Thermal behaviour of major food constituents

In food many physicochemical effects can be observed in the temperature range between -50°C and 300°C. These thermal phenomena may be either endother-mic, such as melting, gelatinisation, denaturation, evaporation, or exothermic such as crystallization, oxidation, fermentation. Glass transitions are observed as a shift in the baseline; this information is associated with water content and water activity determinations.

Specific heats (cp) of foods can be calculated (Gekas, 1992) on the basis of the specific heats of the main constituents (by proportional additivity of the respective masses), but they can also be determined by DSC. The basic principles of such measurements have been described (Raemy and Lambelet, 1982) and many values of food specific heats are given in the literature (Mohsenin, 1980).


The main phenomena observed with carbohydrates are release of crystallization water, melting, decompo sition, gelatinisation of starch in the presence of water, retrogradation of the gel as well as glass transition, relaxation and crystallization of amorphous samples (Raemy and Schweizer, 1983; Raemy et al., 1993; Roos, 1995; Blanshard and Lillford, 1993). Tables with melting and decomposition temperatures as well as enthalpies are given in the literature (Raemy and Schweizer, 1983).

Figure 1 shows a typical calorimetric curve of amorphous sucrose with glass transition and relaxation at 50°C, crystallization at 90°C as well as melting above 150°C. Figure 2 presents calorimetric curves of amorphous sucrose at increasing water activities: Tg and crystallization temperature diminish with increasing water activities.

- Exo ^



65 J/g



65 J/g S §■

\ A relaxation

\ / 3

■ \J

Fig. 1 Typical calorimetric heating curve of (freeze-dried) amorphous sucrose showing glass transition and relaxation at 50°C, crystallization at 90°C and melting above 150°C (Setaram DSC 111, 2°C/min). From Raemy et al., 1993

I etnperature ''

Fig. 1 Typical calorimetric heating curve of (freeze-dried) amorphous sucrose showing glass transition and relaxation at 50°C, crystallization at 90°C and melting above 150°C (Setaram DSC 111, 2°C/min). From Raemy et al., 1993

Glass transition indicates that amorphous carbohydrates change from the glassy state to the rubbery state during heating. Glass transition and relaxation are often superimposed phenomena. Glass transition is a reversible phenomenon observed in DSC experiments as a change in baseline level, whereas relaxation is a non-reversible endothermic transition. Thus, performing two consecutive DSC experiments on the same sample can distinguish between these two phenomena (with MDSC only one scan is necessary as it separates reversible from non-reversible transitions). Glass transition is of particular interest in relation to storage of frozen products and food powders, and also for gas retention in powders foreseen to foam when dissolved. The gas, normally nitrogen, is for example introduced into an amorphous disaccharide matrix at a temperature above Tg, i.e. when the disaccharide are in a rubbery state, and encapsul ated bel ow Tg when the disaccharide are in the glassy state (Vuataz, 2002; Schoonman et al., 2002). Glass transitions are more difficult to observe by DSC in starch than in mono- or disaccharides, but accurate Tg values as a function of water content can be found in the literature (Zeleznak and Hoseney, 1987; Parker and Ring, 2001).

20 40 60 so 100

Temperature ["C]

Fig. 2 Calorimetric heating curves of amorphous sucrose at different water activities (Micro-DSC III, 1°C/min) showing the effect of increasing water activities on glass transition + relaxation and on crystallization. From Raemy et al., 1993

20 40 60 so 100

Temperature ["C]

Fig. 2 Calorimetric heating curves of amorphous sucrose at different water activities (Micro-DSC III, 1°C/min) showing the effect of increasing water activities on glass transition + relaxation and on crystallization. From Raemy et al., 1993

Thermogravimetry can be very useful for studying release of crystallization water, by indicating which endothermic transition observed by calorimetry corresponds to a mass loss.

Gelatinisation of starch-water systems is an endothermic non-reversible phenomenon easily observed by DSC (Ghani et al., 1999). Retrogradation, which is a slow and low energy recrystallization process, can be followed by isothermal microcalorimetry (Raemy et al., 1990; Silverio et al., 1996), but is more often characterized after a storage period by measuring the melting transition of the retrogradated gels (Karim et al., 2000). Penetrometry, DMA or DMTA provide complementary information on gelatinisation and retrogradation, which are associated with rheological modification of the products (Roulet et al, 1988).


Calorimetry (DTA, DSC) has been a method of choice to characterize the thermal properties of triacylglycerols (TAGs) for over 50 years, in particular their polymorphic behaviour.

Overall we can define three types of applications for lipids: determination of the thermodynamic parameters of the liquid/solid phases, monitoring of crystallization kinetics, determination of lipid quality and oxidative stability.

Thermodynamic parameters

The most common use of DSC in the lipid field has been to identify the polymorphism of TAGs and fat (Arishima et al., 1991; Dimick and Manning, 1987; Garti and Sato, 1988; Huyghebaert and Hendrickx, 1971; Loisel et al., 1998; Lovegren et al., 1976; Merken and Vaeck, 1980; Minato et al., 1997; Rousset, 1997; Rousset and Rappaz, 1996; Sato et al., 1989; Sato, 1996; Spigno et al., 2001; Wille and Lutton, 1966). This is done by measuring the melting enthalpy (between 50 and 200 J/g) and the melting temperature (pure components) or temperature range (complex mixiures like fat) of the phases present. As the polymorphism ofTAGs is monotropic, only one form is thermodynamically stable. Thus, observing all polymorphs is not always easy (see next paragraph). A solution is to crystallize the fat using a wide range of cooling rates (e.g., between 0.5°C/min and 50°C/min). Figure 3 presents the melting curves of the five most stable phases of cocoa butter; the least stable form I could not be observed due to the limited cooling capacity of the apparatus used. Changes in melting temperature and enthalpy have also been correlated to fat composition (Chaiseri and Dimick, 1989; Tan et al, 2000).

Fig. 3 DSC heating curves of 5 polymorphs of cocoa butter: II(a), III(ß'2), IV(ß'i), V(ß2) and VI(ßi) (Mettler FP900, 5°C/min). From Rousset, 1997

For binary or ternary mixtures of TAGs or fats, DSC has been used to determine phase diagrams or iso-solid diagrams, by identifying the temperature stability domains of the various phases formed (Ali and Dimick, 1994a; Ali and Dimick, 1994b; Culot, 1994; Elisabettini et al., 1998; Knoester, 1972; Koyano et al, 1992; Loisel et al, 1998; Muhammad and Dimick, 1994; Rousset et al., 1998; Timms, 1980; Timms, 1994). As previously mentioned, DSC is often used in combination with XRD for unambiguous identification.

The solid fat content (SFC) curve represents the ratio of solid to liquid in a partially crystallized lipid as a function of temperature. SFC curves are currently used in the industry for fat selection and quality control. They can be obtained from the calorimetric melting curve by sequential peak integration (Lambelet et al., 1986; Kaiserberger, 1989; Bhaskar, 1998). This determination requires precise knowledge of the melting enthalpy of each phase for each fraction present in the sample, which is very difficult to assess for most fats.

As already mentioned, specific heat is another parameter that can be determined by DSC. It has been measured for various TAGs and fats (Roberts and Pearce, 1983; Rousset, 1997).

Kinetic parameters

A second domain where DSC is useful is the study of the crystallization kinetics of TAGs and fats, and of the formation and stability of their various polymorphs as a function of time and temperature (Rousset et al, 1996; Rousset et al., 1997). For these experiments, 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. Crys talli zation is then achieved either isothermally after quenching at the desired temperature or at constant cooling rate. Kinetic information has been obtained by achieving measurements either isothermally (Koyano et al., 1989; Koyano et al., 1991; Metin and Hartel, 1998; Rousset, 1997; Ziegleder, 1985b; Ziegleder, 1990), or under various cooling conditions (Cebula and Smith, 1991; Kawamura, 1980). Complex thermal paths like tempering stages were also studied by calorimetry to understand precisely the mechanisms that induce the formation of stable crystalline forms (Rousset and Rappaz, 1997).

Precise kinetic parameters can be determined from isothermal crystallization experiments. The variation of SFC as a function of time is obtained by sequential integration of the crystallization peak. This SFC function is used to estimate parameters of crystallization with the help of the Avrami model or more complex ones (Foubert et al., 2002; Kloek et al., 2000; Rousset, 2002). Nucleation induction times can also be determined from isothermal crystallization experiments. This is the time needed before nucleation can appear and is a useful indicator of the nucleation rate, being inversely proportional to it (Rousset and Rappaz, 2001). These kinetic parameters are necessary for the modelling and prediction of crystallization with analytical or numerical models (Rousset, 2002). These models are tools to know how to crystallize lipids in the desired form.

In kinetic studies, DSC signal assignment may be ambiguous and need to be combined with XRD (Keller et al., 1996) and even synchrotron XRD if transformations are rapid in regard to the acquisition time. Even better, new experiments simultaneously combining DSC and synchrotron XRD revolutionise the study of crystallization (Kalnin et al., 2002; Lopez et al, 2000; Lopez et al, 2001a; Lopez et al., 2002; Ollivon et al, 2001). DSC can also be used simultaneously with microscopy to identify morphologies associated with polymorphs and cooling conditions (Rousset et al., 1998; Rousset and Rappaz, 1996).

Kinetic studies also help to understand the effect of compositional changes (TAGs or minor components) on crystallization (Cebula and Smith, 1992; Garti et al., 1988; Tan et al., 2000; Vanhoutte et al., 2002b; Vanhoutte et al., 2002a; Wahnelt et al., 1991).

However, as the samples are not agitated, results from DSC crystallization studies are often difficult to interpret in terms of process operating conditions (Rousset and Rappaz, 2001; Ziegleder, 1985a; Ziegleder, 1988b).

Quality control

DSC crystallization curves have been used to assess the quality of oils, in particular of heated oils (Gloria and Aguilera, 1998; Tan and Man, 1999). Similarly, contamination (adulteration) of fats can be detected by calorimetry either during crystallization or melting of lipid mixtures (Kowalski, 1989; Lambelet and Ganguli, 1983; Bringer et al., 1991; Marikkar et al., 2002).

Lipid oxidation is an exothermic phenomenon that can be followed, at least at elevated temperatures, by DSC or preferably by isothermal calorimetry (Tan and Man, 2002; Raemy et al., 1987; Kowalski, 1989). Measurements can be performed under a static air atmosphere or, better, under oxygen flow or oxygen pressure. DSC isothermal experiments allow induction times of oxidation to be determined. Tables of oxidation induction times measured by isothermal heat flux calorimetry around 100°C are reported in the literature (Raemy et al., 1987). For edible oils DSC induction times 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) or the efficiency of food antioxidants (Raemy et al., 1987; Irwandi et al., 2000; Tan and Man, 2002) in bulk lipids.


The main phenomena observed by DSC during heating of protein solutions (egg white or dairy proteins such as ß-lactoglobulin or a-lactalbumin) are endother-mic transitions commonly called denaturation (Privalov and Khechinashvili, 1974; Harwalkar and Ma, 1990; Ferreira et al., 1997). As an example Fig. 4 presents a Micro-DSC curve of lyzozyme-depleted egg white showing thermal denaturation.

These transitions, although often seen as a single peak in DSC experiments, are composed of data from changes in conformational state of the proteins (unfolding, denaturation) and subsequent aggregation. Whereas the native-to-denatured change in the protein state is a co-operative phenomenon that is accompanied by significant heat uptake, change in hydrophobic interactions during protein aggregation is an exothermic process (Privalov and Khechinashvili, 1974; Biliaderis, 1983; Hayakawa et al., 1996; Gotham et al, 1992). Although very rare, an exothermic component due to protein aggregation was observed following endothermic unfolding denaturation (Rossi and Schiraldi, 1992). The endo-


A.____j_ JX4-C




1 \ 1 _ -—

91.3'C --


i i



Denaiuration Enlhaipy: - 1.71 J/g

■ i i

i t i ■ * i i i i i i i

i i i



i i i i i i nt: Micro DSC Ifl Se(aram !

Sample w«tftbt: 0,4071 g t i i i

4 1


Temperature [*C]

Temperature [*C]

Fig. 4 Micro-DSC heating curve (1°C/min) of lyzozyme-depleted egg white (first run minus second run). From Ferreira et al., 1997

thermic nature of DSC curves recorded during thermal treatments of protein solutions are an indication of the large contribution of denaturation as compared to aggregation. In fact, reported values of enthalpy changes during aggregation of proteins are small, for example 1-5 J/g for aggregation of whey proteins induced by CaCl2 or proteolysis (Ju et al., 1999). Though a significant underlying exothermic contribution of protein aggregation cannot always be ruled out, especially at high protein concentrations, the thermal effect due to aggregation is generally of such small amplitude in relation to the endotherm produced by de-naturation that it is ignored (Donovan and Ross, 1973). In this sense, the temperature of the endothermic transition appearing in DSC analysis of protein solutions is indicative of thermal stability of the protein, and the surface of the peak corresponds to the denaturation enthalpy. For example the influence of hydration on the denaturation temperatures and enthalpies of lyzozyme has been given in the literature (Gregory, 1995). In the same way, the amount of protein that has been denaturated, e.g. during a technological process, can be determined by comparing the surface of denaturation transition to the total denaturation enthalpy (Arntfield and Murray, 1981).

Thermal denaturation of some proteins, e.g. egg (Ferreira et al., 1997) or muscle proteins (Wright and Wilding, 1984; Togashi et al., 2002) is a multi-step process. Thus, thermal denaturation of rabbit (Wright and Wilding, 1984) and walleye pollack (Togashi et al, 2002) myosins was shown to occur via three co operative endothermic processes associated with unfolding of different domains in the molecule.

DSC has also been used to study mixed protein systems. For example, the presence of gluten has been shown to shift the apparent transition temperature of whey protein towards lower temperatures; this was interpreted as gluten modification of the thermal stability of the environment of whey proteins (Lupano, 2000).

Glass transition and oxidation are primarily observed with dry proteins. As heat exchange associated with the glass transition of proteins is small, this transition is rather detected based on changes in rheological parameters obtained, for example by DMTA (Pouplin et al., 1999). However, the sharp decrease in mechanical properties occurring when the sample passes through the glass transition zone depends on the frequency of the forces applied to the sample in the DMTA experiment. Thus, values of Tg obiained from DMTA are not always consistent with those obtained from DSC (Hartel, 2001).


Thermal analysis and calorimetry allow mainly the observation of crystallization (undercooling), melting (of ice) and vaporisation. Since the enthalpies corresponding to these phenomena are quite high (333 J/g for ice melting and 2255 J/g for water vaporisation) they can easily be studied by standard DSC in samples with low water content. It must here be remembered that under undercooling conditions crystallization enthalpies diminish with decreasing temperature (Franks, 1982).


In many foods (beer, ice cream, etc.) air is an important constituent (if we consider volume and not mass). But due to its low density compared to the other solid or liquid constituents, air does not change the thermal properties of foods, with exception of thermal conductivity as air is a good insulator.

Was this article helpful?

0 0
Metabolism Masterclass

Metabolism Masterclass

Are You Sick And Tired Of All The Fat-Burning Tricks And Trends That Just Don’t Deliver? Well, Get Set To Discover The Easy, Safe, Fast, And Permanent Way To Mega-Charge Your Metabolism And Lose Excess Fat Once And For All! This Weight Blasting Method Is Easy AND Natural… And Will Give You The Hot Body And Killer Energy Levels You’ve Been Dreaming Of.

Get My Free Ebook

Post a comment