenation and interesterification
effects on the oxidative stability and melting point of soybean oil. J Agric Food
Chem, 54:(16)60116015.
Dian N.L.H.M., Sundram K., and Idris N.A. 2006. DSC study on the melting properties of palm oil, sunflower oil, and palm kernel olein blends before and after
chemical interesterification. J Am Oil Chem Soc, 83:(8)739745.
Dilworth S.E., Buckton G., Gaisford S., and Ramos, R. 2004. Approaches to determine the enthalpy of crystallization, and amorphous content, of lactose from
isothermal calorimetric data. Int J Pharm, 284:8394.
Dimick P.S. and Manning D.M. 1987. Thermal and compositional properties of cocoa
butter during static crystallization. J Am Oil Chem Soc, 64:16631669.
Elisabettini P., Lognay G., Desmedt A., Culot C., Istasse N., Deffense E., and Durant
F. 1998. Synthesis and physicochemical characterization of mixed diacid triglycerides that contain elaidic acid. J Am Oil Chem Soc, 75:285291.
Ellepola S.W. and Ma C.Y. 2006. Thermal properties of globulin from rice (Oryza
sativa) seeds. Food Res Int, 39:(3)257264.
Ferreira M., Hofer C., and Raemy A. 1997. A calorimetric study of egg white proteins.
J Thermal Anal, 48:683690.
229
Forte L., Andrieux K., Keller G., Grabielle-Madelmont C., Lesieur S., Paternostre
M., Ollivon M., Bourgaux C., and Lesieur P. 1998. Sodium taurocholate-induced
lamellar-micellar phase transitions of DPPC determined by DSC and x-ray diffraction. J Therm Anal Calorim, 51:773782.
Foubert I., Vanrolleghem P.A., Vanhoutte B., and Dewettinck K. 2002. Dynamic
mathematical model of the crystallization kinetics of fats. Food Res Int,
35:945956.
Fournier I., Barwicz J., and Tancrde P. 1998. The structuring effects of amphotericin
B on pure and ergosterol- or cholesterol-containing dipalmitoylphosphatidylcholine bilayers: A differential scanning calorimetry study. Biochim Biophys Acta,
1373:7686.
Fujisawa S. and Kadoma Y. 2006. Comparative study of the alkyl and peroxy radical
scavenging activities of polyphenols. Chemosphere, 62(1):7179.
Gao D. and Rytting J.H. 2006. Use of solution calorimetry to determine the extent of
crystallinity of drugs and excipients. Int J Pharm, 151:183192.
Garti N. and Sato K. 1988. Crystallization and Polymorphism of Fats and Fatty Acids.
Marcel Dekker: New York.
Garti N., Schlichter J., and Sarig S. 1988. DSC studies concerning polymorphism of
saturated monoacid triglycerides in the presence of food emulsifiers. Fett Wiss
Technol, 90:295299.
Giron D. 1995. Thermal analysis and calorimetric methods in the characterisation of
polymorphs and solvates. Thermochim Acta, 248:159.
Giron D., Mutz M., and Garnier S. 2004. Solid-state of pharmaceutical compounds
Impact of the ICH Q6 guideline on industrial development. J Therm Anal Calorim,
77:709747.
Giuffrida F., Destaillats F., Egart M.H., Hug B., Golay P.-A., Skibsted L.H., and
Dionisi, F. 2006. Activity and thermal stability of antioxidants by DSC and ESR
spectroscopy. Food Chem, 101:(3)11081114.
Gloria H. and Aguilera J.M. 1998. Assessment of the quality of heated oils by
Differential Scanning Calorimetry (DCS). J Agric Food Chem, 46:13631368.
Gouveia A.F., Duarte C., Beirao da Costa M.L., Bernardo-Gil M.G., and MoldaoMartins, M. 2006. Oxidative stability of olive oil flavoured by Casicum frutescens
supercritical fluid extracts. Eur J Lipid Sci Technol, 108:(5)421428.
Grinberg V.Y., Grinberg N.V., Mashkevich A.Y., Burova T.V., and Tolstoguzov V.B.
2002. Calorimetric study of interaction of ovalbumin with vanillin. Food
Hydrocolloids, 16:333343.
Haines P.J., editor. 2002. Principle of Thermal Analysis and Calorimetry. Royal
Society of Chemistry: Cambridge, UK.
Hancock B.C. 2002. Disordered drug delivery: Destiny, dynamics, and the Deborah
number. J Pharm Pharmacol, 54:737746.
Hancock B.C. and Dalton C.R. 1999. The effect of temperature on water vapor sorption by some amorphous pharmaceutical sugars. Pharm Devel Technol,
4:125131.
Hancock B.C. and Parks M. 2000. What is the true solubility advantage for amorphous
pharmaceuticals? Pharm Res, 17:397404.
230
Hancock B.C. and Shamblin S.L. 1998. Water vapour sorption by pharmaceutical
sugars. Pharm Sci Technol Today, 1:345351.
Hancock B.C. and Zografi G. 1997. Characteristics and significance of the amorphous
state in pharmaceutical systems. J Pharm Sci, 86:112.
Haque M.K. and Roos Y.H. 2006. Differences in the physical state and thermal
behavior of spray-dried and freeze-dried lactose and lactose/protein mixtures.
Innovative Food Sci Emer Technol, 7:6273.
Harjunen P., Lehto V.P., Koivisto M., Levonen E., Paronen P., and Jarvinen K. 2004.
Determination of amorphous content of lactose samples by solution calorimetry.
Drug De Ind Pharm, 30:809815.
Hartel R.W. 2001. Crystallization in Foods, pp. 3490. Aspen Publishers: Gaithersburg,
USA.
Harwalkar V.R. and Ma C.Y. 1990. Thermal Analysis of Foods. Elsevier Applied
Sciences: London.
Hemminger W. and Hhne G. 1984. Calorimetry: Fundamentals and Practice. Verlag
Chemie: Weinhein, Germany.
Hemminger W. und Cammenga H.K. 1989. Methoden der Thermischen Analyse.
Springer Verlag: Berlin, Germany.
Herrera M.L. and Anon M.C. 1991. Crystalline fractionation of hydrogenated sunflower seed oil. II. Differential scanning calorimetry (DSC). J Am Oil Chem Soc,
68(11):799803.
Hogan S.E. and Buckton G. 2000. The quantification of small degrees of disorder in
lactose using solution calorimetry. Int J Pharm, 207:5764.
Huyghebaert A. and Hendrickx H. 1971. Polymorphism of cocoa butter, shown by
Differential Scanning Calorimetry. Lebensm-Wiss U Technol, 4:5963.
Irwandi J., Man Y.B., Kitts D.D., Bakar J., and Jinap S. 2000. Synergies between
plant antioxidant blends in preventing peroxidation reactions in models and food
oil systems. J Am Oil Chem Soc, 77:945950.
Kaisersberger E. 1989. DSC investigations of the thermal characterization of edible
fats and oils. Thermochim Acta, 151:8390.
Kaletun G. 2007. Prediction of heat capacity of cereal flours: A quantitative empirical correlation. J Food Eng, 82(2):589594.
Kalnin D., Garnaud G., Amenitsch H., and Ollivon M. 2002. Monitoring fat crystallization in aerated food emulsions by combined DSC and time-resolved synchrotron x-ray diffraction. Food Res Int, 35:925934.
Katainen E., Niemela P., Harjunen P., Suhonen J., and Jarvinen K. 2005. Evaluation
of the amorphous content of lactose by solution calorimetry and Raman spectroscopy. Talanta, 68:15.
Kawamura K. 1980. The DSC thermal analysis of crystallization behavior in palm
oil. II. J Am Oil Chem Soc, 57:4852.
Keller G., Lavigne F., Loisel C., Ollivon M., and Bourgaux C. 1996. Investigation
of the complex thermal behavior of fats. J Thermal Anal, 47:15451565.
Kloek W., Walstra P., and van Vliet T. 2000. Crystallization kinetics of fully hydrogenated palm oil in sunflower oil mixtures. J Am Oil Chem Soc, 77:389398.
231
232
Marabi A., Raemy A., Bauwens I., Burbidge A.S., Wallach R., and Saguy I.S. 2008.
Effect of fat content on the dissolution enthalpy and kinetics of a model food
powder. J Food Eng, 85(4):518527.
Marikkar J.M.N., Lai O.M., Ghazali H.M., and Che Man Y.B. 2002. Compositional
and thermal analysis of RBD palm oil adulterated with lipase-catalyzed interesterified lard. Food Chem, 76:249258.
Marti E., Kaisersberger E., and Emmerich W.D. 2004. New aspects of thermal analysis. J Therm Anal Calorim, 77:905934.
Marti E., Kaisersberger E., and Moukhina E. 2006. Heat capacity functions of polystyrene in glassy and in liquid amorphous state and glass transition. J Therm Anal
Calorim, 85:(2)505525.
Medout-Marere V., Malandrini H., Zoungrana T., Douillard J.M., and Partyka S.
1998. Thermodynamic investigation of surface of minerals. J Pet Sci Eng,
20:223231.
Mellier A. 1988. Infrared study of phospholipid hydration. Chem Phys Lipids,
46:5158.
Merken G.V. and Vaeck S.V. 1980. Study of polymorphism of cocoa butter by
Differential Scanning Calorimetry. Lebensm-Wiss U Technol, 13:314317.
Metin S. and Hartel R.W. 1998. Thermal analysis of isothermal crystallization kinetics in blends of cocoa butter with milk fat or milk fat fractions. J Am Oil Chem
Soc, 75:16171624.
Miller B. 1982. Thermal analysis. John Wiley & Sons: New York.
Miller D.P. and de Pablo J.J. 2000. Calorimetric solution properties of simple saccharides and their significance for the stabilization of biological structure and
function. J Phys Chem, B104:88768883.
Miller D.P., de Pablo J.J., and Corti H. 1997. Thermophysical properties of trehalose
and its concentrated aqueous solutions. Pharm Res, 14:578590.
Minato A., Ueno S., Yano J., Smith K., Seto H., Amemiya Y., and Sato K. 1997.
Thermal and structural properties of sn-1,3-dipalmitoyl-2-oleoylglycerol and sn1,3-dioleoyl-2-palmitoylglycerol binary mixtures examined with synchrotron
radiation X-ray diffraction. J Am Oil Chem Soc, 74:12131220.
Miyagawa K., Ogawa I., and Yamano H. 1995a. Calorimetric measurements on the
swelling of green tea. Thermochim Acta, 257:1319.
Miyagawa K., Ogawa I., and Yamano H. 1995b. Calorimetric measurements on the
swelling of seaweed. Thermochim Acta, 257:7582.
Mohsenin, N.N. 1980. Thermal properties of foods and agricultural materials.
Gordon and Breach: New York.
Morgan F., Appolonia-Nouzille C., Baechler, R., Vuataz G., and Raemy A. 2005.
Lactose crystallization and early Maillard reaction in skim milk powder and whey
protein concentrates. Le Lait, 85:315323.
Ollivon M., Loisel C., Lopez C., Lesieur P., Artzner F., and Keller G. 2001. Simultaneous
examination of structural and thermal behaviors of fats by coupled X-ray diffraction
and Differential Scanning Calorimetry techniques: Application to cocoa butter
polymorphism. In: Crystallization and Solidification Properties of Lipids, N.
Widlak, R. Hartel and S. S. Narine, editors, pp. 3441. AOCS Press: Champaign.
233
zilgen S., Simoneau C., German J.B., McCarthy M.J., and Reid, D.S. 1993.
Crystallization kinetics of emulsified triglycerides. J Sci Food Agric,
61:101108.
Pinto S.S. and Diogo H.P. 2006. Thermochemical study of two anhydrous polymorphs of caffeine. J Chem Thermodyn, 38:15151522.
Privalov P.L. and Khechinashvili N.N. 1974. A thermodynamic approach to the
problem of stabilization of globular protein structure: A calorimetric study. J Mol
Biol 86:665684.
Pyda M. 2002. Conformational heat capacity of interacting systems of polymer and
water. Macromolecules, 35:40094016.
Qiu H. and Caffrey M. 1999. Phase behavior of the monoerucin/water system. Chem
Phys Lipids, 100:5579.
Raemy A. 1981. Differential thermal analysis and heat flow calorimetry of coffee and
chicory products. Thermochim Acta, 43:229236.
Raemy A. 1988. Une mthodologie dinvestigation des ractions exothermiques, de
lauto-inflammation et de lexplosion de poussires adapte aux produits alimentaires, pp. C3.1C3.3. C.A.T.: Lille, France.
Raemy A. 1992. From thermal analysis to safety science. J Thermal Analysis,
38:437443.
Raemy A. 2001. La mesure des ractions exothermiques des aliments par analyse
thermique diffrentielle sous pression et calorimtrie diffrentielle programme.
In: Calorimtrie et Analyse Thermique, AFCAT, editor, pp. 6364. Hammamet,
TN.
Raemy A. and Lambelet P. 1982. A calorimetric study of self-heating in coffee and
chicory. J Food Technol, 17:451460.
Raemy A. and Leliger J. 1982. Thermal behaviour of cereals studied by heat flow
calorimetry. Cereal Chem, 59:189191.
Raemy A. and Ottaway M. 1991. The use of high pressure DTA, heat flow, and
adiabatic calorimetry to study exothermic reactions. J Thermal Anal, 37:
19651971.
Raemy A. and Schweizer T.F. 1983. Thermal behaviour of carbohydrates studied by
heat flow calorimetry, J Thermal Anal, 28:95108.
Raemy A. and Gardiol M. 1987. Paramtres thermodynamiques et scurit des oprations industrielles. Association Scientifique Internationale du Caf (ASIC), 12e
Colloque, Montreux (CH), pp. 320330.
Raemy A. and Lliger J. 1985. Self-ignition of powders studied by high pressure
differential thermal analysis. Thermochim Acta, 85:343346.
Raemy A. and Lambelet P. 1991. Thermal behaviour of foods. Thermochim Acta,
193:417439.
Raemy A., Appolonia-Nouzille C., Frossard P., Sagalowicz L., and Leser M.E. 2005.
Thermal behaviour of emulsifier-water systems studied by micro-DSC. J Therm
Anal Calorim, 80:439443.
Raemy A., Frlicher I., and Leliger J. 1987. Oxidation of lipids studied by isothermal
heat flux calorimetry. Thermochim Acta, 114:159164.
234
Raemy A., Hurrell R., and Lliger J. 1983. Thermal behavior of milk powders studied
by differential thermal analysis and heat flow calorimetry. Thermochim Acta,
65:8192.
Raemy A., Kaabi C., Ernst E., and Vuataz G. 1993. Precise determination of low
level sucrose amorphism by microcalorimetry. J Thermal Analysis, 40:437444.
Raemy A., Kaabi C., and MacInnes W.M. 1990. Mise en vidence de la rtrogradation
de lamidon par microcalorimtrie isotherme. In: Calorimtrie et Analyse
Thermique, AFCAT editor, pp. 7378. Clermont-Ferrand: France.
Raemy A., Lambelet P. and Garti N. 2000. Thermal behaviour of food and food
constituents. In: Thermal Behavior of Dispersed Systems, N. Garti, editor, pp.
477505. Marcel Dekker: New York.
Raemy A., Lambelet P., and Rousset P. 2004. Calorimetric information about food
and food constituents. In: The Nature of Biological Ssystems as Revealed by
Thermal Methods, D. Lrinczy, editor, pp. 6998. Kluwer Academic Publishers:
Dordrecht.
Raemy A., Lambelet P., and Lliger J. 1985. Thermal analysis and safety in relation
to food processing. Thermochim Acta, 95:441446.
Roduit B. 2002. Prediction of the progress of solid state reactions under different
temperature modes. Thermochim Acta, 388:377387.
Roos Y. 1995. Phase Transition on Foods. Academic Press: New York.
Roos Y. and Karel M. 1991. Applying state diagrams to food-processing and development. Food Technol, 45:6671.
Rouquerol J., Wads I., Lever T.J., and Haines P.J. 2007. Developments in nomenclature. In: Handbook of Thermal Analysis and Calorimetry, Vol. 5, Further
Advances, Techniques and Applications, P. Gallagher and M. Brown, editors, pp
2162. Elsevier: Amsterdam.
Rousset P. 1997. Etude Exprimentale et Modlisation de la Cristallisation de
Triacylglycrols et du Beurre de Cacao. Thesis 1718, EPFL, Lausanne, Switzerland.
Rousset P. and Rappaz M. 1996. Crystallization kinetics of the pure triacylglycerols
glycerol-1,3-dipalmitate-2-oleate, glycerol-1-palmitate-2-oleate-3-stearate, and
glycerol-1,3-distearate-2-oleate. J Am Oil Chem Soc, 73:10511057.
Rousset P. and Rappaz M. 1997. Alpha-melt-mediated crystallization of 1-palmitoyl2-oleoyl-3-stearoyl-sn-glycerol. J Am Oil Chem Soc, 74:693697.
Rousset P. and Rappaz M. 2001. Experimental study and computer modeling of the
dynamic and static crystallization of cocoa butter. In: Crystallization and
Solidification Properties of Lipids, N. Widlak, R. Hartel and S. Narine, editors,
pp. 96109. AOCS Press: Champaign, IL.
Rousset P., Rappaz M., and Minner E. 1998. Polymorphism and solidification kinetics
of the binary system POS-SOS. J Am Oil Chem Soc, 75:857864.
Rousset P. 2002. Modeling crystallization kinetics of triacylglycerols. In: Physical
Properties of Lipids, A.G. Marangoni and S.S. Narine, editors, pp. 136. Marcel
Dekker: New York.
Salvetti G., Tognoni E., Tombari E., and Johari G.P. 2007. Excess energy of polymorphic states or glass over the crystal state by heat of solution measurement.
Thermochim Acta, 285:243252.
235
Sato K. 1996. Polymorphism of pure triacylglycerols and natural fats. In: Advances
in Applied Lipid Research, Volume 2, F.B. Padley, editor, pp. 213268. IJAI Press:
London.
Sestak J. 1984. Thermophysical properties of solids. Elsevier: Amsterdam.
Silverio J., Svensson E., Eliasson A.C., and Olofsson G. 1996. Isothermal microcalorimetric studies on starch retrogradation. J Thermal Anal, 47:11791200.
Spagnolo D.A., Maham Y., and Chuang K.T. 1996. Calculation of contact angle for
hydrophobic powders using heat of immersion data. J Phys Chem, 100:6626
6630.
Spigno G., Pagella C., and Faveri D. 2001. DSC characterization of cocoa butter
polymorphs. Ital J Food Sci, 13:275284.
Tlgyesi F., Szgyi M., and Gyrgyi S. 1985. DSC study of the influence of chemical
environment on the structure of lyotropic liquid crystals. Thermochim Acta,
93:3740.
Tan C.P. and Man Y.B. 1999. Quantitative differential scanning calorimetric analysis
for determining total polar compounds in heated oils. J Am Oil Chem Soc,
76:10471057.
Tan C.P. and Man Y.B. 2000. Differential scanning calorimetric analysis of edible
oils: Comparison of thermal properties and chemical composition. J Am Oil Chem
Soc, 77:143155.
Tan C.P. and Man Y.B. 2002. Recent developments in Differential Scanning
Calorimetry for assessing oxidative deterioration of vegetable oils. Trends Food
Sci Technol, 13:312318.
Tan C.P., Man Y.B., Selamat J., and Yusoff M.S. 2002. Comparative studies of oxidative stability of edible oils by DSC and oxidative stability index methods. Food
Chemistry, 76:385389.
Terada K., Kitano H., Yoshihashi Y., and Yonemochi E. 2000. Quantitative correlation between initial dissolution rate and heat of solution of drug. Pharm Res,
17:920924.
Timms R.E. 1994. Physical chemistry of fats. In: Fats in Food Products, D.P.J. Moran
and K.K. Rajah, editors, pp.127. Blackie & Son: Glasgow.
Toro-Vazquez J. F., Rangel-Vargas E., Dibildox-Alvarado E., and Charo-Alonso M.
A. 2005. Crystallization of cocoa butter with and without polar lipids evaluated
by rheometry, calorimetry, and polarized light microscopy. Eur J Lipid Sci
Technol, 107:(9)641655.
Ulkowski M., Musialik M., and Litwinienko G. 2005. Use of Differential Scanning
Calorimetry to study lipid oxidation. I. Oxidative stability of lecithin and linolenic
acid. J Agric Food Chem, 53:(23)90739077.
Unterhaslberger G., Schmitt C., Sanchez C., Appolonia-Nouzille C., Raemy A. 2006.
Heat denaturation and aggregation of B-lactoglobulin enriched WPI in the presence of arginine HCl, NaCl and guanidinium HCl at pH 4.0 and 7.0. Food
Hydrocolloids, 20:100610019.
Vanhoutte B., Dewettink K., Foubert I., Vanlerberghe B., and Hyughebaert A. 2002a.
The effect of phospholipids and water on the isothermal crystallization of milk fat.
Eur J Lipids Sci Technol, 104:490495.
236
Vanhoutte B., Foubert I., Duplacie F., Huyghebaert A., and Dewettinck K. 2002b.
Effect of phospholipids on isothermal crystallization and fractionation of milk fat.
Eur J Lipid Sci Technol, 104:738744.
Vu P.L., Park R.K., Lee Y.J., Kim, Y.M., Nam, H.Y., Lee J.H., Akoh C.C., Lee, K.T.
2007. Two-step production of oil enriched in conjugated linoleic acids and diacylglycerol. J Am Oil Chem Soc, 84:(2)123128.
Vuataz G. 2002. The phase diagram of milk: A new tool for optimizing the drying
process. Le Lait, 82:485500.
Wads I. 1997. Isothermal microcalorimetry near ambient temperature: An overview
and discussion. Thermochim Acta, 294:111.
Wahnelt S., Meusel D., and Tulsner M. 1991. Influence of isomeric diglycerides on
phase transitions of cocoa butterinvestigations by isothermal DSC. Fett Wiss
Technol, 93:174178.
Widmann G. and Riesen R. 1987. Thermal Analysis: Terms, Methods, Applications.
A. Hthig Verlag: Heidelberg, Germany.
Wille R. and Lutton E. 1966. Polymorphism of cocoa butter. J Am Oil Chem Soc,
43:491496.
Willumeit R., Kumpugdee M., Funari S.S., Lohner K., Pozo Navas B., Brandenbourg
K., Linser S., and Andr J. 2005. Structural rearrangement of model membranes
by the peptide antibiotic NK-2. Biochem Biophys Acta, 1669:125134.
Wong S.M., Kellaway I.W., and Murdan S. 2006. Enhancement of the dissolution
rate and oral absorption of a poorly water soluble drug by formation of surfactantcontaining microparticles. Int J Pharm, 317:6168.
Yan N.X., Maham Y., Masliyah J.H., Gray M.R., and Mather A.E. 2000. Measurement
of contact angles for fumed silica nanospheres using enthalpy of immersion data.
J Colloid Interface Sci, 228:16.
Zoungrana T., Douillard J.M., and Partyka S. 1994. Assessment of the surface-tension
of various divided solids. J Thermal Anal, 41:12871293.
Chapter 10
Shelf Life Prediction of Complex
Food Systems by Quantitative Interpretation
of Isothermal Calorimetric Data
Simon Gaisford, Michael A.A. ONeill, and Anthony E. Beezer
Introduction
Qualitative Studies
Quantitative Studies
Empirical Model Fitting
Modeling Based on Reaction Kinetics
Reactions That Proceed to Completion
Calculation of the Initial Calorimetric Signal 0
Calculation for the Reaction Order
Calculation for the Total Heat Released for Complete Reaction
Calculation for Reaction Half-Life
Calculation for Rate Constant
Calculation for Reaction Enthalpy
Reactions That Proceed to a Point of Equilibrium
Test for Complete Reaction
Determination of K
Calculation of QT
Summary
References
237
239
245
246
249
252
252
252
253
254
254
255
255
255
255
256
261
261
Introduction
The analysis of foods and food components presents a considerable
challenge, not least because they may contain many ingredients, be
237
238
difficult to handle, and may derive from one or more natural sources
(which inherently introduces batch-to-batch variability in composition). Indeed, in many respects foods may be considered as complex
biomaterials. While qualitative analyses may suffice for some quality
assurance protocols (e.g., texture, feel, density), it is often desirable to
obtain quantitative data; this is particularly true where food authenticity (i.e., whether a food substance conforms to its label claim) is
required. In this context, it is often difficult or complex to use classical
assay techniques. It is not necessarily straightforward, for example, to
quantify banned additives in foodstuffs using high-performance liquid
chromatography or spectroscopy because of the need to isolate the
analyte prior to analysis. A discussion of the merits of various analytical tools for determination of food authenticity can be found in Reid
et al. (2006).
We have long argued that calorimetry, in particular isothermal calorimetry, is ideally suited to the study of complex samples because it
offers many unique advantages. First, the measured parameter is heat.
This is advantageous because heat can be considered as a universal
indicator of change (and note here that change in this context can mean
both physical and chemical processes). Thus, it will unquestionably be
the case that a sample can in principle be studied with calorimetry.
Whether a meaningful interpretation can be made depends only upon
the magnitude of the heat change and the number of events occurring
(discussed further below). Second, the instrument requires no sample
treatment or preparation. The entire sample is housed within an ampoule
and monitored in situ; or, if the sample is too large for the ampoule, a
fraction is enclosed that is representative of the whole. Thus, the need
to isolate a particular analyte, as in a chromatographic assay, is obviated. Finally, the technique does not require optical clarity of a sample
and is invariant to physical form, which means that any complex material can be studied in its entirety.
There are two principal calorimetric techniques; isothermal calorimetry (IC) and differential scanning calorimetry (DSC). With the former,
the sample is monitored at a constant temperature; and with the latter,
the sample is subjected to a controlled temperature ramp (usually
increasing). Unsurprisingly many of the calorimetric studies of foodstuffs have used DSC (see Raemy et al. 2004; Schiraldi 2004) or
temperature-modulated DSC (De Meuter et al. 1999). This is, perhaps,
not unexpected since such instruments are uniquely well suited to
239
Qualitative Studies
Quantitative data interpretation usually requires some prior knowledge
of the properties of the system under investigation (e.g., number of
reacting systems, reaction pathway) as well as a model with some
factual basis. While for simple (one- or two-component) systems quantitative interpretation may be possible, inevitably with more complex
systems there will be instances in which qualitative outcomes can be
indicative of change despite the absence of detailed interpretation of
the experimental data. Much of the literature in this field thus reports
qualitative data, and it is here that this discussion starts.
Qualitative analyses are usually applied to complex systems because
in such cases it is not technically possible to interpret multifaceted
power-time data. Consequently, this discussion starts with applications
of calorimetry to bioprocesses and bioprocessing in which microorganisms are used in the production of, or as an ingredient in, foodstuffs.
It is not a simple matter to measure efficacy of bioprocesses in situ
because of the heterogeneity of the system (which may change viscosity, density, and optical transparency). Conventional microbiological
assay techniques, such as plating and counting, are time consuming,
do not provide information on real-time growth in the actual process
environment, do not compensate for the fraction of the microbial load
that is dead or not viable, and exclude any contribution to efficacy
caused by physical effects (such as thickening of the medium).
240
241
Figure 10.1. Power-time data showing the growth curves for six repeats of P.
aeruginosa.
A good example of the very valuable data that can be produced from
qualitative analysis of isothermal calorimetric data is provided by
studies of the growth of yeasts on a variety of substrates (Perry et al.
1981, 1983). Yeasts play an important role in many food industries,
including baking and brewing, and are used in other industries as well,
for example, as a means of producing ethanol from renewable sources
or for use in single-cell protein (SCP) production. The authors took
three commercial strains of S. cerevisiae (D1, distilling; D2, pressed
baking; and D3, active dried baking) and two NCYC strains (87, distilling and 239, brewing) and analyzed their growth curves with flow
microcalorimetry (an isothermal calorimeter equipped with a cell that
allows medium to be flowed through from an external reservoir). In a
glucose medium the growth curves of the three baking strains showed
little differentiation, although the growth curve of the brewing strain
was notably different. In a maltose medium, good differentiation was
observed between all strains. One immediate outcome from these data
is the ability to use a simple medium (maltose) to characterize the
properties of a new strain of S. cerevisiae and select it as appropriate
for either baking or brewing.
When complex media are considered the utility of calorimetric study
increases further. Perry et al. (1983) showed that growth of yeast in
242
glucose-maltose media did not show sequential use of the carbohydrates; rather, both sugars were exhausted after the initial growth
phase. Discrimination of strains was achieved by adding maltotriose
to the medium.
Molasses, a waste product of the sugar industry, is a raw material
used in both baking and brewing because it is highly suitable for yeast
growth. The suitability of a particular batch of molasses is dependent
upon the technical processes by which the sugar was manufactured and
also by agronomic factors of the cane and beet production. As a result,
the growth of yeast in a particular molasses batch can be highly variable, and there will be a direct impact on the production costs and
quality of the baked or brewed product. A method that allows rapid
assessment of the quality and suitability of a batch of molasses is thus
highly desirable and difficult to achieve by classical physical and
chemical means. It is known that only part of the carbohydrate reservoir in molasses is of nutritive value to the yeast. As a consequence,
analytical data are not sufficient to characterize molasses batches from
a bioavailability point of view. Perry et al. (1981) show how interpretation of calorimetric growth curves of S. cerevisiae in molasses samples
could be used rapidly to identify optimal growth media, growth curves
in molasses identified as poor being distinct from growth in molasses
typed as adequate-to-good.
A similar approach was used more recently by Alklint et al. (2005)
to predict the shelf life of carrot juice. Here, growth of the mesophilic
and psychrotrophic flora in the carrot juice was monitored after manufacture at 17 C (the highest permitted temperature for elevated stability tests of chilled foodstuffs in Sweden) with an isothermal calorimeter.
It was found that the heat outputs recorded correlated with plate counts,
indicating that the calorimetric approach was valid. The initial cause
of spoilage was found to be the same at 17 C as at 8 C (the maximum
permitted storage temperature of chilled foodstuffs in Sweden), and
the data were found to be suitable for predicting shelf lives.
An area of growing interest is foods that offer some health benefits,
so-called functional foods. Prime among these are products that aim
to modulate the microflora of the gastrointestinal (GI) tract. Bacterial
numbers vary along the human GI tract, increasing from about 103 g1
of gastric content to about 106107 g1 of content at the terminal ileum
(Gorbach et al. 1967). The colon, in particular, is a complex and
diverse microbial ecosystem, bacterial numbers reaching 10111012 g1
243
of gut content (Cummings and Macfarlane 1991). This makes the colon
the most metabolically active organ in the body. In addition, the number
of bacterial species is large (several hundred), creating a diverse
microflora. The numerically predominant bacteria (30%) are the
bacteroides, although other genera, such as bifidobacteria, eubacteria,
lactobacilli, streptococci, and clostridia, are also present (Fooks et al.
1999).
The microflora has many roles and is important to general health
and well-being. A primary function is in fermenting undigested foodstuffs that have not been absorbed in the upper GI tract. Typically,
substrates are carbohydrates (including starches, dietary fiber, and oligosaccharides). The principal fermentation products are short-chain
fatty acids (SCFA); these are subsequently absorbed by the body and
metabolized, which contributes to the energy gain of the host
(Cummings 1995) and means the relationship between host and microflora is symbiotic. Another important function is to inhibit the growth
of pathogenic organisms. However, in the absence of sufficient carbohydrate, certain bacteria, such as clostridia, switch to protein fermentation, which produces harmful nitrogenous metabolites (including
biogenic amines, indoles, and ammonia).
To minimize this effect, the body excretes a number of mucins,
which are high in carbohydrates and encourage the growth of certain
microbial species. The UK food market is currently replete with products designed to modulate the gut microflora; usually, such products
are supplemented with either prebiotics or probiotics. Probiotics are
live bacteria of species deemed to be beneficial to health when ingested;
usually, lactic acid bacteria (LAB) are indicated and they are commonly found in yogurts, where they convert lactose to lactic acid,
giving the product its distinctive sour taste and acting as a preservative.
Prebiotics are nondigestible food ingredients that stimulate the growth
of one or more beneficial bacteria in the colon and were first defined
by Gibson and Roberfroid (1995). Several potential nondigestible
foodstuffs have been investigated for prebiotic efficacy, but to date the
only substances for which credible data showing a favorable effect are
available are the oligosaccharides (Delzenne and Roberfroid 1994).
However, it is not a simple matter to show a beneficial prebiotic
effect in vivo, primarily because of the sheer complexity of the gut
microflora and its effect on physiological response. Typically,
an increase in the number of bifidobacteria excreted in feces has
244
245
time. This qualitative outcome was supplemented with data from plate
counting and pH measurements, which confirmed the increasing power
signal derived from an increase in the microbial population.
Galindo et al. (2005) investigated the change in the metabolic
response of foodstuffs pre- and postprocessing by isothermal calorimetry. Here, minimal processing operations were considered, such as
peeling, grating or shredding. The effect of these operations can impact
the quality of the product through several means; changes in respiration
rate, increased biochemical reactions through wounding stress and
microbiological storage. The rate of many of these processes is dependent upon the surface area of the product, and the authors found relationships between surface area and thermal power for several vegetables
(carrots, rutabagas, and potatoes). It was also possible to monitor enzymatic browning of potatoes and the effect of browning inhibitors, such
as citric acid or ascorbic acid.
Quantitative Studies
Quantitative (which may mean simply determining the number and
nature of reaction processes through to recovery of descriptive reaction
parameters, such as rate constants, enthalpies, and activation energies)
interpretation of complexity in isothermal calorimetric data is demanding. Primarily, this is because, as noted, heat is a universal accompaniment to chemical and physical change. Its ubiquitous nature means that
the measured signal is a composite of the powers arising from each of
the individual events occurring (which can involve physical, as well
as chemical, change). Any meaningful analysis thus has the primary
objective of determining the number of processes contributing to the
overall data. This number alone is a useful basis for quantitative interpretation, because it could be indicative of a reaction pathway and
hence could give some indication of mechanism. Once the number of
steps is known, the data must be deconvoluted into their component
parts. Once the individual processes are identified, analysis to recover
quantitative reaction parameters is usually much more straightforward.
Over the past 15 years, a number of methods have been proposed for
quantitative analysis of calorimetric data, from simple model fitting to
model-free chemometric analysis. Here, these approaches are reviewed
and illustrated with examples of application to food products.
246
y = y0 A.e t
(10.1)
where x and y are the plotted variables, y0 is the initial value of y, and
A and t are constants. For example, some calorimetric data are represented in Figure 10.2. The exact process that gave rise to these data is
not important, but it shall be assumed that they represent the heat
output of a partially completed reaction. The data can be fitted to
Equation 10.1 by least-squares minimization to determine the equation
parameters that describe them. Once these values have been determined, it is a simple matter to extend the data to the time at which the
power signal falls to zero (shown by the dotted line in Figure 10.2).
The area under the dotted line thus represents the total heat, Q, which
Figure 10.2. The fit of calorimetric data to an exponential model and the subsequent
extrapolation to power = 0.
247
248
calibration curve was found to be linear to 2.68 g/l; above this concentration, nonlinearity was found as a result of the inverse reaction also
catalyzed by the fumarase. The authors compared the assay with a
classical spectrophotometric approach and noted the calorimetric technique gave equivalent answers but with no requirement for sample
isolation or cleanup. L-malic acid was successfully quantified in a
range of beverages (red and white wines, soft drinks, and apple juice)
and solid food products (apples, mandarins, and powder for making
carbonated water).
The use of ascorbate oxidase to quantify ascorbic acid (vitamin C)
concentrations has attracted much attention. Antonelli et al. (2002)
found a calibration curve of calorimetric response versus ascorbic acid
to be linear between ascorbic acid concentrations of 3270 mg/l.
Similarly, ONeill (2004) used this approach to investigate the quality
of fresh orange juices and determined the kinetics of ascorbic acid
degradation (discussed further below).
A derivative technique that can offer some useful insight into the
properties of food substances is solution calorimetry. In this experiment, a solid (usually) sample is introduced to a solvent, and the heat
change upon dissolution is recorded (Royall and Gaisford 2005). The
technique is useful because small changes in the physical form of a
material, such as a change in polymorph or percentage of amorphous
content, will result in a change of dissolution enthalpy. Marabi et al.
(2007) have shown that solution calorimetry can be used to study the
dissolution of two food powders, maltodextrin and skimmed milk. As
the moisture content of the powder increased a concomitant decrease
in the exothermic dissolution, heat was seen. The effect was reversible
if the moisture content was reduced unless crystallization occurred in
the sample. The authors used real-time video analysis to follow dissolution kinetics and related these to the calorimetric data.
However, it is possible to interpret the calorimetric data quantitatively to recover dissolution parameters. Conti et al. (2006), in a study
of the dissolution of various hydroxymethylcellulose (HPMC) polymers, demonstrated how it is possible to convert the calorimetric dissolution data into a plot of swelling ratio; this was achieved by assuming
that the total heat output during the experiment, Q (obtained by integration of the power-time data), corresponded to complete swelling, while
the heat output to any time t (qt) corresponded to the fraction of swelling that had occurred to that point. Hence, a plot of qt/Q versus time
249
Figure 10.3. Power-time data for the dissolution of various grades of HPMC into
water. Reprinted from Conti et al. (2006), with permission from Elsevier.
gave a set of data that represented the swelling response; typical plots
of the dissolution profile of the polymer and the swelling ratios are
shown in Figures 10.3 and 10.4. The swelling curves were then analyzed using a power-law model. It was shown that for all polymers
dissolution occurred immediately following hydration of the polymer,
although there was a rate dependence upon both polymer grade and
solution pH.
Modeling Based on Reaction Kinetics
The next step in complexity from the use of empirical models is to fit
data to models based on reaction kinetics. Clearly, for this approach
to be valid, the mechanism of reaction must be known prior to analysis.
This approach is also not appropriate for those processes that involve
physical change (although if the chemical reaction data can be isolated
from any physical change data this approach is legitimate). While the
use of such models has been documented in full elsewhere (Gaisford
and ONeill 2006), it is appropriate to discuss the simplest case, a
single step, A P reaction process here; logical extension of the
analysis can be made to more complex reaction schemes. The rate of
disappearance of reactant A, or the buildup of product P, is given by
250
Figure 10.4. The swelling profiles of various grades of HPMC calculated from the
power-time data shown in Figure 10.1. Reprinted from Conti et al. (2006), with permission from Elsevier.
dA
= kAn
dt
(10.2)
A = A0 x
(10.3)
dx
n
= k ( A0 x )
dt
(10.4)
Since
Then
Where dx/dt is the rate of reaction, k is the rate constant, A0 is the initial
quantity of reactant A that is available for reaction, x is
the quantity of reactant A reacted at time t, and n is the order of reaction. It should be noted that from a mathematical perspective, n may
have any value, integral or nonintegral, but to have meaning as a
kinetic model only integral values are considered. Immediately, therefore, it is possible to check the validity of a model against real data,
for the values of any reaction orders obtained should be integral. For
any given reaction that has gone to completion, the total heat evolved
251
(10.5)
q = x H
(10.6)
It follows that
where q is the heat evolved at time t. Substituting q/H for x and Q/H
for A0 in Equation 10.4 and rearranging gives
dq
n
= = k H 1 n (Q q )
dt
(10.7)
where is the calorimetric power (in watts). Assuming n 1, integration of Equation 10.7 gives
1
(Q q ) = [ k t H 1 n (n 1) + Q1 n ]1 n
(10.8)
= k H 1 n [ k t H 1 n (n 1) + Q1 n ]1 n
(10.9)
252
t1 =
k H 1 n
1 n
n
Q1 n
k H 1 n ( n 1)
(10.10)
and
t2 =
k H 1 n
1 n
n
Q1 n
k H 1 n ( n 1)
(10.11)
and therefore,
1 n
t 2 ( 2 ) n
=
t1 ( )1nn
1
(10.12)
253
It follows that the order of reaction may be determined from knowledge of the value of t2/t1. This is most easily achieved through the use
of a suitable mathematical worksheet. The values of 1 and 2 are
converted into percentages of the initial calorimetric signal (0) and,
by using the worksheet, a table of values of t2/t1 calculated from
Equation 10.12, can be constructed as a function of the rate constant
for a particular pair of (1/0) and (2/0) ratios. The experimental t2/t1
constant may then be compared with the table of t2/t1 values, and the
order of reaction can be determined.
(10.7)
(10.13)
(10.14)
Then,
n
1 (Q q1 )
=
2 (Q q2 )n
(10.15)
1 n = (Q q1 )
2
(Q q2 )
Setting
(10.16)
254
1n
=R
(10.17)
(q1 Rq2 )
(10.18)
1 R
(2n1 1)
(10.19)
[( n 1) kA0n1 ]
(10.20)
255
If the ratio of two data points 1 and 2 at times t1 and t2 are taken,
Equation 10.21 is obtained:
1 (1 + kA0 t2 )
=
2 (1 + kA0 t1 )
(10.21)
(10.22)
[ A]
[ B]
(10.23)
256
[ B] =
Q
H
(10.24)
[ A] = AT [ B ]
(10.25)
QT Q
H H
(10.26)
QT Q
K= H H
Q
H
(10.27)
Q
(QT Q )
(10.28)
Hence,
[ A] =
Thus,
And therefore,
K=
257
(10.29)
(10.30)
K1
K
will be equal to 2 , for temperatures T2 and T3, if the
K2
K3
temperatures T1, T2, and T3 are such that Equation 10.31 is true.
The ratio
T1T2
T2T3
=
(T2 T1 ) (T3 T2 )
(10.31)
K1 (QT Q1 ) K 2 (QT Q2 )
=
=
=
Q3
Q2
K2
K3
(QT Q2 )
(QT Q3 )
(10.32)
Q2 Q3Q1
(10.33)
(10.34)
258
(10.35)
(H G )
T
= S
(10.36)
(10.37)
The calculated values of K will allow an internal check on the validity of the procedure through the derived value of H. A plot of
ln K versus 1 will yield a straight line with slope H . If the procedure
T
R
is invalid, the mechanism changes with T, or there is a significant
temperature dependency of heat capacity (over the temperature range
used), the vant Hoff plot will not be linear, and/or the value of H
will not match that derived from Equation 10.34.
Since Q and QT are known and if H is accurately recovered, then
it is possible quantitatively to determine the reactable material content
in a heterogeneous sample.
Although there are fewer examples in the literature, kinetic interpretation of calorimetric data for foodstuffs has been reported. Tortoe
et al. (2007) report a kinetic analysis of the osmotic dehydration of
apples, bananas, and potatoes. In osmotic drying, the foodstuff is
placed in a concentrated sugar solution; water is then drawn out of the
foodstuff as a result of osmosis. A side effect of osmotic drying is the
ingress of sugar into the foodstuff, resulting in a dried, sweetened
material. For all apple samples, a three-phase signal was observed. The
initial, large, signal was assumed to be rapid transfer of water and sugar
from the cut surfaces of the food. The second phase represented movement of water from intracellular spaces, and the final phase represented
movement of water from extracellular spaces. For banana and potato
samples, a two-phase signal was observed. In all cases, the processes
259
were first order, which meant that a simple plot of ln (power) versus
time allowed recovery of the rate constants from the gradients of the
respective phases. The rate constants changed in magnitude in the
order k1 > k2 > k3 and increased with temperature. Arrhenius plots of
the data suggested a linear relationship for Golden Delicious
apples and potatoes and nonlinear behavior for Cox apples and
bananas.
As noted, ONeill (2004) studied the kinetics of ascorbic acid in
freshly squeezed orange juice. This is more complex than simple
studies of the oxidation of ascorbic acid because spoilage of fresh
orange juice generally results both from oxidation of ascorbic acid
and from degradation of pectin by pectin methyl esterase. Although it
is still the subject of debate, it is thought that the degradation of
pectin results in protein agglomerates that settle, resulting in clarification of the juice. Clarification is a major cause of spoilage in orange
juice, and strenuous efforts are made to eliminate it. This can be done
in a number of ways. Classically, the juice is pasteurized to denature
the protein as well as destroy the microbial flora. However, heat
treatment adversely affects the flavor and aroma of the juice and is
generally unsatisfactory if the juice is to be marketed as fresh-squeezed.
Alternatively, the juice is frozen during storage and transportation,
reducing the activity of the enzyme and other degradative processes.
This is costly and usually only viable for concentrated juice. An ideal
solution would be some natural additive that does not adversely affect
the desired quality parameters for orange juice but does inhibit pectin
methyl esterase, in quantities that are not harmful, reducing cloud
destabilization. Selection of such a material starts with the ability
quantitatively to monitor the reaction processes directly within orange
juice samples.
Both the oxidation reaction and the enzyme reaction are pH dependent, with the oxidation reaction favoring alkaline conditions and the
enzyme reaction favoring more neutral conditions. By buffering the
orange juice at pH 7.0, ONeill (2004) held the enzyme reaction at
almost optimal conditions, while the oxidation reaction was more
favored compared with the naturally more acidic pH of orange juice.
The calorimetric data revealed two distinct first-order phases (Figure
10.5). The first reaction progressed with a rate constant of 3.9
(1.0) 105 s1, and the second progressed with a rate constant of 2.7
(0.3) 105 s1.
260
Figure 10.5. Observed calorimetric output for buffered orange juice. Reprinted from
ONeill (2004), with permission.
0
kH
(10.38)
261
Summary
The study of foods and food ingredients is complicated, largely because
physical form impacts the selection and use of any analytical assay
technique. Because it measures a universal property and is invariant to
physical form, calorimetry offers an exciting opportunity for the investigation of foodstuffs. The full capabilities of calorimetry have not yet
been exploited in this area, although the number of reported applications is increasing. Here, it has been shown that calorimetry can be
used to study a huge variety of samples, from simple ingredients (such
as ascorbic acid) to complex biological processes. Data interpretation
ranges from qualitative to quantitative, the analysis methods being
dependent upon the complexity of the sample. However, a combination
of clever experimental design, sample preparation, and data analysis
mean that quantitative outcomes are increasingly available from calorimetric data matrices, and the application of the technique to foods
and food ingredients can only continue to increase.
References
Alklint, C., Wads, L., and Sjholm, I. 2005. Accelerated storage and isothermal
microcalorimetry as methods of predicting carrot juice shelf-life. J Sci Food Agri,
85:281285.
Antonelli, M.L., Spadaro, C., and Tornelli, R.F. 2008. A microcalorimetric sensor
for food and cosmetic analyses: L-malic acid determination. Talanta, 74:
14501454.
Antonelli, M.L. DAscenzo, G. Lagan, A., and Pusceddu, P. 2002. Food analyses:
A new calorimetric method for ascorbic avid (vitamin C) determination. Talanta,
58:961967.
Bakri, A., Janssen, L.H.M., and Wilting, J. 1988. Determination of reaction rate
parameters using heat conduction microcalorimetry. J Thermal Anal, 33:
185190.
Beezer, A.E., Newell, R.D., and Tyrrell, H.J.V. 1976. Application of flow microcalorimetry to analytical problemspreparation, storage and assay of frozen inocula
of saccharomyces cerevisiae. J Appl Bacteriol, 41:197207.
Conti, S., Gaisford, S., Buckton, G., and Conte, U. 2006. Solution calorimetry to
monitor swelling and dissolution of polymers and polymer blends. Thermochim
Acta, 450:5660.
Cosgrove, R.F. 1979. Long-term storage of microorganisms used in antimicrobial
effectiveness tests. J Assoc Off Anal Chem, 62:11881190.
262
Cummings, J.H. 1995. Short chain fatty acids. In: Human Colonic Bacteria: Role in
Nutrition, Physiology and Pathology, G.R. Gibson and G.T. Macfarlane, editors.
CRC Press: Boca Raton, FL.
Cummings, J.H. and Macfarlane, G.T. 1991. The control and consequences of bacterial fermentation in the human colon. J Appl Bacteriol, 70:443459.
Delzenne, N. and Roberfroid, M.B. 1994. Physiological effects of non-digestible
oligosaccharides. Lebens-Wiss Technol, 27:16.
De Meuter, P., Rahier, H., and Van Mele, B. 1999. The use of modulated temperature
differential scanning calorimetry for the characterisation of food systems. Int J
Pharm, 192:7784.
Fooks, L.J., Fuller, R., and Gibson, G.R. 1999. Prebiotics, probiotics, and human gut
microbiology. Int Dairy J, 9:5361.
Forte, L., Vinci, G., and Antonelli, M.L. 1996. Isothermal microcalorimetry as a
useful tool for fat determination in food. Anal Let, 29:23472362.
Gaisford, S. and ONeill, M.A.A. 2006. Pharmaceutical isothermal calorimetry.
Informa Healthcare, New York.
Gibson, G.R. and Roberfroid, M.B. 1995. Dietary modulation of the human colonic
microbiota: Introducing the concept of prebiotics. J Nut, 125:14011412.
Gmez-Galindo, F., Rocculi, P., Wads, L., and Sjholm, I. 2005. The potential of
isothermal calorimetry in monitoring and predicting quality changes during processing and storage of minimally processed fruits and vegetables. Trends Sci
Technol, 16:325331.
Gorbach, S.L., Nahas, L., and Lerner, P.I. 1967. Studies of intestinal microflora. I.
Effects of diet, age, and periodic sampling on numbers of faecal microorganisms
in man. Gasteroenterology, 53:845855.
Marabi, A., Mayor, G., Raemy, A., Bauwens, I., Claude, J., Burbidge, A.S., Wallach,
R., and Saguy, I.S. 2007. Solution calorimetry: A novel perspective into the dissolution process of food powders. Food Res Int, 40:12861298.
ONeill, M.A.A. 2004. PhD dissertation. University of Greenwich: London.
Ouwehand, A.C., Derrien, M., de Vos, W., Tilhonen, K., and Rautonen, N. 2005.
Prebiotics and other microbial substrates for gut functionality. Cur Opi Biotechnol,
16:212217.
Perry, B.F., Beezer, A.E., and Miles, R.J. 1979. Flow microcalorimetric studies of
yeast growth: Fundamental aspects. J Appl Bacteriol, 47:527537.
Perry, B.F., Beezer, A.E., and Miles, R.J. 1981. Microcalorimetry as a tool for evaluation of complex media: Molasses. Microbios, 32:163172.
Perry, B.F., Beezer, A.E., and Miles, R.J. 1983. Characterization of commercial yeast
strains by flow microcalorimetry. J Appl Bacteriol, 54:183189.
Qin, C., Li, H., Xiao, Q., Liu, Y., Zhu, J., and Du, Y. 2006. Water-solubility of chitosan and its antimicrobial activity. Carbohydrate Polymers, 63:367374.
Raemy, A., Lambelet, P., and Rousset, P. 2004. Calorimetric information about food
and food constituents. In: The Nature of Biological Systems as Revealed by
Thermal Methods, Dnes Lrinczy, editor, pp 6998. Klewer Academic Publishers:
London.
263
Reid, L.M., ODonnell, C.P., and Downey, G. 2006. Recent technological advances
for the determination of food authenticity. Trend Food Sci Technol, 17:344353.
Riva, M., Fessas, D., and Schiraldi, A. 2001. Isothermal calorimetry approach to
evaluate the shelf life of foods. Thermochim Acta, 370:7381.
Royall, P.G. and Gaisford, S. 2005. Application of solution calorimetry in pharmaceutical and biopharmaceutical research. Curr Pharm Biotechnol, 6:215222.
Schffer, B., Szakly, S., and Lrinczy, D. 2004. Examination of the growth of probiotic culture combinations by the isoperibolic batch calorimetry. Thermochim
Acta, 415:123126.
Schiraldi, A. 2004. Thermal analyses and combined techniques in food physical
chemistry. In: The Nature of Biological Systems as Revealed by Thermal Methods,
Dnes Lrinczy, editor, pp 6998. Klewer Academic Publishers: London.
Tortoe, C., Orchard, J., Beezer, A.E., and ONeill, M.A.A. 2007. Potential of calorimetry to study osmotic dehydration of food materials. J Food Eng,
78:933940.
Willson, R.J. 1995. Ph.D. dissertation. University of Kent: Canterbury.
Willson, R.J., Beezer, A.E., Mitchell, J.C., and Loh, W. 1995. Determination of
thermodynamic and kinetic parameters from isothermal heat conduction
microcalorimetry: Applications to long-term reaction studies. J Phys Chem
99:71087113.
Chapter 11
Use of Thermal Analysis to Design and
Monitor Cereal Processing
Alberto Schiraldi, Dimitrios Fessas, and Marco Signorelli
Introduction
Starch
Proteins
Nonstarch Carbohydrates
Process Applications
Conclusions
References
265
268
272
276
278
285
285
Introduction
The term thermal analysis (TA) means the record of any physical property during a given thermal treatment under strict temperature control.
The main physical property that is monitored in this way is enthalpy
(and the related property, heat capacity); the relevant thermal analysis
is named calorimetry and is performed in a few well-defined conditions, each with a specific name: isothermal calorimetry (IC), differential scanning calorimetry (DSC), temperature-modulated DSC
(TMDSC), and modulated adiabatic scanning calorimetry (MASC).
Another physical property that is traditionally included in the TA
realm is mass; the relevant analysis is named thermogravimetry (TGA)
and is currently used to monitor the mass loss upon heating the sample.
Its trace is often transformed into that of its time derivative (DTG).
The mechanical properties of a given sample, like Youngs modulus,
E, elastic and storage moduli, G and G, respectively, are those
265
266
267
Fortunately, this is not true, provided that the operators are well
trained and adequately educated in physical chemistry. Suitably performed TA indeed provides a real insight into the processes that occur
within a food system during a thermal treatment, as the relevant change
of the involved physical property is directly related to the laws of
thermodynamics and kinetics. For example, any thermal effect, H,
when coupled with a temperature value, T, is a measure of the stability
of the system, thanks to the relationship,
G = H
T2
T T p
(11.1)
(11.2)
268
Starch
Starch is a supramolecular substance that nature assembles within the
seeds of cereals and pseudocereals, legumes, and tubers. Starch may not,
therefore, be referred to as a chemical compound: it has neither a definite molecular mass nor a fusion point, and it does not react with other
substances until its granular structure is rotten and its glucose polymers, amylose and amylopectin, are exposed to the surrounding environment. Starch chemistry starts with surface processes that take place
at the pores and defects of the granule structure, which remains practically unaffected at temperatures below 45 C, even with excess water.
An aqueous suspension of starch granules is easy to prepare and
investigate with TA techniques. A superficial wetting takes place when
starch granules are dispersed in excess water. The process is exothermic and can be detected with isothermal calorimetry (IC) and suitable
mixing cells (Riva, Piazza, and Schiraldi 1991).
DSC (and related variations) equipment (Liu and Shi 2006) is instead
enough to monitor changes that take place when the starch suspension
is heated (Figure 11.1).
Across a 30 C range, starting from an onset temperature that depends
on the vegetal origin of the starch investigated (e.g., 45 , 50 , and 65 C,
for potato, wheat and rice, respectively), water enters the granule and
disaggregates the internal crystal regions that are mainly formed by the
side branches of amylopectin molecules. The whole starch granule is
transformed in a swollen jelly ghost of the original hard and birefringent
body. A gentle stirring turns the starting suspension into a dispersion of
amylopectin gel and amorphous insoluble amylose. The two glucose
polymers are mutually incompatible (Kalichevsky and Ring 1987),
which means that they may not stay in the same phase, being competitors
for the available moisture.
Because of this, the system is rather heterogeneous and unstable. On
further heating, the amylopectin gel turns into a sol, while around
90 C, nucleation of amylose crystals can take place. The whole process,
currently dubbed starch gelatinization, is therefore a multistep, fully
269
Figure 11.1. DSC trace of a suspension of starch (Merck product) granules in excess
(72.5%, w/w) water, recorded on heating (upper curve) and cooling (lower curve) at
5 C min1 scanning rate (authors unpublished data).
270
Figure 11.2. DSC traces of suspensions of starch (Merck product) granules at various
water contents (72.5%, 61.1%, and 42.5%, w/w) recorded on heating at 5 C min1
scanning rate (authors unpublished data).
271
Figure 11.3. Endothermic effect related to the fusion of amylopectin crystals formed
within the starch gel phases of a bread dough (data are given as excess heat capacity,
namely heat flux, divided by the scanning rate sample mass product). Dashed curve
describes the trend of the of the relevant enthalpy versus the storage time. The records
were obtained at 2 C min1 heating rate. Modified from Riva et al. (Riva, Fessas, and
Schiraldi 2000).
two DSC traces obtained from the same aqueous starch suspension:
the dashed profile refers to the first heating (2 C min1) run, during
which starch gelatinization and fusion of amylose-lipid complexes
occurred, whereas the full line profile refers to the heating run performed after a suitable annealing treatment (Wasserman et al. 2007)
that favors the formation of amylose crystals. The latter shows the
endothermic effect related to the fusion of amylopectin crystals, while
a second endothermic peak with onset at a much higher temperature
is related to the fusion of amylose crystals. The growth of these crystal
phases prevents the formation of amylose-lipid complexes (no signal
appears in the expected temperature range).
These amylose crystals are resistant to the amylase assisted hydrolysis: the hosting system has therefore received the misleading name
of resistant starch, which meets industrial and commercial needs
rather than the chemical truth.
A starch gel heated in an open crucible releases its moisture with a
diffusion-limited mechanism (Fessas and Schiraldi 2005). This means
that solvation water can be easily exchanged with water engaged in
the structure.
272
Figure 11.4. DSC traces obtained from an aqueous starch suspension (heating rate
2 C min1; data are transformed in apparent heat capacity, dividing the heat flux by
sample mass heating rate). The dashed line refers to the first heating. The heavy
line corresponds to the record of the reheating run performed after a suitable annealing
treatment (Wasserman et al. 2007). Notice that the fusion of the amylopectin crystals
(formed during the annealing) occurs at lower temperature, while no signal relevant
to amylase-lipid complexes appears and that a new endotherm occurs at much higher
T corresponding to the fusion of amylose crystals (formed during the annealing).
Proteins
In most cereals, both globular and networking proteins are present. The
former, dubbed albumins and globulins after Osborne (Osborne 1924),
can be either enzymes or carriers, are soluble in aqueous media, and
273
are therefore easily extractable. The latter tend to form wide threedimensional meshes that entrap aqueous phases and separated bodies,
like starch granules. Gluten is the most important representative of this
family: it is not soluble in water and therefore can be separated by
washing a dough loaf with hot water to wash away starch carbohydrates and globular proteins. Minor amounts of nonsoluble compounds
remain entrapped in the gluten meshes and represent the unavoidable
contaminants of any gluten preparation. Details about the chemistry
of the cereal proteins are beyond the scopes of this chapter; suffice it
to say that when a dough is prepared from a cereal flour, globular
proteins play mainly a surfactant role that is crucial in stabilizing the
air bubbles formed in a leavening loaf, whereas gluten is responsible
for the overall rheological behavior of the system.
Both globular proteins and gluten fix water molecules, although in
rather different ways. The former are normally solvated at the surface
polar groups and modify their own solvation shell when unfolding and
denaturation take place. Gluten instead uses water molecules as bridges
between the next neighboring chains (Belton 1999) and develops an
extended network, due to a large number of hydrogen bonds (Figure
11.5).
Because of this, gluten can entrap large amounts of interstitial water
within its meshes. Some disulphide bonds provide more robust interand intrachain links and affect the overall extensibility of the network.
Any elongation strain squeezes the interstitial water out of the meshes
so that the next neighboring chains become closer to one another; when
the strain is allowed to relax, water can reoccupy the interstitial regions
and make the meshes swell back to the starting size. This view has
been recently perfected (Belton 2005; Kontogiorgos and Goff 2006),
and it is still adequate to suggest general guidelines for our understanding of the competition of various flour components for the available
water that undergoes displacement between the coexisting aqueous
phases of a cereal flour dough from its early mixing, to proofing and
baking, or freezing and storing.
This a fundamental issue to be considered. Because of the thermodynamic incompatibility (Tolstoguzov 2003) between proteins and
carbohydrates, as well as between different carbohydrates (Liu and Shi
2006), a flour dough is indeed a dispersed system in which several
aqueous phases coexist that can exchange the solvent between one
another.
274
Figure 11.5. Naive picture of the water bridges and direct hydrogen bonds between
hydro-compatible biopolymer chains (which can be referred to as carbohydrates or
gluten proteins).
This partition of water can be detected with thermogravimetry investigations (Fessas and Schiraldi 2005): the related DTG trace shows a
broad signal that can be deconvoluted in two or more components,
each relevant to a given water fraction (Figure 11.6).
The water fraction that sustains the evaporation at mild temperatures
mainly belongs to the imbibing moisture (in the earliest steps) and by
the separated (because of the thermodynamic incompatibility of their
solutes, such as carbohydrates and globular proteins) aqueous phases:
the solvent that evaporates from one aqueous phase is quickly replaced
by the water migrating from any neighboring aqueous phase. As a
result, the dehydration of the samples looks like a single process governed by the core-to-surface diffusion of moisture (Fessas and Schiraldi
2005). The rest of the water (about 15% of the starting overall dough
moisture) tends to remain close to the network forming polymers (mainly
gluten) and can be stripped away only at temperatures above 100 C.
Starch gelatinization is mainly sustained by the former water fraction when the dough is being baked. Because the starch gelatinization
275
Figure 11.6. Deconvolution of the DTG trace obtained from a wheat dough sample.
The high-temperature peak can be related to the water fraction tightly trapped within
the gluten meshes, while the lowest broad peak is related to the imbibing water that
is released through a Fickian diffusion mechanism. The other minor peaks should be
referred to as contributions from moderately bound water fractions.
276
Figure 11.7. Deconvolution of the DTG trace of a manually mixed dough with 42%
moisture content (30 mg sample, 2 C/min heating rate) and starch gelatinization of
a dough undergoing heating at the same heating rate in a open DSC pan. Modified
from Fessas and Sachiraldi (Fessas and Schiraldi 2000).
Nonstarch Carbohydrates
Cereal and pseudo-cereal flours contain some nonstarch carbohydrates
coming from various regions of the seed. A fraction of them is not
water extractable and therefore is segregated from the aqueous phases
of the dough: the role of this fraction has little relevance to the physical
behavior of the other dough components (safe for some sensorial properties of the final products).
277
Figure 11.8. DTG traces of dough prepared with buckwheat (a), wheat (b), and 50%
(w/w) buckwheat/wheat flour (c) (authors unpublished data).
278
meshes, although the temperature gap between the two maxima, which
is related to the looseness of the gluten network, is smaller than in the
trace of a wheat flour dough. This effect is mainly related to nonstarch
and nongluten proteins, which form separate phases (droplets) that do
not allow gluten to attain an extended and tight reticulation (Fessas
and Schiraldi 1998).
Process Applications
The information collectable with TA investigations can be directly
used to improve food formulations and process conditions. Of great
interest are the use of the so-called state diagrams that highlight the
role of the glass transition temperature as a border between high- and
low-molecular-mobility regions, which have been collected in Rooss
book (Roos 1995). Few other examples are helpful to the reader.
Any given thermal treatment, such as cooking, baking, frying, etc.,
corresponds to a thermal history experienced by the system. A direct
reproduction of such history can be difficult to plan when using TA
equipment. Nonetheless, one can inscribe any thermal history in a TTT
(time temperature transformation) diagram that can be defined on the
experimental basis of TA evidence of a given transformation responsible for a specific signal in the TA record. Starch gelatinization monitored with DSC is a good example. The relevant signal is an endothermic
shouldered peak, the area of which, once divided by heatingrate starch-mass product, corresponds to the specific enthalpy change,
H (joules per gram of starch units), accompanying the gelatinization.
The peak area swept at any given T within the (Tonset, Tend) range is
related to the progress degree, , achieved at that T:
(T ) =
(11.3)
279
(11.4)
Since starch gelatinization is an irreversible transformation, the relevant DSC signal reflects the process kinetics and is affected by the
heating rate, = dT/dt. The higher the heating rate, the higher the end
temperature of the signal (there is also a shift of the apparent onset
temperature that is of minor concern for the present discussion,
however). One therefore has to define the temperature range where
starch gelatinization takes place for each and the progress degree
within that range:
d
d dT
d
Q = H = H = H
dt
dT dt
dT
(11.5)
where Q is the recorded heat flux per unit mass of starch and H is
the related enthalpy change.
Once the DSC runs at the selected heating rates are performed, the
increase of on sweeping the related peaks can be determined: more
precisely, one has to detect the temperatures at which attains some
given levels (say = 0.1, 0.2, 0.3, etc.) for each considered (Figure
11.9).
The TTT diagram is a T-versus-t plot, where the straight lines corresponding to the various heating rates considered in the experimental
design of the DSC investigations have to be drawn first.
The selected iso- temperatures (see above) have to be marked
along the corresponding straight line in the TTT diagram. Eventually,
a map of iso- points is obtained that can be used to draw iso- curves
(Figure 11.10).
Remember that the maximum attainable does not depend on ,
being mainly related to the available moisture. This is a third fundamental parameter to be accounted for. The experimental design must
therefore include DSC runs performed with samples of a different
moisture content, and a third axis can be added to the TTT diagram to
account for it.
In the three-dimensional TTT diagram, the iso- loci are surfaces.
For a more practical approach, samples can be partially dehydrated by
0.5
140
120
0.5
T / C
100
0.8
0.6
80
0.4
60
0.1
0.2
40
20
0
0
50
100
150
200
250
t / min
Figure 11.10. Iso-moisture (excess water) TTT plane relevant to starch gelatinization
in bread dough samples. The dotted straight lines correspond to different heating
temperatures (0.5 , 1 , 2 , 5 C/min), while the data points are the drawn from the
-versus-T plots to evidence the attainment of a given level (values reported at the
right-hand side). The tie lines connecting these points are the iso- curves. Modified
from Fessas and Schiraldi (Fessas and Schiraldi 2000).
280
281
Figure 11.11. Bread dough samples: starch gelatinization extent achieved in DSC
open pans. Modified from Fessas and Schiraldi (Fessas and Schiraldi 2000).
heating them in open DSC pans within the furnace of the instrument
at a given heating rate. The run has to be stopped at a temperature, Ti,
that can represent a rough average of the real thermal history experienced by the product. The DSC sample pan is then quickly cooled to
room temperature, sealed, and again heated at the same rate to evaluate
the residual starch gelatinization. The relevant moisture content of the
sample at Ti (and in the second DSC run with sealed pans) can be
eventually determined as the mass loss after an overnight rest at 105 C
after piercing the cover of the sample pan with a needle (Figure 11.11).
This experiment allows determination of the fraction of starch gelatinization that occurred during the first (open pans at variable moisture
content) and the second phase (sealed pans at constant moisture content)
(Fessas and Schiraldi 2000). An ideal experiment should in principle
be performed with a TGA-DSC combined instrument. Unfortunately,
this cannot be a realistic choice, since the vaporization enthalpy of
water (2.3 kJg1) is 2 orders of magnitude larger than the enthalpy of
starch gelatinization. A superposition of separate investigations can
nonetheless be performed (see Figure 11.7).
The corresponding TTT diagram (Figure 11.12) is indeed a curved
section of the three-dimensional plot. On the same diagram, any
thermal history actually experienced by the system, as well as the
282
T / C
100
0.7
0.6
0.5
0.4
80
0.3
0.2
0.1
60
a
40
20
thermal histroy
10
15
t / min
20
25
30
Figure 11.12. TTT diagram corresponding to the data reported in Figure 11.11. The
figure is a projection of a surface of the three-dimensional TTT (moisture-vs.temperature-vs.-time) diagram.
moisture loss that occurred in the real process, can be represented with
a process path. The corresponding curve crosses the iso- surfaces
(Figure 11.13).
The highest attained may not regress and can indeed be referred
to as the maximum progress of starch gelatinization achievable at the
end of the thermal history considered. Obviously, one must take into
account that real systems have a much larger mass than a DSC sample
and therefore experience temperature and moisture gradients on
cooking or baking. For a more detailed simulation of the process, one
therefore must first record the thermal history and the moisture changes
in each region of the real system and then draw the relevant information about the extent of the starch gelatinization achieved.
Another practical use of the information drawn from TA investigations concerns the effects of mechanical treatment, such as kneading,
and layering, that affect the tightness of the gluten network of wheatbased products. Mechanical stresses squeeze some water out of the
gluten phase, thus allowing protein moieties to come closer to one
another and, possibly, to form direct links (mainly hydrogen bonds
[Belton 1999; Belton 2005]). If the product is baked just after the
experience of such mechanical stresses, more water evaporates on
baking, and the final product is harder and brittle. This occurs in bis-
283
Figure 11.13. Sketched representation of the real thermal history experienced by the
product (sigmoid curve) that crosses the iso- surfaces in the TTT diagram. The figure
is a projection of a surface of the three-dimensional TTT (moisture-vs.-temperaturevs.-time) diagram. Modified from Fessas and Schiraldi (Fessas and Schiraldi 2000).
284
A
Figure 11.14. Pictures of the alveolar crumb structure of bread from wheat (A) and
a wheat-buckwheat blend added with soluble polysaccharides (B).
than the other water fractions of the dough. Some visual evidence of
this picture is provided by light microscopy investigations (Autio and
Salmenkallio-Marttila 2001; Autio and Laurikainen 1997) that show
islets of nonstarch hydrocolloids dispersed within the gluten meshes,
apparently hindering the contacts between them and preventing the
formation of a tight network. If this situation is not perturbed by
mechanical stresses, the water loss during baking is smaller than for a
dough with no extra hydrocolloids. The result is a product with broader
crumb alveoli (Fessas and Schiraldi 1998). The nonstarch hydrocolloid
water sinks keep the final product softer for a longer period, thus
acting as antistaling ingredients (Fessas and Schiraldi 2001a). One may
define an optimal shift of the DTG peak that corresponds to the
desired tightness of the final network by adjusting the amounts of extra
hydrocolloids added to the recipe.
The case of bread prepared from a blend of wheat and buckwheat
flours can be a suitable example (Fessas et al. 2008), where the nonstarch carbohydrates were those water-extracted from the buckwheat
hull in a separate process and added as an aqueous solution to the
dough. The bread prepared from a blend of wheat and dehulled buckwheat had a crumb with alveolar distribution quite similar to that of
the crumb of the bread obtained from wheat flour (Figure 11.14),
although with a 45% greater density.
In this modified dough, the effects of soluble nonstarch polysaccharides has been tuned through the counterbalancing action of
285
globular proteins that play the role of surfactants that stabilize the
dough matrix/air interface (Fessas and Schiraldi 1998).
Conclusions
Thermal analysis is very suitable for monitoring the transformations
that take place in cereal-based food products. It requires, however,
specific training for operators because the records obtained from a food
sample usually have to be mathematically treated to unveil the single
contributions coming from largely overlapped events relevant to different components or phases of the system.
Combination with other experimental techniques is nonetheless of
great help, especially when water partition and displacements are
involved in the changes induced by the process.
Specific applications to simulate thermal treatments and adjust the
recipe of a given product can be drawn from either DSC or TGA data.
References
Autio, K. and Laurikainen, T. 1997. Relationships between flour/dough microstructure and dough handling and baking properties. Trends Food Sci Technol,
8:181185.
Autio, K. and Salmenkallio-Marttila, M. 2001. Light microscopic investigations of
cereal grains, doughs, and breads. Lebensm-Wiss U Technol, 34:1822.
Belton, P.S. 1999. On the elasticity of wheat gluten. J Cereal Sci, 29:103107.
Belton, P.S. 2005. New approaches to study the molecular basis of the mechanical
properties of gluten. J Cereal Sci, 41:203211.
Biliaderis, C.G., Page, C.M., Maurice, T.J., and Juliano, B.O. 1986. Thermal characterization of rice starches: A polymeric approach to phase transitions of granular
starch. J Agric Food Chem, 34:614.
Bulpin, P.V., Welsh, E.J., and Morris, E.R. 1982. Physical characterization of amylose-fatty acid complexes in starch granules and in solution. Starch/Staerke,
34:335339.
Courtin, C.M. and Delcour, J.A. 1998. Wheat-Derived Arabinoxylans. J Agric Food
Chem, 46:40664073.
Eliasson, A.C. 1994. Interactions between starch and lipids studied by DSC.
Thermochim Acta, 246:343356.
Fessas, D. and Schiraldi, A. 1998. Texture and staling of wheat bread crumb: Effects
of water extractable proteins and pentosans. Thermochim Acta, 323:1726.
286
287
Riva, M., Piazza, L., and Schiraldi, A., 1991. Starch gelatinization in pasta cooking:
Differential flux calorimetry investigations. Cereal Chem, 68:622627.
Riva, M., Schiraldi, A., and Piazza, L. 1994. Characterization of rice cooking:
Isothermal and differential scanning calorimetry investigations. Thermochim Acta,
246:317328.
Roduit, B. 2000. Computation aspects of kinetic analysis. Thermochim Acta,
355:171180.
Roos, Y.H. 1995. Phase Transitions in Foods. Academic Press: New York.
Schiraldi, A. 2003. Phenomenological kinetics: An alternative approach. J Therm
Anal Calorim, 72:885900.
Schiraldi, A. and Fessas, D. 2003. Classical and Knudsen thermogravimetry to check
states and displacements of water in food systems. J Therm Anal Calorim,
71:221231.
Schiraldi, A., Piazza, L., and Riva, M., 1996. Bread staling: A calorimetric approach.
Cereal Chem, 73:3239.
Schwartzberg, H.G. 1976. Effective heat capacities for the freezing and thawing of
food. J Food Sci, 41:152156.
Slade, L. and Levine, H. 1991. Beyond water activity: Recent advances based on an
alternative approach to the assessment of food quality and safety. CRC Crit Rev
Food Sci Nutr, 30:115359.
Tolstoguzov, V.B. 2003. Some thermodynamic considerations in food formulation.
Food Hydrocolloids, 17:123.
Vittadini, E., Dickinson, L.C., Lavoie, J.P., Pham, X., and Chinachoti, P. 2003. Water
mobility in multicomponent model media as studied by 2H and 17O NMR. J Agric
Food Chem, 51:16471652.
Vodovotz, Y., Hallberg, L., and Chinachoti, P. 1996. Effect of aging and drying on
thermomechanical properties of white bread as characterized by Dymanic
Mechanical Analysis (DMA) and Differential Scanning Calorimetry (DSC).
Cereal Chem, 73:264270.
Wasserman, L.A., Signorelli, M., Schiraldi, A., Yuryev, V., Boggini, G., Bertiniand,
S., and Fessas, D. 2007. Preparation of wheat resistant starch: Treatment of gels
and DSC characterization. J Therm Anal Calorim, 87:153157.
Yuryev, V.P., Krivandin, A.V., Kiseleva, V.I., Wasserman, L.A., Genkina, N.K.,
Fornal, J., Blaszczakb, W., and Schiraldi, A. 2004. Structural parameters of amylopectin clusters and semi-crystalline growth rings in wheat starches with different
amylose content. Carbohydrate Res, 339:26832691.
Zobel, H.F. 1988. Starch crystal transformations and their industrial importance.
Starch/Staerke, 40:4450.
Chapter 12
Importance of Calorimetry in Understanding
Food Dehydration and Stability
Yrj H. Roos
Introduction
Phase and State Transitions of Food Components
Calorimetric Glass Transition Measurement
Dielectric and Mechanical Relaxations
Thermal Analysis in Characterization of Food
Systems
The Frozen State of Foods Systems
State Diagrams and Dehydration
Spray-Drying
Freeze-Drying
Glass Transition and Stability of Dehydrated Materials
Conclusions
References
289
292
293
297
298
301
303
305
306
307
308
309
Introduction
Dehydration involves removal of solvent water from dissolved and
hydrated food components. The process requires heat for evaporation
or sublimation of water and concentration of food solids at high levels.
This results in water removal to an almost anhydrous state of food
components. Traditional dehydration processes are based on empirical
knowledge of food material properties and processing needs to achieve
desired product characteristics. Advanced dehydration processes, such
289
290
291
292
293
Equilibrium
Liquid
(Pressure)
C
oo
lin
g
R
ap
id
Cooling
Crystallization
ow
Sl
tin
g
g
in
ol
Co
GLASS
ea
Heating
g
in
at
He
Equilibrium
Liquid
Cooling
RUBBER
Nonequilibrium
Solid
Heating
CRYSTAL
a
Pl
Calorimetric
Measurements
eh
yd
ra
tio
n
a
iz
tic
n
tio
a
t
ur
at
io
n
Equilibrium
Solid
So
lu
bi
liz
at
io
n
SOLUTION
MELT
294
Exotherm
Onset
Midpoint
Time-dependent
changes
C p
Endset
Endotherm
T
Figure 12.2. Schematic representation of typical DSC curves obtained for amorphous materials in heating over their glass transition temperature range. The glass
transition temperature, Tg, is often taken from the onset temperature of the glass
transition or as the midpoint value corresponding to 50% change in heat capacity
occurring over the glass transition. Glass transition may involve an exotherm or an
endotherm corresponding to differences in glass formation (heating/cooling rates;
solvent removal/sorption rates).
295
V
H
S
Anomalous Changes
In Thermodynamic
Properties Depending
on Glass Properties
Non-equilibrium
State
Equilibrium
State
Non-equilibrium
State
Liquid
Supercooled
liquid
Hm
Crystal
Glass
100-150C
Tg
Tm
Figure 12.3. Thermodynamic states of materials. The equilibrium liquid state occurs
at and above the equilibrium melting temperature, Tm. At lower temperatures, the
crystalline solid state may exist at equilibrium. Noncrystalline systems can be supercooled liquids or they may become solidlike supercooled materials (glasses) below
the glass transition temperature, Tg. However, the glassy state is a nonequilibrium and
time-dependent state. It may have various molecular arrangements with different
enthalpy, H, entropy, S, and volume, V, states. Melting of crystals involves heat of
melting, Hm, but the glass transition involves no latent heat of the transition.
Storage modulus or
dielectric constant
Increasing frequency
Mechanical and
dielectric relaxations
Loss modulus or
dielectric loss
and relaxations
Tg
relaxation
TEMPERATURE
Figure 12.4. Mechanical and dielectric relaxations of materials in the glassy state
and around the glass transition.
296
297
298
299
Glass transition
ENDOTHERMAL HEAT FLOW
Tg
Melting
T cr
Tm
Crystallization
TEMPERATURE
300
120
60
0.6
40
20
0
Tg
-20
-40
Critical Water
Activity (25C)
Critical Water
Content (25C)
-60
0
10
20
30
0.4
0.2
WATER ACTIVITY
TEMPERATURE (C)
80
-80
0.8
Extrapolated
GAB Sorption
Isotherm
Time-dependent
crystallization
100
0
40
Figure 12.6. Glass transition and water sorption behavior of lactose. Water sorption
results in lactose crystallization as the glass transition at the observation temperature
is exceeded, and the water activity and water content become higher than the critical
values.
301
302
Initial Tg
Unfrozen state
Annealing time
t0
Partial freeze-concentration
Tm
t1
t2
T'g
Maximal freeze-concentration
T'm
t3
Annealing temperature
TEMPERATURE
Figure 12.7. A schematic representation of time-dependent ice formation in freezeconcentrated food systems. A rapidly cooled system shows a glass transition corresponding to the solute-water ratio of the solution when no ice is formed. Ice formation
may be achieved by annealing (isothermal holding) at a temperature below Tm. This
can be detected from an increasing glass transition temperature, Tg, for the freezeconcentrated unfrozen solute phase and an increasing size of the ice-melting endotherm. At maximum ice formation, more ice cannot be formed, and the glass transition
occurs at initial water content-independent temperature, Tg, and is followed by onset
of ice melting at Tm.
TEMPERATURE
Tm
20
1.3x10 5
1.0x10
303
Supercooled
liquid
Tm
low
sf
u
co
olid
Vis
ys
Tg
ation
ion
-relax
ansit
r
t
s
Glas
0
ss
Gl a
Thermal plasticization
Water plasticization
1.0
Figure 12.8. Schematic state diagram showing the decrease in glass transition with
increasing water content and decreasing relaxation times at increasing levels of
thermal or water plasticization. Maximum ice formation takes place time dependently
over the temperature range from the glass transition temperature of the maximally
freeze-concentrated unfrozen phase, Tg, and onset of ice melting in the maximally
freeze-concentrated unfrozen phase, Tm. Equilibrium ice melting occurs according to
the equilibrium ice-melting temperature, Tm, curve.
50
-50
-100
Glass transition
Solubility
range
(equilibrium mixture
Time-dependent
of - and
crystallization
-lactose)
Supercooled
Calorimetric
transition
Tg
liquid
temperatures
T'm
T'g
Temperature (C)
100
Tg
Glass
-150
0.0
Gla
s
304
0.2
0.4
0.6
0.8
Weight Fraction of Lactose
C'g
1.0
Figure 12.9. State diagram of lactose with the glass transition temperature, Tg, curve,
and transition temperatures for maximally freeze-concentrated lactose solutions
(Tg is the glass transition temperature of a maximally freeze-concentrated solution,
and Tm is onset temperature of ice melting in a maximally freeze-concentrated
solution).
305
Spray-Drying
Spray-drying is an efficient dehydration method for a large variety of
liquid materials and slurries, which can be converted to small liquid
droplets using rotating disks or pressure nozzle atomizers. The tiny
droplets can be dehydrated in hot air within seconds, which allows
continuous production of free-flowing powders. Although the principle
of the process is relatively simple, phase and state transitions of food
solids have a significant impact on whether the materials can be spraydried successfully or whether the powders have free-flowing properties
in handling, packaging stages, storage, and use.
It has been suggested that formation of the glassy state from solids
in spray-drying, particularly the glass-forming properties of carbohydrates, have a correlation with spray-drying behavior of fruit juices and
materials rich in sugars (Bhandari and Howes 1999). These materials
are often extremely difficult to dehydrate because the solids tend to
stick on drier surfaces and cake inside dehydration- and powderhandling equipment. Stickiness is probably the most important property in establishing criteria for the suitability of food materials to
spray-drying. Studies of glass transitions of dehydrated sugars and
high-sugar products have confirmed that stickiness is related to the
glass transition of amorphous powders (Roos and Karel 1990). Based
on the knowledge of phase and state transitions in dehydration of
liquids with dissolved substances, it may be assumed that the rapid
removal of water causes vitrification of the liquid droplets (i.e., solids
in the droplets dehydrate and form glasslike structures) within a short
time and formation of a solid particle surface (Roos 2004). Glass transition of the solids at the surface layer of a drying droplet is a key
parameter in defining stickiness behavior of the particles and formation
of liquid bridges occurring in subsequent caking as a result of rapid
decrease in surface viscosity above the glass transition. Therefore,
materials with an anhydrous glass transition below room temperature
cannot be spray-dried, as they cannot be converted to a solid state at
room temperature.
The viscosity changes resulting from the glass transition also can be
used to control agglomeration of fine particles and in the manufacturing of instant powders (Roos 1995). In such processes, it is essential
to allow controlled stickiness on particle surfaces and adhesion of
particles to form clusters. The formation of clusters is followed by
306
Freeze-concentrated
unfrozen solute phase
Freeze-dried
glass solute
membranes
Ice
Pores
Freeze-drying
below T m
id
Sol
Flo
w
Freeze-drying
above T m
Collapsed
liquid
Figure 12.10. The role of onset temperature of ice melting, Tm, in successful freezedrying and liquid flow resulting in collapse as ice temperature exceeds Tm in a freezedrying process.
removal of water and cooling to solidify the surfaces into the glassy
state. The final product will have larger particles and remain free
flowing and stable at appropriate storage conditions.
Freeze-Drying
The Tg and Tm temperatures of biological materials are extremely
important determinants of appropriate operation parameters in freezedrying (Roos 2004). Freeze-drying requires that dissolved substances
are freeze-concentrated to an almost solid state and that the highly
viscous state is retained throughout the dehydration process; that is,
the material should consist of solid ice and a freeze-concentrated, solid,
glassy, unfrozen phase (Figure 12.10). Ice melting above Tm has a
dramatic plasticization effect in a freeze-concentrated system and
results in liquid flow. The effect of state transitions and ice melting
above Tm are described for freeze-drying in Figure 12.10. Accordingly,
the highest allowable pressure (or ice temperature) in freeze-drying is
defined by the initial melting temperature of ice in the system. At
conditions allowing melting, flow may occur, and some dehydration
occurs from the liquid state; the process no longer can be referred to
as freeze-drying (Figure 12.10).
307
Stability Zone
Solid
Glass Transition
Fermis Model
(M. Peleg)
CRITICAL ZONE
VISCOUS FLOW
Increasing Diffusion
Structural Transformations
SOLID
Crispness
RELAXATION TIME
Months
Days
Hours
Minutes
Seconds
Glassy State
Flow
LIQUID
Critical Zone
Mobility Zone
Highly time-dependent
Instant changes
Years
Hardening, Cracking
308
Figure 12.11 Changes in relaxation times as a result of thermal or water plasticization in food systems. Around and above the glass transition rapidly appearing liquidlike properties of the materials result in dramatic changes in mechanical properties
and diffusion. The changes in mechanical properties around glass transition may be
modeled using the Fermi relationship (Peleg 1993).
the crystalline form of lactose produced is dependent on the crystallization conditions (Jouppila et al. 1997; Haque and Roos 2005). Most
crystals formed are anhydrous, but at the higher storage humilities
increasing amounts of -lactose monohydrate is formed. Crystallization
of amorphous lactose in sealed packages and in bulk storage also
results in an increase in water activity and acceleration of most deteriorative changes, such as browning reactions and oxidation.
Conclusions
Thermal and calorimetric properties of food and biological materials
at various water contents are extremely important determinants of
their dehydration and stability characteristics. Calorimetric measurements provide data for selection of appropriate dehydration parameters
and manipulation of solids composition to enhance dehydration and
improve storage stability. Several dehydrated materials, particularly
spray-dried and freeze-dried, exist as amorphous, glassy solids.
Formation of a solid structure contributes to the success of dehydration
processes and the quality characteristics of dehydrated materials.
Knowledge of glass transitions and ice-melting properties of sensitive
309
310
Chapter 13
High-Pressure Calorimetry and
Transitiometry*
Stanislaw L. Randzio and Alain Le Bail
Introduction
High-Pressure Calorimetry
Scanning Transitiometry
Applications
Water in Pork Muscle
Frozen Water Ratio in Gelatine Gels
Pressure Shift Freezing
Gelatinization of Starch
Phase Stability of Systems Containing Lipids
Conclusions
References
311
313
317
324
324
326
329
330
336
337
338
Introduction
The discovery by Bridgman in 1914 (Bridgman 1914a) of a pressureinduced coagulation of egg white enhanced research on the influence
of pressure on biological and food systems. From the results obtained
over the years, it became evident that the use of pressure can lead to
developments of food products with new properties, such as highpressure-processed jams prepared by nonheating food processing,
which entered the market in Japan in 1992 (Hayashi 2002). Other
important technological applications are concerned with high-pressure
processing of food, especially freezing-thawing and crystallization
*M. Malecki is acknowledged for technical assistance in the preparation of this
chapter.
311
312
313
High-Pressure Calorimetry
Figure 13.1 presents a schematic diagram of a high-pressure calorimetric system developed by LeBail et al. (Le Bail et al. 2001; Zhu et al.
2004). The high-pressure calorimetric system is composed of a differential calorimetric detector made from 220 thermocouples interconnected between the two calorimetric vessels, high-pressure pump,
hydraulic fluid reservoir, pressure detector (Asco Instruments, France),
and a circulating liquid thermostat. The pressure in the system was
controlled by proportional-integral-derivative computer software
through a stepping motor and a gear box of the high-pressure pump
(Nova Swiss, Switzerland). The processing of the calorimetric signal
was performed with computer software. The investigated substance
was always placed in a small flexible plastic pouch. A detailed view
of the calorimetric system elements can be seen in Figure 13.2. The
high-pressure vessels connected to the high-pressure hydraulic system
by flexible stainless steel capillary tubing can be easily introduced into
the cavities of the differential calorimetric detector, placed in the calorimetric block, which in turn is surrounded with a copper coil in which
Figure 13.1. Schematic diagram of a differential high-pressure calorimeter: (1) differential calorimetric detector, (2, 3) calorimetric vessels, (4) high-pressure pump, (5)
hydraulic fluid reservoir, (6) pressure detector, (7) circulating liquid thermostat, (8),
pressure control, (9) calorimetric signal processing, (10) investigated substance in a
flexible plastic pouch.
314
Figure 13.3. High-pressure calorimetric experimental vessel: (1) body of the vessel
made from stainless steel, (2) polyethylene pouch containing the investigated substance, (3) nitrile O-ring placed on a plug, (4) threaded plug holding the closing in
place.
315
316
latent heat of fusion equal to the value derived from Bridgmans highpressure volumetric data (Bridgman 1912) within a 3% agreement,
which is a good validation of the instrument correctness. It is worth
noting that the Calvet-type calorimeters are also suitable instruments
for measurements performed under typical processing conditions of
high hydrostatic pressure and temperature of various processes under
well-determined pressure and temperature conditions that are important for food science and technology. Such instruments have been used
either as a single calorimetric detector (Chourot, LeBail, and Chevalier
2000) or as a differential calorimetric device (Randzio, Grolier, and
Quint 1994). In differential mounting, a Setaram C80 calorimeter was
used in an upside-down position. This permitted using differential
calorimetric vessels fixed on a laboratory table and connected to the
high-pressure pump and pressure detectors with rigid stainless steel
tubing, allowing performance of direct calorimetric measurements up
to 400 MPa. This was the first pressure-controlled scanning calorimeter
with linear pressure variations at rates from 0.5 kPa/s (30 kPa/min) to
0.2 MPa/s (12 MPa/min).
When performing high-pressure calorimetric measurements, one
should realize that the significance of the calorimetric signal depends
on the use of the pressure-transmitting fluid. If the pressure is transmitted to the calorimetric vessel through the substance under investigation
(liquid, liquid suspension, liquid emulsion, or even a paste), the heat
developed as a result of pressure variation is proportional to the coefficient of thermal expansion of the substance under investigation
(Randzio 1985). This is because the mass of the substance contained
in the calorimetric vessel varies, m = VE/V, where VE is the internal
volume of the calorimetric vessel and V is the molar or specific volume
of the substance under investigation, the latter one being pressuredependent. This is a kind of open mass vessel (quasi-constant volume),
in which the substance under investigation entirely fills the calorimetric vessel and at least a part of the external tubing connecting to the
pressure generator. In investigating solid samples, the pressure must
be transmitted by a fluid. Thus, the thermal effect developed due to a
pressure variation is composed of two contributions. The first is proportional to (dV/dT )p of the solid sample under investigation and the
second to the thermal expansion coefficient of the pressure-transmitting fluid. However, an exact separation of the two contributions
requires careful treatment of the data (Rodier-Renaud et al. 1996).
317
Scanning Transitiometry
Figure 13.5 presents schematically a relatively new technique called
scanning transitiometry (Randzio 1996). The function of scanning
transitiometry consists of scanning one of the three variables ( p, V, or
T ) when the second is kept strictly constant. During the scanning, the
variations of the dependent variables and the associated calorimetric
signal are simultaneously recorded. From these two quantities and the
scanned variable, two thermodynamic derivatives, thermal and mechanical, are simultaneously determined for the system under study. Figure
13.6 presents four thermodynamic situations covered by scanning transitiometry, where from the state variables and the heat effect one can
determine four pairs of thermodynamic derivatives (Randzio 1997).
Each of the situations has specific applications, which prove its particular utilities. In studying food systems, the most useful is the use of
temperature as a scanned variable at constant pressure and the use of
pressure as a scanned variable at constant temperature. In the former
case, the output variables are heat capacity at constant pressure and
thermal expansion. After proper integration, one obtains enthalpy and
volume variations caused by the applied temperature change. In the
latter case, the output variables are pressure derivative of entropy and
compressibility. After proper integration, one obtains heat of transition
and associated volume change. Sometimes, especially for transitions
volume 15 cm3
5 +106 5 +103 cm3/s
temperature
pressure
heat flux
223673K
0.15mK/s
COMPUTER
0.1400MPa
0.0010.05MPa/s
107101W
318
OUTPUTS
T = const
P = f(t)
PVT
Thermal
Mechanical
T = const
V = f(t)
PVT
Thermal
Mechanical
P = const
T = f(t)
PVT
Thermal
Mechanical
V = const
T = f(t)
PVT
Thermal
Mechanical
(S/P)T = (V/T)P
(V/P)T
(S/P)T = (P/T)V
(V/P)T
(H/T)P
(V/T)P
(U/T)V
(P/T)V
319
calorimetric detector
upper entries
heat.-cool. shield
air
cone plug
TDIFF
calorimetric
signal
TSECURITY
80W
300W
ampoule
temp.
control
calorimetric detector
investigated
substance
computer
spring
data acquisition
&
process control
high pressure
tubing
heat insulation
calorimetric
block
dry air flow
cooling fluid
pump
Pt 100
measuring vessel
step motor
control
pressure
detector
reference vessel
320
321
Figure 13.8. Transitiometric vessels: (a) standard high-pressure vessels; (b) the
vessels kept in a holder to facilitate reproducible closing and opening of the vessel
by a dynamometric wrench; (c) a sample of a gelatinized sample of 50% aqueous
suspension of wheat starch pushed out after a transitiometric isobaric experiment
under 90 MPa.
The method presented here is rather simple and safe in practice. The
total volume of the liquid phase under pressure is only about 20 ml; the
energy accumulated in it is rather small and not dangerous. The mercury
used as a hydraulic fluid is always contained in a closed space. In case of
a leak, the mercury is collected on a special protecting plate. The calorimetric vessels are conveniently and reproducibly closed and opened
with a torque wrench with the vessels placed in a specially designed
holder. Figure 13.8ac presents a view of transitiometric vessels: standard high-pressure vessels, the vessels kept in a holder to facilitate
reproducible closing and opening of the vessel by a dynamometric
wrench, and a sample of a gelatinized sample of a 50% aqueous suspension of wheat starch pushed out after a transitiometric experiment.
Because of the high sensitivity of the instrument, some precautions
must be taken to ensure valid measurements. In the case of the calorimetric signal, the main precaution is to carefully compensate the
thermal balance of the differential calorimetric vessels. It is also important to keep the initial mercury level always in the same position, just
above the entry to the calorimetric detector zone. Adjustment of the
mercury level is easily done with the motorized pump. With respect to
the volumetric component, it is very important that displacement
322
(13.1)
The mean deviation between Equation 13.1 and the calibration data is
1.4%. The reproducible resolution of the calorimetric detector varies
from 1.3 107 W at 303 K to 1.6 107 W at 430 K. As reported previously, for properly designed experimental vessels, the energetic
calibration constant of the calorimetric detector does not depend on
pressure (Randzio, Grolier, and Quint 1994).
The volumetric calibration of the high pressure pump was performed
by weighing 11 mercury samples displaced by known numbers of
motor steps. Each motor step corresponded to a displacement of
(5.22 0.03) 106 cm3.
The volumetric calibration of the high pressure pump and both
energetic and temperature calibrations of the calorimetric detector
were verified by test measurements using isobaric fusion of benzene,
for which both enthalpy and volume of the transition are exactly known
(Bridgman 1914b). Figure 13.9 presents an example of such measurements performed at a scanning rate of 2.5 mK s1 at 100 MPa. The mean
results from eight independent measurements gave the following:
fusV(100 MPa) = 0.1038 0.0028 cm3g1 [the respective literature
323
120
10
Endo
100
80
50
60
70
40
90
110
130
299
20
b
301
303
305
Temperature (K)
30
307
0
309
Figure 13.9. Example of simultaneous transitiometric traces (heat flux and volume
variations) of isobaric melting at 100 MPa of 0.3858 g of benzene used as a verification test for thermal and volumetric calibrations of the transitiometer used in the
present study: (a) heat flux, (b) dV/dT.
324
Figure 13.10. Volume variations during isobaric temperature scans at various pressures with only hydraulic liquid (mercury) present in the system. The data at 10 MPa
and 100 MPa are shifted by +0.3 mm3K1 and 0.3 mm3K1, respectively, to avoid
overlapping.
325
326
20
40
60
80
Pork
100
0.1
Water
50
100
Pressure (MPa)
150
200
Figure 13.11. Comparison of thawing heat flux of pure ice and frozen pork muscle
induced by pressure scans at a rate of 5 kPa/s (0.3 MPa/min) at 263.1 K.
10
20
30
268.0K
40
0.1
25
50
175
200
250
Figure 13.12. Thawing heat flux of frozen pork muscles induced by a pressure scan
at a rate of 5 kPa/s (0.3 MPa/min) at various temperatures.
sure of 111.7 MPa. At the transition, the temperature scan was perturbed
because of a rapid release of heat during fast crystallization of ice.
Frozen Water Ratio in Gelatine Gels
The same experimental procedure as that described above for pure
water and water in pork muscle has also been used with gelatine gels
containing 2% and 10% of dry gelatine (Chevalier-Lucia et al. 2003).
327
259
258
1.5
Temperature
1.0
0.5
257
Temperature (K)
2.0
Heat flux
0.0
0.5
256
0
1200
3600
Time (s)
6000
Figure 13.13. Heat flux of crystallization of water in pork muscle induced by cooling
under pressure of 111.7 MPa.
(13.2)
(13.3)
328
400
300
200
100
253
258
263
268
273
Temperature (K)
Figure 13.14. Evolution of the latent heat of pure water (square, experimental data;
dash, Bridgmans data) and of gelatine gels (circle, 2%; diamond, 10% in dry matter)
according to the melting temperatures.
Figure 13.15. Evolution of the ratio of latent heat of melting gelatine gels and latent
heat of melting water as a function of melting temperature (or respective pressure):
circle, 2% gelatine gels; diamond, 10% gelatine gels.
329
Nucleation
Supercooling
Water
Cooling
F
Cooling
Temperature (K)
273
Ice-I
C
G
D
Depressurization
252
0.1
Pnuc
330
300
253.1
200
Heat flux
100
0
252.1
0
Pressure
100
0
Temperature (K)
400
1200
2400
Time (s)
3600
251.1
4800
Figure 13.17. Typical profiles of pressure, heat flux, and temperature during pressure-shift freezing of pork muscle under pressure of 199 MPa at 253.1 K.
Temperature (K)
40
Ice ratio and regression curve
273
30
20
253
10
50
100
150
Pressure (MPa)
200
0
250
50
Phase-change curve
0.1
331
Figure 13.18. Ratios of ice crystal to sample mass instantaneously formed after
depressurization during pressure-shift freezing of pork muscle (74.2% moisture
content) at various initial pressures, with temperatures slightly higher than the corresponding phase-change points of water.
332
Endo
46.8
Heat flux difference (mW/g)
28
30
56.0
32
60.3
34
64.8
36
38
40
300
310
320
330
340
350
360
370
380
Temperature (K)
Figure 13.20. Transitiometric traces (heat flux and dV/dT per gram of dry starch)
obtained simultaneously and under the process conditions of pressure and temperature
by scanning temperature at a rate of 2.5 mK s1 at various pressures for a starch-water
emulsion (56.0 wt % total water content). A, exothermic transformation; N, endothermic transition.
333
334
Figure 13.21. Division of dV/dT for the main endothermic transition into a positive
dV/dT due to swelling and a negative dV/dT due to the transition itself.
335
0.1
10
60
100
3.52 0.07
3.12 0.12
2.65 0.07
2.45 0. 7
0.788 0.038
0.647 0.040
0.586 0.035
1.53 0.10
1.22 0.05
0.683 0.036
320.5 0.5
319.5 0.5
318.1 1.0
317.9 0.6
Source: Data at 0.1 MPa are from Randzio et al. (Randzio, Flis-Kabulska, and Grolier 2002).
336
337
Figure 13.22. (a) PT phase diagram of lipids (cocoa butter, palm oil, copra oil) and
water determined from high-pressure calorimetric measurements. (b) Pressure dependence of latent heat of fusion of lipids (cocoa butter, palm oil, copra oil) and water
determined from high-pressure calorimetric measurements.
338
References
Bridgman, P.W. 1912. Water in the liquid and five solid forms under pressure. Proc
Am Acad Arts Sci, 47:439.
Bridgman, P.W. 1914a. The coagulation of albumin by pressure. J Biol Chem, 19:
511.
Bridgman, P.W. 1914b. Change of phase under pressure. I. Phase diagrams of eleven
substances. Phys Rev, 3:153.
Cheftel, J.C., Levy, J., and Dumay, E. 2000. Pressure-assisted freezing and thawing:
Principles and potential applications. Food Rev Int, 16:453.
Cheftel, J.C., Thiebaud, M., and Dumay, E. 2002. Pressure assisted freezing and
thawing: A review of recent studies. High Press Res, 22:601.
Chevalier, D., Le Bail, A., and Ghoul, M. 2002. Freezing and ice crystals formed in
cylindrical model food. Part II. Comparison between freezing at atmospheric pressure and pressure shift freezing. J Food Eng, 46:287.
Chevalier, D., Sequeira-Munoz, A., Le Bail, A., Simpson, B.K., and Ghoul, M. 2001.
Effect of freezing conditions and storage of ice crystals and drip volume in turbot
(Scophthalmus maximus), evaluation of pressure shift freezing vs. air-blast freezing. Innov Food Sci Emerg Technol, 1:193.
Chevalier-Lucia, D., Le Bail, A., Ghoul, M., and Chourot, J.M. 2003. High pressure
calorimetry at sub-zero temperature: Evaluation of the latent heat and frozen water
ratio of gelatine gels. Innov Food Sci Emerg Technol, 4:361.
Chourot, J.M., LeBail, A., and Chevalier, D. 2000. Phase diagram of aqueous solution
at high pressure and low temperature. High Press Res, 19:191.
Douzals, J.P., Perrier-Cornet, J.M., Coquille, J.C., and Gervais, P. 2001. Pressuretemperature phase transition diagram for wheat starch. J Agric Food Chem, 49:
873.
339
Fennema, O.R. 1973. Nature of freezing process. In: Low Temperature Preservation
of Foods and Living Matter, O.R. Fennema, W.D. Powrie, and E.H. Marth, editors,
pp. 151222. Marcel Dekker: New York.
Hayashi, R. 2002. High pressure in bioscience and biotechnology: Pure science
encompassed in pursuit of value. Biochim Biophys Acta, 1595:397.
Hayert, M., Le Bail, A., Rigenbach, M.H., and Gruss, E. 2003. High pressure calorimetry as a tool to monitor phase transitions in foods: Application to water and
selected lipids. In: Advances in High Pressure Bioscience and Biotechnology II,
R. Winter, editor. Springer Verlag: London.
Kaletun, G. and Breslauer, K.J. 1996. Construction of wheat-flower state diagram.
J Therm Anal, 47:1267.
Kalichevsky, M.T., Knorr, D., and Lillford, P.J. 1995. Potential applications of highpressure effects on ice-water transitions. Trends Food Sci Tech, 6:253.
Knorr, D. 1999. Process assessment of high-pressure processing of foods: An overview. In: Processing Foods: Quality Optimization and Process Assessment, F.A.R.
Oliveira and J.C. Oliveira, editors, p. 249. CRC Press: Boca Ration, FL.
Le Bail, A., Boillereaux, L., Davenel, A., Hayert, M., Lucs, T., and Monteau, J.Y.
2003. Phase transition in foods: Effect of pressure and methods to asses or control
phase transition. Innov Food Sci Emerg Technol, 4:15.
Le Bail, A., Chevalier, D., Chourot, J.M., and Monteau, J.Y. 2001. High pressure
calorimetry. Comparison of two systems (differential vs. single cell). Application
to the phase change of water under pressure. J Therm Anal Calorim, 66:243.
Pham, Q.T. 1987. Calculation of bound water in frozen food. J Food Sci, 52:210.
Randzio, S.L. 1985. Scanning calorimeters controlled by an independent thermodynamic variable: Definitions and some metrological problems. Thermochim Acta,
89:215.
Randzio, S.L. 1996. Scanning transitiometry. Chem Soc Rev, 25:383. Retrieved from:
http://www.transitiometry.com
Randzio, S.L. 1997. State variables in calorimetric investigations: Experimental
results and their theoretical impact. Thermochim Acta, 300:29.
Randzio, S.L. 1998. Recent developments in calorimetry. Ann Rep Prog Chem, Sect
C, 94:433.
Randzio, S.L. 2002. Recent developments in calorimetry. Ann Rep Prog Chem, Sect
C, 98:157.
Randzio, S.L., Flis-Kabulska, I., and Grolier, J.P.E. 2002. Re-examination of phase
transitions in the starch-water system. Macromolecules, 35:8852.
Randzio, S.L., Grolier, J.P.E., and Quint, J.R. 1994. An isothermal scanning calorimeter controlled by linear pressure variations from 0.1 to 400 MPa. Calibration and
comparison with the piezothermal technique. Rev Sci Instrum, 65:960.
Randzio, S.L. and Orlowska, M. 2005. Simultaneous and under the process conditions
of pressure and temperature analysis of thermal and volumetric properties of starch
gelatinization over wide pressure and temperature ranges. Biomacromolecules,
6:3045.
Rodier-Renaud, L., Randzio, S.L., Grolier, J.P.E., Quint, J.R., and Jarrin, J. 1996.
Isobaric thermal expansivities of polyethylenes with various crystallinities over
340
the pressure range from 0.1 MPa to 300 MPa and over the temperature range from
303 K to 393 K. J Polym Sci Pol Phys Ed, 34:1229.
Rubens, P. and Heremans, K. 2000. Pressure-temperature gelatinization phase diagram
of starch: An under the process conditions of pressure and temperature Fourier
transform infrared study. Biopolymers, 54:524.
Smeller, L. 2002. Pressure-temperature phase diagrams of biomolecules. Biochim
Biophys Acta, 1595:11.
Stute, R., Heilbronn, R.W., Boguslawski, S., Eshtagi, M.N., and Knorr, D. 1996.
Effects of high pressure treatment on starches. Starch/Strke, 48:399.
Zhu, S., Bulut, S., Le Bail, A., and Ramaswamy, H.S. 2004. High-pressure differential
scanning calorimetry (DSC): Equipment and technique validation using water-ice
phase-transition data. J Food Proc Eng, 27:359.
Zhu, S., Ramaswamy, H.S., Le Bail, A. 2004. High-pressure differential scanning
calorimetry: Evaluation of phase transitions in pork muscle at high pressures. J
Food Proc Eng, 27:377.
Zhu, S., Ramaswamy, H.S., and Le Bail, A. 2005. High-pressure calorimetric evaluation of ice crystal ratio formed by rapid depressurization during pressure-shift
freezing of water and pork muscle. Food Res Int, 38:193.
Chapter 14
Calorimetric Analysis of Starch Gelatinization
by High-Pressure Processing
Kelley Lowe and Gnl Kaletun
Introduction
Gelatinization of Starch by Heat
Gelatinization of Starch by High Hydrostatic Pressure
High-Pressure Processing of Wheat Starch Suspensions
Storage of Gelatinized Starch
Conclusions
References
341
342
344
344
347
348
349
Introduction
Starches are used in many food products to increase viscosity or to
form gels. Because starch is insoluble in water, a mixture of starch
with water forms a suspension. Starch granules in suspension swell
with heat, and the viscosity of the suspension increases, depending on
starch concentration. Thermal processing changes the physicochemical
properties of starch, such as increased water solubility and development of viscoelastic behavior (Fennema 1996; Jobling, 2004).
Because starch affects the texture of food products, characterization
of aqueous starch suspension behavior and its interaction with other
food additives under conditions relevant to food processing and storage
is important for assessment of food stability. This chapter provides
a review of starch characterization studies by calorimetry relevant
341
342
0.32
65.72C
2.416J/g
0.34
0.36
70.33C
58.30C
1.673J/g
0.38
0.40
63.63C
0.42
45
Exo Up
50
55
60
65
70
75
80
85
90
Universal V2.6D TA Instruments
Temperature (C)
Figure 14.1. DSC thermograms of 15% (w/w) native corn and wheat starch suspensions. Corn starch (solid line) and wheat starch (dashed line).
30
40
50
60
70
80
90
100
Universal V2.6D TA Instruments
Temperature (C)
Figure 14.2. DSC thermograms of wheat starch (A), wheat starch-sugar (B), and
wheat starch-sugar-milk protein (C). Total solids: 30% for all cases.
343
344
345
Sample preparation
A Hydrostatic High Pressure Unit (Quintas QFP-6; ABB Autoclave
Systems, Columbus, OH) with a maximum pressure of 900 MPa was
used to gelatinize the starch suspensions at ambient temperature. A
water-ethylene glycol (1 : 1 vol/vol) mixture was used as the pressuretransmitting fluid. The rate of pressurization was 400 MPa/min, with a
pressure release time of less than 20 s. During pressurization, the pressure, temperature, and time were kept constant using an automatic
device and recorded throughout the cycle using a data logger. Fifty
milliliters of 30% starch suspension was placed in a sterile, polyethylene bag and sealed under vacuum. The sample bags were placed inside
the Hydrostatic High Pressure Unit. HHP processing was performed
at 25 C for 15 min at a range of pressures from 100 MPa to 700 MPa.
The pressure-treated samples were analyzed using a DSC (model
2090; TA Instruments, New Castle, DE) to determine the degree of
gelatinization. Samples (5055 mg) of HHP-treated starch were placed
in a high-volume stainless steel crucible, and the thermograms were
recorded from 1 C to 100 C at a heating rate of 5 C/min. Each thermogram was analyzed to calculate the onset and the peak temperatures
and the enthalpy of the endothermic transition corresponding to the
melting of ungelatinized starch.
Results
The degree of gelatinization of a 30% starch solution is directly related
to the level of pressure applied at 25 C. Figure 14.3 shows that the
peak area corresponding to melting of ungelatinized starch decreases
progressively as the level of applied pressure is increased. In addition,
as the pressure level increases, the thermal stability of the ungelatinized starch phase decreases, indicating changes in the crystal structure
of starch, although some light microscopy studies show intact starch
granules after pressure treatment (Douzals et al. 1998; Stolt et al.
2001). However, some microscopic observations showed some swelling of starch granules (Douzals et al. 1996), which could lead to the
observation of reduced thermal stability.
The data in Figure 14.3 were analyzed further to calculate the percentage of gelatinized starch as a function of applied pressure (Figure
14.4). Percent starch gelatinization data from Douzals et al. (1996) and
Stute et al. (1996) for wheat starch also are included in Figure 14.4.
Although the initial starch concentrations are different, 16% (Douzals
0.30
600
500
0.32
400
0.34
300
200
0.36
100
0.38
0.40
50
Exo Up
55
60
65
70
75
80
Universal V2.6D TA Instruments
Temperature (C)
Figure 14.3. Gelatinization endotherm of 30% wheat starch suspension after HHP
processing at various pressure levels.
100
Percent gelatinization
80
60
40
20
0
0
100
200
300
400
500
600
Pressure (MPa)
346
347
et al. 1996), 25% (Stute et al. 1996), and 30% (w/w) dry solids (this
work), the trend seems to be similar. A fairly sharp increase in the
extent of gelatinization starts at 300 MPa for all starch data, and wheat
starch suspensions are completely gelatinized after a treatment at 500
600 MPa. The results suggest that starch can be partially or completely
gelatinized by increasing pressure at constant temperature.
Douzals et al. (2001) further characterized the behavior of the wheat
starch-water suspensions at 5% dry starch concentration over a pressure range of 0.1 and 600 MPa and over the temperature range of 20
and 96 C. The microscopic measurements of the loss of birefringence
of the granules calibrated by DSC was used to determine the extent of
gelatinization. The data were used to develop the pressure-temperature
(P-T) gelatinization diagram. The P-T gelatinization diagram is similar
to the P-T diagram of denaturation of proteins. The P-T diagram indicates that starch can be gelatinized under various pressure-temperature
combinations; however, the gelatinized starch can have different properties based on the gelatinization conditions.
For corn starch, Zuo et al. (1999) report the presence of native starch
after a treatment at 700 MPa for 2 min, but complete gelatinization after
5 min, which indicates that starch gelatinization by pressure is a kinetically controlled process. This issue becomes significant for investigation of starch kinetics at HHPs. The starch gelatinization kinetics
should be decoupled from the rate of increase of pressure with
time. In fact, Stolt et al. (2001) state that treatment of barley starch
suspension at 550 MPa with a zero-minute holding time exhibited
almost complete gelatinization, indicating pressurization at a rate of
20 MPamin1 was sufficiently slow for complete gelatinization.
348
Conclusions
An understanding of the effect of high pressure on the properties of
starch-based food systems under conditions relevant to food storage
are necessary to predict storage stability of such systems so that HHPprocessing protocols can be optimized for successful development of
HHP-processed commercial food products.
Studies on the retrogradation properties of starch in the presence of
common food ingredients are also essential for optimization of processing conditions so as to improve the physical stability and textural
characteristics of HHP-processed food products. Because DSC instruments operating under conditions relevant to HHP-processing conditions are not commercially available, starch samples have to be treated
before performing DSC on the samples. The thermodynamic basis of
starch gelatinization involves initial and final states that can be experimentally defined and energetic and/or structural differences that can
be measured using calorimetry. The comparison of various final states
as a function of exposure to various levels of pressure, starting from
the same initial state, makes it possible to predict the effectiveness of
HHP to produce gelatinized starch. However, development of DSC
equipment working at high pressures will allow one to perform pressure scans to determine the pressure dependence of the gelatinization
event, to investigate the kinetics of starch gelatinization at constant
pressure, and to separate the irreversible and reversible events.
349
References
Douzals J.P., Marechal P.A., Coquille J.C., and Gervais P. 1996. Microscopic study
of starch gelatinization under high hydrostatic pressure. J Agric Food Chem,
44(5):14051409.
Douzals J.P., Perrier-Cornet J.M., Gervais P., and Coqulle J.C. 1998. High-pressure
gelatinization of wheat starch and properties of pressure-induced gels. J Agric
Food Chem, 46:48244829.
Douzals J.P., Perrier-Cornet J.M., Coqulle J.C., and Gervais P. 2001. Pressuretemperature phase transition diagram for wheat starch. J Agric Food Chem,
49:873876.
Ezaki S. and Hayashi R. 1992. High-pressure effects on starch: Structural changes
and retrogradation. In: High Pressure and Biotechnology, Vol. 224, Balny C.,
Hayashi R., Heremans M.P., editors, pp. 163165. Colloque INSERM/John Libbey
Eurotext: Montrouge, France.
Famelart M.H., Chapron L., Piot M., Brule G., and Durier C. 1998. High pressureinduced gel formation of milk and whey concentrates. J Food Eng, 36:149164.
Fennema O.R. 1996. Food Chemistry, 3rd edition, pp. 201204. Marcel Dekker: New
York.
Jobling S. 2004. Improving starch for food and industrial application. Cur Opin Plant
Biol, 7:210218.
Jouppila K., Kansikas J., and Roos Y.H. 1998. Factors affecting crystallization and
crystallization kinetics in amorphous corn starch. Carbohyd Polym, 36:143149.
Messens W., Van Camp J., and Huyghebaret A. 1997. The use of high pressure to
modify the functionality of food proteins. Trends Food Sci Technol, 8:107112.
Roos Y.H. 1995. Phase Transitions in Food. Academic Press: San Diego, CA.
Stolt M., Oinonen S., and Autio K. 2001. Effect of high pressure on the physical
properties of barley starch. Innov Food Sci Emerg Technol, 1:167175.
Stute R., Heilbronn R.W., Klingler R.W., Boguslawski S., Eshtiaghi M.N., and Knorr
D. 1996. Effects of high pressures treatment on starches. Starch/Staeke,
48:399408.
Zuo C., Ma, C., and Zhang S. 1999. Effects of high hydrostatic on gelatinization of
corn starch. In: Proceedings of the International Conference on Agricultural
Engineering (ICAE), Vol. 4, pp. 9193 Beijing, China.
Chapter 15
Use of Calorimetry to Evaluate Safety
of Processing
Hans Fierz
Scope
Concepts
Severity: Adiabatic Temperature Rise
Probability: Time to Maximum Rate
Critical Conditions
Autocatalysis
Differential Scanning Calorimetry
Screening
Comparison of Open and Closed Measurement Methods
Estimation of q(T )
Isoconversional Methods
High-Sensitivity Calorimetry
Adiabatic Measuring Methods
Dewar Vessels
Accelerating Rate Calorimeter
Reactions with Oxygen
Screening Test
Determination of Self-Ignition Temperature
Applications
Formation of Hot Spots in Dryers
Storage and Hot Discharge
Prevention of Molasses Incidents
Transport Safety
Conclusion
References
351
352
352
353
353
354
356
356
356
357
357
360
361
361
361
362
363
363
364
364
364
364
365
365
366
366
352
Scope
Food, as does every other chemical, shows chemical reactivity, and if
handled in bulk can be dangerous. Numerous incidents involving
wheat, milk powder, coffee, or molasses are known, due to, for example,
self-heating, self-ignition of hot spots, or dust explosions.
This chapter focuses on the methodology for characterizing the
thermal consequences of exothermic decompositions in bulk and the
correspondent safety risks. Of course, there are desired synthetic exothermic reactions in bulk, with food presenting a thermal risk, as for
example the hydrogenation of fats. These cases are rather rare and may
not justify a treatment from a specialized point of view such as food
chemistry.
However, the methodology of how to treat and quantify the risk of
both desired reactions and decompositions is not specific to food chemistry, but was developed for process chemistry in general. A thorough
discussion of this topic can be found in books about process safety
(Stoessel 2008).
One of the major risks in food production is dust explosions in mills
or dryers, for example, in sugar refining. Because this chapter deals
with applications of calorimetry, we consider dust explosions only
insofar as hot spots serve as ignition sources. Further literature about
this topic can be found in Bartknecht (1981).
Concepts
The risk of an exothermic decomposition can be defined as the product
of its severity and its probability, both of which can be discussed within
the framework of the so-called adiabatic scenario.
Adiabatic means that there is no heat exchange at all between the
system and the surroundings. A situation with no or negligible heat
exchange could occur in various cases, including failure of heat transfer (breakdown of stirrer or cooling system) during storage of a liquid
or storage of reactive solids in bulk.
The latter situation can occur either intentionally, for example, when
a solid is stored in drums or containers at elevated temperature, or
unintentionally in a dryer or a mill when after a technical incident the
product is no longer agitated and the product cannot be discharged.
353
Qr
(15.1)
cp
Q
dT q
=
= r r
dt c p
cp
(15.2)
RT 2
Ea
cp
q (T )
(15.3)
354
Temperature [C]
60
50
t/8
40
t/4
30
t/2
20
t
10
0
0
10
20
30
40
50
60
70
Time [min]
The time until the system explodes is a function of the initial heat
release rate q and thus of the temperature (Figure 15.1). Note that in
the beginning, the temperature increase rate is only moderate. The
TMRad is therefore related to the probability of an incident. At very
long times, countermeasures may be taken, for example, discharging
the container or filling it with water. A very short time, however, does
not allow any action to be taken, and the thermal explosion cannot be
avoided.
Critical Conditions
Critical heat release rate
The adiabatic case is the worst-case assumption. Any real system loses
heat to its surroundings either by convection, in the case of liquids, or
by conduction, in solids. A steady temperature will be obtained when
the heat release rate of the reaction equals this heat loss rate, which is
the definition of the critical heat release rate qcrit. Any heat release rate
higher than this will lead to heat accumulation, to a temperature
increase, and finally to a thermal explosion.
355
RT 2
Ea
q(T )
(15.4)
Critical temperature
As the rate is related to the temperature via Arrhenius law, there is also
a critical temperature for a given bulk volume. Details can be found
in Gray and Lee (1967). The critical temperature, Tcrit, is therefore a
function of the layer thickness, r, and its physical properties (density
, thermal conductivity , geometry ), as well as of the macrokinetics
of the decomposition characterized by (q(T), Ea). Figure 15.2 shows
a typical dependency of the critical radius of the temperature.
Many methods originally developed by trial and error use, in fact,
the concept of critical temperatures (see below).
1.00E+00
1.00E-01
unstable, thermal explosion
1.00E-02
stable, no thermal explosion
1.00E-03
0
50
100
Temperature [C]
150
200
Figure 15.2. Dependence of the critical radius r of the temperature T for a typical
reaction as described in the text. Combinations of T and r in the overcritical region
will lead to a thermal explosion; in the undercritical region, the system will be stable.
356
Autocatalysis
The concepts described above rely on the assumption that the heat
release rate does not depend on the conversion. There are, however, many cases in which the heat release rate initially increases with
increasing conversion, reaches a maximum, and then decreases
again. This behavior is called formal autocatalysis or, in short,
autocatalysis.
For such cases, the equations previously discussed for the TMRad
and the critical layer thickness can be applied only with caution. It is
therefore important to identify this type of formal reaction. However,
using temperature-programmed thermoanalytical measurements for
this is not obvious and requires experience: Bou-Diab and Fierz (2002)
describe an identification approach.
Differential Scanning Calorimetry
Differential scanning calorimetry (DSC) is often used either as a
screening tool or to establish the thermal kinetics of a decomposition.
Here, the change in heat flow as function of the oven temperature is
recorded.
Becasue starting materials and products may be volatile, correct
results are obtained only by using closed and pressure-tight crucibles.
Measurements in which the samples are allowed to lose mass can show
quite different behavior from those made in closed crucibles. An
example describing this situation is given for saccharose in the
Comparison of Open and Closed Measurement Methods section below.
Typically, pressure-resistant gold-coated 40 l crucibles can be used,
which can withstand 400 C/220 bars.
Screening
From one temperature-programmed run, both the reaction and decomposition energy and its temperature range can be deduced (ASTM
E537-98).
As often is the case in calorimetry, the determination of the baseline
is difficult because, for reactions, the recorded signals usually cover a
broad temperature range. Interpolation of the baseline over such a
broad range and therefore determination of the decomposition potential
2.5
2
1.5
1
0.5
0
-0.5
-1
-1.5
120
Mass loss in %
100
80
TGA
SDTA
60
357
40
20
0
-20
DeltaT [K]
0
50
0
45
0
40
0
35
0
30
0
25
0
20
0
15
0
10
50
Temperature [C]
can be quite difficult. To cool down the sample after a run and to
perform a second scan without removing the sample crucible from the
sensor helps establish the baseline.
Comparison of Open and Closed Measurement Methods
Figure 15.3 shows a typical thermogravimetric trace of sugar (saccharose). Thermogravimetric analysis is based on mass loss of the sample;
the sample crucible is open to the atmosphere and thus the method is
called an open method. At 220 C, there is a sudden decrease in mass
followed by a gradual mass loss until, at 450 C, all the sample has
evaporated. At the same time, the instrument used records a qualitative
heat flow signal, which indicates a peak heat release rate at 350 C.
The same substance measured in DSC in a closed crucible shows a
quite high decomposition potential at lower temperatures (240 C)
(Figure 15.4).
Estimation of q(T)
The detection limit of a modern DSC apparatus is about 1 W/kg. If this
value is set equal to qcrit and used to calculate the critical radius rcrit in
Equation 15.4, a value of about 0.1 m results, which is equivalent to a
cubic container of approximately 8 L, depending on the assumptions
358
1
0.5
0
-0.5
-1
0
50
100
150
200
250
300
350
400
Temperature [C]
Figure 15.4. DSC measurement of saccharose at 4 K/min in a closed pressureresistant crucible. Note that the decomposition occurs at lower temperatures than in
Figure 15.3.
made. In other words, lower heat release rates that may lead to a
thermal explosion in a bigger volume may remain undetected.
Extrapolation of heat release rates to lower temperatures is thus
often necessary. A conservative estimate (that is an estimate giving
high values) for q at lower temperatures can be obtained as follows,
Once a baseline is drawn, the deviation of the signal from the baseline at any temperature T is proportional to the heat release rate q(T ).
If this is determined at the beginning of the signal, the influence of the
conversion can be neglected, and q(T ) can be used as reference value
(Figure 15.5). In practice, a value of 20 W/kg for q(T ) is a good compromise between a value too small to be measured and one so high
that there is already noticeable conversion.
To estimate heat release rates at other temperatures, a value for the
activation energy Ea is needed. This is generally not known a priori.
To overcome this problem, a very low value of the activation energy,
for example 50 kJ/mol, can be used, which in the case of real decompositions is practically never encountered.
This can now be used to calculate a too-short and thus conservative
value of the TMRad (Figure 15.6) or of the critical radius rcrit. If acceptable TMRad values or critical volumes are then obtained at the desired
temperature, no further action has to be taken. If not, more refined
kinetic methods are needed.
Thus, from one measurement both the severity Tad and the probability of a runaway TMRad can be estimated. This approach is in
practical use in many laboratories and has been verified theoretically
(Keller et al. 1997). Also, by comparing actual adiabatic experiments
to the results of the described method, it has been shown that the results
140
120
100
80
60
20 Watt/kg
40
20
0
0
50
100
150
200
250
300
Temperature [C]
Figure 15.5. Estimation of q(T) values from a DSC measurement. Shown is a schematic DSC thermogram and the determination of the heat release rate at small
conversion.
Figure 15.6. Influence of different activation energies on the safety margin of the
extrapolated TMRad. A measured heat release rate (15 W/kg at 160 C, upper right)
can be used to extrapolate the heat release rate corresponding to a TMRad of 24 h using
a high activation energy (lower curve) or a low activation energy (upper curve).
Extrapolated temperatures of 100 C and 50 C, respectively, result.
359
360
361
362
Dewar vessels are usually made of glass and are not pressureresistant. Gases and vapors must be allowed to escape from the system,
and because this is normally an endothermic process, it also will
contribute to the heat loss. The sensitivity of such an arrangement
can be improved (1) by minimizing the temperature difference between
Dewar and oven by adapting the temperature of the oven to that of
the Dewar and (2) by placing the Dewar vessel in an autoclave to suppress the evaporation of gases and vapors (Grewer 1994). The use of
containment is also advisable for reasons of safety and industrial
hygiene.
This improvement requires the use of rather sophisticated experimental facilities, and especially in the case of autoclaves, a room with
concrete walls and with remote control should be used.
meas
TMR ad
(15.5)
363
and
Tadcorr = Tadmeas
(15.6)
where
= 1+
mb c p,b
ms c p , s
(15.7)
364
Applications
Formation of Hot Spots in Dryers
Many solid food products are produced by spray-drying of water-based
solutions. Such a solution is sprayed into a stream of hot air. The water
is evaporated, and the solute is converted to a powder. Examples are
the production of powdered milk or coffee whitener.
Reactions with oxygen can lead either to product deterioration or to
smoldering if there are deposits. Deposits not only occur in the dryer
itself but also in equipment downstream, for example, in filters. As
discussed previously, if the heat release rate due to the oxidation in
such a product layer is overcritical, the formation of hot spots that can
trigger a dust explosion is unavoidable. In general, the self-ignition
temperature should therefore be determined.
Storage and Hot Discharge
Sometimes the product is discharged from a dryer to drums or containers while still hot. In a dryer, heat produced by an ongoing decomposition reaction can be removed by convection because the product is
agitated. In a container, however, this agitation is lacking, and heat is
removed by conduction only. This is a much less efficient mechanism
of heat transfer. The product can therefore self-heat, the self-ignition
temperature can be reached, and a fire can occur.
If the decomposition kinetics are known, it is possible to calculate
the critical temperature for a given container size. Alternatively, one
can establish a temperature limit using a series of isothermal Dewar
experiments by starting at a relatively high temperature where a signal
is observed and then lowering the temperature in steps of 10 K until
no signal is observed. From this temperature a safety margin, for
365
366
367
Keller, A., Stark, D., Fierz, H., Heinzle, E., and Hungerbhler, K. 1997. Estimation
of the time to maximum rate using dynamic DSC experiments. J Loss Prevent
Process Ind, 10:3141.
Opfermann, J. and Hdrich, W. 1995. Prediction of the thermal response of hazardous
materials during storage using an improved technique. Thermochim Acta,
263:2950.
Pastr, J., Wrsdrfer, U., Keller, A. and Hungerbhler, K. 2000. Comparison of
different methods for estimating TMRad from dynamic DSC measurements with
ADT 24 values obtained from adiabatic Dewar experiments. J Loss Prevent
Process Ind, 13:717.
Platje, T., Wittenberg, A., and Timmermans, A. 2006. Study of the runaway behaviour of technical sucrose solutions. Zuckerindustrie, 131(4):231238.
Raemy, A. and Lambelet, P. 1982. A calorimetric study of self heating in coffee and
chicory. J Food Technol, 17:451460.
Roduit, B. 2000. Computational aspects of kinetic analysis. Part E: The ICTAC
Kinetics Projectnumerical techniques and kinetics of solid state processes.
Thermochim Acta, 355:71.
Rogers, R.R. 1989. The advantages and limitations of adiabatic Dewar calorimetry
in chemical hazard testing. Plant Operations Progress, 8:109.
Stoessel, F. 2008. Thermal Safety of Chemical Processes: Risk Assessment and
Process Design. Wiley-VCH: Weinheim.
Suurkuus, J. and Wads, I. 1982. A multichannel microcalorimetry system. Chem
Scripta, 20:155163.
Townsend, D.I. and Tou, J.C. 1980. Thermal hazard evaluation by an accelerating
rate calorimeter. Thermochim Acta, 37:130.
UNECE Transport Division. 2003. International Recommendations on the Transport
of Dangerous Goods, Manual of Test and Criteria, 4th edition. UNECE: New
York.
Index
369
370
Index
nonthermal treatment of
bacteria in, 16163, 162f
hydrostatic pressure resistance of,
4445, 44f
inactivation of, 10
Lactobacillus plantarum DSC
analysis of, 15558, 157f
Listeria monocytogenes DSC
analysis of, 14953, 151f
antibiotics effect on, 15354,
153f
heat inactivation parameters of,
159
results for, 15253, 152f
sample preparations for,
14950
Mycoplasma laidlawii DSC
analysis of, 149
Staphylococcus aureus
nonthermal treatment,
16263, 162t
Batch high-pressure vessel, mixing
and reaction heat flux
microcalorimeter with, 29,
31t
Batch mixing vessel, high
sensitivity heat flux
calorimeter with, 27, 28t
Batch standard vessel, mixing and
reaction heat flux
microcalorimeter with, 29,
31t
Benzene, transitiometry verification
test using, 322, 323f
Binding
data quantifies high-affinity,
7577
mixing and reaction calorimetry
with, 4142
processes, protein in dilute
solution with, 78, 79, 80f
Blank test heat flow equation, 32
Index
Bovine -lactoglobulin, 9495
Bovine serum albumin (BSA),
protein denaturation affected
by, 104
Broad beans, 11S globulin from,
99100, 101f, 105
BSA. See Bovine serum albumin
Bulk phase system, denaturationaggregation of globular
proteins in, 12429,
126f128f
Butter, heat capacity for, 36t
Butyric acid, melting point of,
171f
Cabbage, heat capacity for, 36t
Calibration, 2325, 24f, 25f
Calvet type calorimeter, 23,
206
food-processing design, 206
heat calibration procedure for
HP-DSC, 60f, 6163, 62f
high pressure calorimetry, 314
HP-DSC, 5763, 60f, 62f
Joule effect in, 2324, 24f, 25f
MICROCALIX, 177
necessity for HP-DSC, 57
parameters for, 58
heating rate, 58
pan type, 58
sample mass, 58
temperature, 58
reference substances for, 58
indium, 58, 60, 61
lead, 58
tin, 58, 59
zinc, 58
scanning transitiometry, 32223
table of corrections for, 57
temperature calibration procedure
for HP-DSC, 5861
uncertainty with HP-DSC, 58
371
Calorimeter
accelerating rate, 36263
symmetrical, two-chamber,
1718
Calorimetry. See also Calvet type
calorimetry; Differential
scanning calorimetry; Heat
flux calorimeters; Heat flux
microcalorimetry; High
pressure calorimetry; High
pressure differential
scanning calorimetry;
High-sensitivity calorimetry;
High sensitivity heat flux
calorimeter; Isothermal
calorimetry; Isothermal
solution calorimetry;
Isothermal titration
calorimetry;
Microcalorimetry; Mixing
and reaction calorimetry;
Mixing and reaction heat
flux microcalorimeter;
Pressure calorimetry
advantages for using, 7
applications of, 1545
under controlled relative
humidity, 45
food dehydration understood
with, 289309
calorimetric glass transition
measurement for, 29396,
294f, 296f
dielectric and mechanical
relaxations with, 296f,
29798
freeze-drying for, 290, 3067
freezing in, 3013, 302f, 303f
glass transition and stability of,
3078, 308f
phase and state transitions of,
290, 29293, 293f
372
Index
Calorimetry (continued)
spray-drying for, 290, 3056,
306f
state diagrams with, 3037,
304f
thermal analysis in, 298301,
299f, 300f
food industry interest in, 226
food-processing design in, 2026
alternating DSC, 204
Calvet type calorimetry, 203,
206
differential scanning
calorimetry, 2026
differential thermal analysis,
203
dynamical mechanical analysis,
225
dynamical mechanical thermal
analysis, 225
methods, 2056
modulated DSC, 204
samples, 206
techniques, 2035, 204f
food-processing safety evaluated
with, 35166, 354f, 355f,
357f359f
adiabatic measurement
methods for, 36163
applications for, 36466
concepts for, 35256, 354f,
355f
critical conditions in, 35456,
355f
critical heat release rate in,
35455
critical temperature, 35556,
355f
estimation of q(T) in, 35760,
359f
formation of hot spots in
dryers, 364
Index
C80 technique v. vessel/heating
mode with, 31t
cereal with, 12
endothermic/exothermic effects
of, 19t
gelatinization of starch-water
systems, 207
glass transition with, 294
hydrophilic component of,
291
thermal analysis of cereal
nonstarch, 27678, 277f
thermal behavior of food
constituents in, 2068, 207f,
208f
Carboxymethylcellulose,
denaturation temperature of
11S globulin with, 106t
Carp, heat capacity for, 36t
-Carrageenan
denaturation temperature of 11S
globulin with, 106t
denaturation temperature of
-lactoglobulin with, 110
-Carrageenan
denaturation temperature of 11S
globulin with, 106t
denaturation temperature of
-lactoglobulin with, 110
Carrots, isothermal traces at
temperatures for, 39f
Caseins, 12223
CB. See Cocoa butter
Cellobiose, calorimetric curves of,
208f
Centre National de la Recherche
Scientifique (CNRS), 176
Cereal, 12
C80 technique v. vessel/heating
mode with, 31t
Cereal processing, thermal analysis
to design/monitor, 26585
373
374
Index
Crystallization
DSC technique v. vessel/heating
mode with, 28t
heating mode with, 36
isothermal, 39
lard, 190, 191f, 191t
lipids, 21011
milk fat, 18990
oil-in-water emulsions, 13236,
135f, 136t
water in pork muscle with, 327f
CSC, 21
Dairy, heat capacity for, 36t
Debye-Hckel approximation, 98,
99t
Dehydration. See Food dehydration
Denaturation
cold, protein in dilute solution
with, 75
defined, 122
DSC technique v. vessel/heating
mode with, 28t
globular proteins in bulk phase
system, 12429, 126f128f
11S globulin, 8992, 91f
alcohols effect on, 99100,
101f
different pH values in, 91f, 92
polysaccharides effect on,
1056, 106t, 110
salts effect on, 9698, 97f, 99t
two-state model to analyze, 89
heat effects of, 88
heating mode with, 3637, 37f
Kunitz inhibitor, polysaccharides
effect on, 10710, 108f
-lactoglobulin, 125
methodological approaches to
study, 89
of protein, 87113, 91f, 97f, 99t,
101f, 103f, 106t, 108f
Index
hydrogen ion buffer selection
in, 77
purity in, 77
two-state, reversible transitions
in, 7677
efficiency ratio of flat-shaped,
21f
Escherichia coli analysis by, 148,
15558, 157f
erythromycin treatment of,
15455, 155f
heat inactivation parameters of,
15960, 160f
nonthermal treatment of,
16263, 162t
foodborne bacteria analysis by,
14764, 151f, 153f, 155f,
157f, 160f, 162t
antibiotics effect on, 15355,
153f, 155f
cold shocking, 148
heat shocking, 148, 151
food-processing design in, 2026
methods, 2056
samples, 206
techniques, 2035, 204f
with XRD, 225
food-processing safety with, 355f,
35661, 357f359f
estimation of q(T) for,
35760, 359f
high-sensitivity calorimetry for,
361
isoconversional methods for,
36061
open v. closed measurement
methods for, 357, 357f, 358f
screening for, 35657
food-processing treatment
evaluation by, 15864, 160f,
162f
antimicrobials in, 16364
375
heat inactivation parameters of
bacteria in, 15861, 160f,
162f
HHP in, 161
nonthermal treatment of
bacteria in, 16163, 162f
glass transition with, 294f
heat flux type of, 2021, 2630,
27f, 28t
heatings role in, 88
high pressure, 5164, 54f56f,
60f, 62f
applications of, 63
calibration of, 5763, 60f, 62f
construction of, 5357,
54f56f
Lactobacillus plantarum analysis
by, 15558, 157f
Listeria monocytogenes analysis
by, 14953, 151f
antibiotics effect on, 15354,
153f
heat inactivation parameters of,
159
results for, 15253, 152f
sample preparations for,
14950
microcalorimetry v., 16, 1925,
20f25f
heat flux microcalorimetry,
1925, 20f25f
Mycoplasma laidlawii analysis
by, 149
power compensated type of,
2022
protein in dilute solution with,
6877, 71f
equations, 7074, 71f
heat capacity change origins
for, 74
information content, 6869
instrumentation, 6970
376
Index
Index
oil-in-water, 13241, 135f, 136t,
137f, 139f, 139t, 140f
anhydrous milk fat, 13341,
135f, 136t, 137f, 139f, 139t,
140f, 18487, 186f, 187f
Avrami equation for, 13839,
139f, 139t
crystallization in, 13236,
135f, 136t
fat crystal growth in, 138
Gompertz model for, 13940,
139t, 140f
ice cream, 133
kinetics of, 13641, 137f, 139f,
139t, 140f
melting of fat droplets in,
13236, 135f, 136t
triacylglycerols, 133
whipped cream, 133
proteins role in, 10
Enthalpy, 265
activation, 129
estimate of apparent denaturation,
112
reaction, 255
vant Hoff enthalpy change,
7274
Entropy, protein heat-induced
transformations with, 129
Enzymatic reactions, mixing and
reaction calorimetry with,
42, 42f, 43f
Enzyme
C80 technique v. vessel/heating
mode with, 31t
DSC technique v. vessel/heating
mode with, 28t
endothermic/exothermic effects
of, 19t
Erythromycin, Escherichia coli
treatment with, 15455, 155f
Escherichia coli
377
378
Index
Index
open v. closed measurement
methods with, 357, 357f,
358f
screening with, 35657
reactions with oxygen in, 36364
determination of self-ignition
temperature for, 364
screening test for, 363
Formal autocatalysis, 356
Fourier transform infrared
spectroscopy, 11
Free protein, denaturation
temperature of 11S globulin
with, 106t
Freeze-drying, 290, 3067
Fruit, heat capacity for, 36t
Galactose, calorimetric curves of,
208f
Gas-flow vessel, mixing and
reaction heat flux
microcalorimeter with, 29,
31t
Gelatin, endothermic/exothermic
effects of, 19t
Gelatin gels, frozen water ratio in,
32629, 328f
Gelatinization
DSC technique v. vessel/heating
mode with, 28t
heating mode with, 38
starch, 12, 274, 27879, 280f
calorimetric analysis by HPP
of, 34149, 343f, 346f
heat in, 342, 343f
high pressure calorimetry on,
33036, 332f334f, 335t
storage of, 34748
thermodynamic data for, 335t
wheat, 34447, 346f
starch-water systems, 207
Gelatin molecules, 122
379
380
Index
Index
triacylglycerols, 133
whipped cream, 133
peak temperatures and heat of
reaction in, 128f
protein solutions with, 11932,
126f128f, 130f, 131t, 132f,
141
activation enthalpy of, 129
denaturation-aggregation of
globular proteins in, 12429,
126f128f
entropy of, 129
kinetics of, 12932, 130f, 131t,
132f
Lumry-Eyring model for, 129,
130f, 131t
protein structures in, 12123
thermodynamics of, 12324,
12932, 130f, 131t, 132f
whey protein isolate in, 125,
126f, 128f, 132f
Heating mode, 3540, 37f, 39f, 40f
aggregation, 3637, 37f
crystallization, 36
denaturation, 3637, 37f
gelatinization, 38
gelation, 3738
isothermal calorimetry, 38
isothermal crystallization, 39
oxidative stability, 38
retrogradation, 28t, 38
scanning calorimetry, 3536
shelf life, 38, 39f
step heating in calorimetry, 40,
40f
Heat of combustion, parameters for
food-processing design of,
225
Heat of solution, parameters for
food-processing design of,
21824, 221f223f
Heat release rate, critical, 35455
381
382
Index
Index
calculation for reaction order,
25253
calculation for total heat
released, 25354
calculation of initial
calorimetric signal 0, 252
calculation of QT, 25660, 260f
determination of K, 25556
empirical model fitting for,
24649, 246f, 249f, 250f
qualitative studies on, 23945,
241f
quantitative studies on, 245
reaction kinetics based model
of, 24951
reactions that proceed to
completion for, 25255
reactions that proceed to
equilibrium for, 25560,
260f
test for complete reaction, 255
Isothermal crystallization, heating
mode with, 39
Isothermal solution calorimetry, 220
Isothermal titration calorimetry
(ITC), 910
dependence of model with, 82
heat of interaction measured with,
84
protein in dilute solution with,
68, 7784, 80f, 83f
binding processes in, 78, 79,
80f
data analysis for, 7982
information content, 7778
instrumentation, 7879, 80f
power compensation design in,
78
range of applicability with,
8283, 83f
shape of titration curve with,
8283, 83f
383
384
Index
Lauric acid
melting point of, 171f
MICROCALIX calibration with,
177
Lead, HP-DSC calibration using, 58
Linoleic acid, melting point of, 171f
Linolenic acid, melting point of,
171f, 172f
Lipids, 169. See also
Triacylglycerols
antioxidant efficacy of, 212
crystallization kinetics of, 21011
emulsifier-water systems with,
21214
emulsions, 214
melting profile of, 209
oxidative stability of, 21112
polymorphism of, 20910, 210f
quality control for, 211
thermal behavior of food
constituents in, 20814,
210f, 213f
Liquids, heat capacity determination
for, 3435, 34f
Listeria monocytogenes
antibiotics effect on, 15354,
153f
DSC analysis of, 14953, 151f
heat inactivation parameters of,
159
results for, 15253, 152f
sample preparations for, 14950
Lumry-Eyring model, 104
protein heat-induced
transformations with, 129,
130f, 131t
Lyophilization, DSC technique v.
vessel/heating mode with,
28t
Maltodextrin (MD), moisture
content of, 22224, 223f
Index
mixing and reaction
calorimetry, 4043, 41f43f
pressure calorimetry, 4345,
44f
Milk
bovine -lactoglobulin of, 9495
DSC technique v. vessel/heating
mode with, 28t
endothermic/exothermic effects
of, 19t
heat capacity for, 36t
isothermal traces at temperatures
for, 39f
protein, 9495, 131
skim milk powder, 22224, 223f
Milk fat
anhydrous, 13341, 135f, 136t,
137f, 139f, 139t, 140f,
18487, 186f, 187f
crystallization properties of,
18990
DSC and XRD with, 18490,
186f, 187f, 189f
globules, 18889, 189f
Mixing and reaction calorimetry,
4043, 41f43f
batch mixing in, 40
binding, 4142
dissolution, 41
enzymatic reactions, 42, 42f, 43f
fermentation, 43
flow mixing in, 41
neutralization, 41, 41f
solubility, 41
Mixing and reaction heat flux
microcalorimeter, 2930, 31t
ampoule mixing vessel for, 30,
31t
batch high-pressure vessel for,
29, 31t
batch standard vessel for, 29, 31t
gas-flow vessel for, 29, 31t
385
386
Index
Oersted law, 25
Oil. See also Lipids
C80 technique v. vessel/heating
mode with, 31t
DSC technique v. vessel/heating
mode with, 28t
endothermic/exothermic effects
of, 19t
oxidative stability of, 21112
Oil-in-water emulsions, 13241,
135f, 136t, 137f, 139f, 139t,
140f
anhydrous milk fat, 13341, 135f,
136t, 137f, 139f, 139t, 140f,
18487, 186f, 187f
Avrami equation for, 13839,
139f, 139t
crystallization in, 13236, 135f,
136t
fat crystal growth in, 138
Gompertz model for, 13940,
139t, 140f
ice cream, 133
kinetics of, 13641, 137f, 139f,
139t, 140f
melting of fat droplets in,
13236, 135f, 136t
triacylglycerols, 133
whipped cream, 133
Oleic acid, melting point of, 171f
One-cell calorimetric principle,
18f
Open measurement method, 357,
357f, 358f
Orange juice, heat capacity for,
36t
Ovalbumin, protein denaturation
effected by, 1024, 103f
Overlapping peaks, interpretation
of, 8
Oxidative stability, heating mode
with, 38
Index
thawing heat flux of, 326f
water in, 32426, 326f, 327f
Postdenaturation aggregation
aggregation rate determined by
denaturation rate with, 111
estimate of apparent denaturation
enthalpy with, 112
irreversible, 110
kinetic parameters of, 111
of protein, 11012
reversible, 110
Potato, heat capacity for, 36t
Power compensation principle, 53
Pressure calorimetry, 4345, 44f
Pressure shift freezing (PSF)
basic procedure of, 329f
heat flux during, 330f
high pressure calorimetry with,
32930, 329f331f
ice crystal/sample mass ratio
formed during, 331f
pressure during, 330f
temperature during, 330f
Propylene glycol, denaturation
temperature of
-lactoglobulin with, 110
Protein
20 amino acids constituting, 120
behavior upon heating of, 89
bovine -lactoglobulin of milk,
9495
calorimetry of dilute solution of,
6784, 71f, 80f, 83f
cold denaturation with, 75
DSC data quantifies highaffinity binding with, 7577
DSC for, 6877, 71f
ITC for, 68, 7784, 80f, 83f
cereal with, 12
conformation stability of, 121
DSC technique v. vessel/heating
mode with, 28t
387
emulsions/foams, role in, 10
free, 106t
heat-induced transformations in
solutions of, 11932,
126f128f, 130f, 131t, 132f,
141
denaturation-aggregation of
globular proteins with,
12429, 126f128f
kinetics of, 12932, 130f, 131t,
132f
protein structures with, 12123
thermodynamics of, 12324,
12932, 130f, 131t, 132f
hydrophilic component of, 291
milk, 9495, 131
polysaccharides thermodynamic
incompatibility with, 10910
postdenaturation aggregation of,
11012
structures, 12123
caseins, 12223
gelatin molecules, 122
polypeptide chains, 121
secondary structures, 121
tertiary structures, 12122
thermal analysis of cereal
processing with, 27276,
274f, 275f
gluten fix of water molecules
for, 273
soluble in aqueous media for,
273
starch gelatinization with, 274
thermal behavior of food
constituents in, 21416,
215f, 216f
thermal denaturation of, 87113,
91f, 97f, 99t, 101f, 103f,
106t, 108f
effects of alcohols on, 99100,
101f
388
Index
Protein (continued)
effects of odorants on, 1024,
103f
effects of pH on, 8995, 91f
effects of polysaccharides on,
10410, 106t, 108f
effects of salts on, 9599, 97f,
99t
reversibility of, 123
two-state model of, 123
thermodynamic compatibility of
denatured/native, 89
ultrasensitive calorimetry to,
8
Protein-protein interactions
(Exothermic reaction),
12728
PSF. See Pressure shift freezing
QT, calculation of, 25660, 260f
Rate constant, calculation for,
25455
RBPC. See Ribulose 1,5
biphosphate carboxylase
Reaction enthalpy, calculation for,
255
Reaction half-life, calculation for,
254
Reaction kinetics, model of based
on, 24951
Reaction order, calculation for,
25253
Retrogradation
DSC technique v. vessel/heating
mode with, 28t
heating mode with, 28t, 38
starch, 270
Ribulose 1,5 biphosphate
carboxylase (RBPC),
different pH values in
denaturation of, 93
Index
precautions for, 32122
pressure detector in, 320
schematic diagram of, 319f
scheme of basic principles of,
317f
temperature and energy scales of,
322
thermodynamic scheme of, 318f
transitiometric vessels for, 321f
Self-ignition temperature, 364
Setschenow equation, 100
SFC. See Solid fat content
Shelf life, 23761
empirical model fitting for
analysis of, 24649, 246f,
249f, 250f
heating mode with, 38, 39f
qualitative studies on, 23945,
241f
quantitative studies on, 245
reaction kinetics based model of,
24951
reactions that proceed to
completion for analysis of,
25255
calculation for rate constant,
25455
calculation for reaction
enthalpy, 255
calculation for reaction halflife, 254
calculation for reaction order,
25253
calculation for total heat
released, 25354
calculation of initial
calorimetric signal 0, 252
reactions that proceed to
equilibrium in analysis of,
25560, 260f
calculation of QT, 25660, 260f
determination of K, 25556
389
390
Index
Starch (continued)
high pressure calorimetry on,
33036, 332f334f, 335t
storage of, 34748
thermodynamic data for, 335t
wheat, 34447, 346f
HHPs effects on, 12
retrogradation, 270
thermal analysis of cereal
processing with, 26872,
289f272f
aqueous suspension of starch
granules for, 268
DSC for, 26872, 289f272f
State diagrams
food dehydration in, 3037, 304f
freeze-drying, 290, 3067
lactose, 304f
spray-drying, 290, 3056, 306f
State transitions, food dehydration
in, 290, 29293, 293f
Stearic acid, melting point of, 171f,
172f
Step heating, calorimetry with, 40,
40f
Sucrose, calorimetric curves of,
207f, 208f
Sugar
C80 technique v. vessel/heating
mode with, 31t
calorimetric curves of sucrose,
207f, 208f
DSC technique v. vessel/heating
mode with, 28t
glass transition with, 294
thermal behavior of food
constituents in, 2068, 207f,
208f
SXRD. See Small-angle X-ray
diffraction
TA. See Thermal analysis
Index
TMDSC. See Temperature
modulated DSC
Total heat released, calculation for,
25354
Transitiometry scanning technique,
12
Transition state theory, 111
Transport safety, 36566
Triacylglycerols (TG)
composition of, 169
DSC and XRD in study of,
16976, 171t, 172t, 173f
fatty acids, 17073, 171t, 172t,
173f
crystallographic/energetic
properties of, 172f
hexagonal, 172, 172f, 173f
melting point of, 171f, 172f
orthorhombic perpendicular,
172, 172f, 173f
triclinic parallel, 172, 172f,
173f
heat-induced transformations
with, 133
main types of, 173f
melting profile of, 209
polymorphism of, 17073, 171t,
172t, 173f, 20910, 210f
Trypsin inhibitor, 94
Vanillin, 102
Vant Hoff enthalpy change, DSC
measured, 7274
Vegetable, heat capacity for, 36t
Water. See also Oil-in-water
emulsions
emulsifier-water systems with
lipids, 21214
frozen water ratio in gelatin gels,
32629, 328f
gluten fix with molecules of, 273
391
392
Index
Yeast
C80 technique v. vessel/heating
mode with, 31t
DSC technique v. vessel/heating
mode with, 28t
endothermic/exothermic effects
of, 19t
Yogurt processing, DSC technique
v. vessel/heating mode with,
28t
Youngs modulus E, 265
Zinc, HP-DSC calibration using,
58