It is well known that the interpretation of any DSC trace requires a preliminary assessment of the base line. The base line of the DSC trace of many food samples can be rather 'irregular', when compared to that underlying the DSC peak of the fusion of Indium (the usual standard compound). This simply means that food samples undergo changes of the heat capacity with no parallel changes of enthalpy. This kind of processes can take place several times in the course of the temperature scan, since they are relevant to different phases of the food system.
It is therefore expedient to split the trace into regions, each relevant to a given main 'signal' (like an endo- or exo-thermic peak), which are to be ana
20 40 GO SO 100 130
Fig. 2 Starch gelatinization in a rice kernel suspended in excess water. The relevant endothermic signal has been singled out from the overall trace by splitting it into four gaussian functions lysed separately from one another. Within each region the base line trend can be tentatively defined with a SP-line (or even a straight line) across peak shaped signals, or with a sigmoid function when a simple transition is supposed to take place. Once the trace is accordingly scaled, it can be split into the minimum number of gaussian functions to attain an acceptable fit (P < 0.05). This treatment allows a tentative estimation of the enthalpy associated to each peak and the progress of the relevant transformation across the respective temperature span (Fig. 2).
The single peak can be finally interpreted according to either a thermodynamic or a kinetic approach. Comparison with DSC signals obtained from pure compounds can be of help to improve the analysis of each gaussian peak.
The glass transition temperature, Tg, is another quantity that can be determined with a DSC investigation. This transition is actually spread over a relatively wide temperature range and corresponds to the relaxation of the translation degrees of freedom within the sample investigated. The process is therefore accompanied by an increase of the heat capacity which produces the endother-mic shift of the base line of the DSC trace. As a typical finding, the glass transition is followed by other phenomena (Fig. 3) that are sustained by the increased molecular mobility, like crystallization of ice (exothermic peak) or enthalpy relaxation (endo- or exothermic peak).
Food sysiems that undergo depletion of liquid solvent (waier) on freezing host a residual liquid phase with increased viscosity where nucleation and growth of crystal phase (ice) is hindered: this liquid forms a glassy phase on further cooling . The same physical interpretation applies to solvent poor systems where the large viscosity hinders the formation of crystals of the 'solute' (e.g., sugars)
20 40 GO SO 100 130
Fig. 2 Starch gelatinization in a rice kernel suspended in excess water. The relevant endothermic signal has been singled out from the overall trace by splitting it into four gaussian functions
. But glass transition is also observed in systems where no solvent is present: the glass transition indeed occurs in every polymer material that is brittle for T < Tg and rubbery for T > Tg. Many foods contain biopolymers like amylose, amylopectin, gluten, etc. Therefore typical Tg shifts of the base line are found in the DSC traces of relatively dry food samples.
It is well known from polymer science that compounds with a small molecular mass can directly affect the molecular mobility of larger molecules and therefore modify the overall viscosity of the system , as revealed by changes of the glass transition temperature. The ubiquitous compound responsible for such effects in foods is water. It therefore is of interest the study of the aqueous binaries of a number of compounds by defining the relevant state diagram in the T-vs.-c(w) plane, where c(w) is the water content. The curve that fits the Tg-vs.-c(w) trend separates the underlying glassy region, where because of the low molecular mobility and high viscosity no transition or reaction can take place, from the upper region of the diagram (Fig. 4) where these changes can occur [5, 6].
A scheme of the various regions of the diagram can be summarized as follows. A liquidus curve fits the freezing points of water-rich binaries: it starts from T = 273.15 K for pure water and bends down with increasing solute content until it intercepts the Tg-vs.-c(w) curve in the point [Tg', c'(w)], that is the lowest temper ature at which a liquid phase can be observed in the presence of ice crystals. At higher solute contents the viscosity of the solution would be too high for any further ice nucleation and the expected eutectic point cannot be attained. Tg' accordingly is the Tg of maximally freeze-concentrated solutions.
For c(w) < c'(w) and T < Tg' the system is an amorphous glass. For c(w) > c'(w) and T < Tg' ice can still nucleate and grow although at a much lower rate. When a sample with c(w) > c'(w) is thawed (at a given rate) from T < Tg, the DSC trace shows a first endothermic shift of the base-line at T = Tg, which can be immediately followed by an exothermic wave that corresponds to ice crystal-
V LIQUID STATE
n. rubbery \
Fig. 4 Schematic view of a state diagram of an aqueous binary system that undergoes glass transition because of the high viscosity that hinder nucleation and growth of crystal phases. See text for lettering
Fig. 4 Schematic view of a state diagram of an aqueous binary system that undergoes glass transition because of the high viscosity that hinder nucleation and growth of crystal phases. See text for lettering lization. When Tg' is attained some ice melting takes place as revealed by an en-dothermic peak: the larger the c(w) the broader the endotherm. An example of such a behaviour is given in Fig. 3. The DSC trace for samples with c(w) < c'(w) shows an endothermic shift of the base-line at T= Tg which can be followed by other signals according to the nature of the solute. In the case of simple compounds, like sugars and pure polysaccharides, a broad endothermic peak is observed which corresponds to the solubilization into the liquid phase. The solubility curve (Tm-ra.-c(w) in Fig. 4). bends down from the melting point of the pure solute (when it actually exists) and crosses the curve of primary ice separation at the point [Tg', c'(w)]: DSC investigations indicate that this intersection may often occur at T > Tg' and c(w) > c'(w), as shown in the diagram in Fig. 4.
A number of applications and/or phenomena of technological interest, like freeze-drying, caking of powders, cryo-preservation, etc., have been described on the basis of the relevant phase diagrams , as well as an application to extrusion processing of flour . A number of papers [9-16] therefore appeared where various experimental approaches to Tg, like DSC, TMA DTMA, NMR, ESR, fluorescence and phosphorescence decay, etc., were reported and sometimes compared to each other. It however should be emphasized that some spec-troscopic techniques, like NMR and ESR, reveal changes related to short range molecular mobility within a s time scale, being practically blind for the macroscopic modifications of viscosity and specific heat at the operator's time scale which can be detected through other approaches, like thermal analyses. The comparison of the results of different techniques should therefore be considered taking into account the relevant time scale involved.
A particular emphasis has been recently given to the application of the Modulated Temperature DSC (MTDSC) to separate reversing from non-reversing heat-flow signals obtained from food systems. In MTDSC a sinusoidal temperature fluctuation is superimposed on a main increase of the temperature at a constant heating rate, po, so that the overall change of T is described by the expression
Where To, t, A, stand for starting temperature, time and amplitude of the temperature fluctuation, respectively, while ro = 2nv accounts for the frequency, v, of the fluctuation. The instantaneous heating rate is therefore
When A < (po v)/2n, P is always positive. The modulated temperature program produces a heat flow trace, HF, that fluctuates with the same frequency, v, and is the sum of two components, dubbed reversible and non reversible, respectively:
HFtot HFrev + HFnon rev.
where HFtot corresponds to the trace that would be obtained in a traditional DSC run performed at po heating rate. A Fourier analy sis allows these components to be separated in the form of two orthogonal heat capacities,
Cp' = |Cp | cos a Cp''= |Cp| sin a where the |C p | is the modulus of the heat capacity corresponding to the ratio between the amplitude of HF oscillation, AHF, and the amplitude of heating rate, (A*ro). The reversible heat flow is HFrev = po Cp'.
A major information drawn from MTDSC is relevant to the heat capacity drop observed at Tg from the stress-relaxation endotherm (non-reversing signal) that is often observed on heating samples previously cooled at subzero temperatures, like frozen doughs . The changes of the relaxation enthalpy are worth determining since they are related to the residual molecular mobility in quenched products and therefore with their stability and shelf-life. It has to be noticed that starch gelatinisation is seen as a totally irreversible process.
When dealing with an aqueous solution of a biopolymer, several conformational changes can take place above the Tg threshold, like formation of entangled chain gel, gel-sol transition, thermosetting, etc., according to the chemical nature of the compound . This often implies large changes of mechanical properties: it can be clearly demonstrated by coupling the DSC record with that of volume dilation (Fig. 5).
An interesting combination of DSC concerns X-ray diffraction. Synchrotron radiation is employed in these investigations. The small (SAXS) and wide (WAXS) angle X-ray scattering are of interest in food systems, like starch gels [18, 19] and fats . The X-ray beam of given wavelength is conveyed toward the calorimetric cell that is a glass capillary (0 ~1.5 mm) of about 20 mL volume. Fig. 6 reports a sketched view of the very complex instrumental apparatus .
Amylose crystallisation can be monitored during the isothermal annealing at adequate temperatures. The exothermic effect is indeed hard to detect, but the growth of B- and V-amylose crystal structure  can be neatly recognised in the relevant X-ray diffraction pattern. The same kind of information comes from the investigation of cocoa butter and milk fat, that contain large amounts of triglycerides showing a monotropic polymorphism related to the time and temperature of annealing .
In the case of amylose the DSC signal mainly concerns the fusion of amylose lipid complexes (Fig. 7), while in the case of triglicerides DSC traces are the resultant of the fusion of many coexisting crystal forms that have different melting points but very close fusion enthalpies (Fig. 8).
The combination of X-ray diffraction with DSC aliows in both cases a much better view of the transformations of interest, that are directly related to the quality of the hosting food, like bread, chocolate, milk fats, etc.
Another combination of DSC with gas analysis allows the characterization of processes with emission of volatile compounds. The instrument, that allows the simultaneous evaluation of the thermal effect and the amount of gas released, requires a fitting with a gas-chromatographer or a mass spectrometer. For this reason Calvet calorimeters are the most suitable. The gas coming out from the open sample cell is conveyed to the gas chromatographer. An independent circulation of inert gas flows through the reference cell. The combination with a mass spectrometer allows a more rapid identification of the released volatiles. The sample and reference materials are under continuous flow of gas during the experiment. The gas collection is performed at normal pressure by means of a capillary, which is kept hot to prevent condensation of the volatiles before the injection into the mass spectrometer inlet.
Simulation of heat treatment is among the aims of investigations on foods. Since most of these treatments, like cooking and baking, are carried out, or practically occur, in isothermal conditions, the investigations aimed to simulate the process should be carried out at constant temperature. Isothermal Calorimetry (IC) can be of help. The best approach requires the use of a calorimeter that can host sufficiently large cells (e.g., about 10 mL); as a standard procedure, one should first thermally equilibrate at the desired temperature the sample to be investigated and the an empty calorimetric vessel where it has to be dropped, in order to avoid initial misbalance of the instrumental output (Fig. 9). The time lag of the instrument is to be accounted before the treatment of the results (see appendix in ). These procedures were employed to characterize several food systems and processes, like starch gelatinization and retrogradation in cereal products, milk pasteurization, egg white denaturation, pasta and rice isothermal cooking, microbial growth, etc. [24-27].
Fig. 9 Isothermal 'cooking' of rice at various temperature. The traces have to be scaled taking into account the time lag of the instrument at each single temperature 
Food spoilage and preparation of particular dairy foods, like cheeses, yoghurt, kyr, etc., are sustained by specific microbial activity. In these cases too IC can be of help when coupled with the traditional microbiology techniques (Fig. 10). Cultures of living organisms, like yeast and fermenting bacteria, are poured into the calorimetric cell under strictly defined conditions, to obtain a calorimet-ric trace that may be directly related  to growth and metabolic rate.
Was this article helpful?