Results and discussion

CHARACTERISATION OF THE TG REGION WITH MDSC

MDSC has several advantages compared to conventional DSC for the determination of Tg of starchy systems. The total heat flow signal can be deconvoluted into heat capacity and non-reversing heat flow (Fig. 1). In the total heat flow signal the Tg can hardly be detected for this sample. Tg can be calculated much easier from the heat capacity curve. The increased sensitivity, resulting from the high instantaneous heating rate, together with the increased resolution, due to the low average (underlying) heating rate, are major benefits of MDSC for the determination of weak and broad thermal transitions, such as the glass transition of starch. This improves the accuracy of the determination of Tg and the change in heat capacity at Tg, ACp [46].

Temperature (°C)

Fig. 1 Determination of Tg with the MDSC heat capacity signal (lower curve). Upper curve: total heat flow

Temperature (°C)

Fig. 1 Determination of Tg with the MDSC heat capacity signal (lower curve). Upper curve: total heat flow

It is however necessary to apply the correct modulation conditions like amplitude, period and heating rate to obtain reliable results. These parameters have been optimised for the pan types used [19].

TG AS A FUNCTION OF THE WATER CONTENT

The influence of the moisture content on Tg of starch is depicted in Fig. 2. The results of this work are in good agreement with literature data [22, 34, 37], also shown for comparison.

Tg decreases by the addition of water. For concentrated starch systems, the addition of one percent moisture decreases Tg with about 15°C.

Several approaches have been proposed to account for the effect of plasticizer on the position of Tg. The Couchman and Karasz approach leads to the following expression for Tg of polymer-diluent mixtures [47]:

with Tg1: the glass transition temperature of the plasticizer (in K); Tg2: the glass transition temperature of the polymer (in K); ACp1: the change in heat capacity at Tg1 (in Jg-1K-1); ACp2: the change in heat capacity at Tg2 (in Jg-1K-1); W1: the weight fraction of the plasticizer, W2: the weight fraction of the polymer.

Equation 1 was tried out for starch-water systems, to estimate Tg of dry starch (Tg2) which cannot be determined experimentally. Tg1 and ACp1 of water are difficult to measure. Literature values are 134K and 1.94 Jg-1K-1, respectively [48]. ACp2 of dry starch was taken 0.38 Jg-1K-1 (extrapolated from own MDSC measurements). Applytng the above mentioned constants and Tg2 of starch as the only parameter to fit the Couchman-Karasz equation to the experimental results of this work (see Fig. 2), an optimised Tg2 value for dry waxy corn starch of 250°C is obtained. The fit for pregelatinised waxy corn starch is shown in Fig. 2. It should be noticed, however, that a ACp2 value as high as 0.47 Jg-1K-1 is reported [30]. Lower values are also reported: 0.42 Jg-1K-1 for amylopectin containing 17% moisture [35], 0.30 Jg-1K-1 for high molecular weight maltodextrins [18], and 0.295 Jg-1K-1 calculated for waxy corn starch casted at 90°C [34]. The highest value reported for ACp2, 0.47Jg-1K-1, gives a Tg2 value for dry starch of 217°C. If 0.30Jg-1K-1 is applied a Tg2 of about 300°C is found. Note however that by using this latter ACp2 value, the fitting is less good.

Fig. 2 Tg values of starch-water mixtures measured with DSC and MDSC: x this work, O K. J. Zeleznak and R. C. Hoseney [22], ■ H. Bizot et al. [34], a M. T. Kalichevsky et al. [37] The line is a fit of Eq. 1 (Couchman-Karasz)

Fig. 2 Tg values of starch-water mixtures measured with DSC and MDSC: x this work, O K. J. Zeleznak and R. C. Hoseney [22], ■ H. Bizot et al. [34], a M. T. Kalichevsky et al. [37] The line is a fit of Eq. 1 (Couchman-Karasz)

In literature, a wide range of values are reported for Tg of dry starch, obtained by extrapolation of materials containing various amounts of water: 125°C [28], 150°C [49], 243°C [18], 240-250°C [29], 285°C [34], 330°C (amylose) [27].

Below a certain concentration, called the maximal freeze concentration (Cg'), phase separation between plasticized starch and water occurs. In DSC, a melting endotherm of the water-rich phase becomes visible which overlaps with the glass transition of the plasticized starch-rich phase. In this work, the phase separation was found to occur at a waxy corn starch concentration of 73w%. The same value for Cg' has been reported by other researchers [10, 18, 21].

For concentrations near the maximal freeze concentration (Cg'), the decrease of Tg by the addition of water levels off (Fig. 2). The minimum Tg value of starch, called Tg', is reported to be 5°C [21] but other researchers have published values between 6°C [18] and 11°C [10].

In this study, the minimum value for Tg of plasticised starch is measured using MDSC. MDSC enables an accurate measurement of Tg during cooling (see Fig. 3), even if phase separation should occur thermodynamically. With MDSC the glass transition of the starch-rich phase can be separated from other thermal events during cooling. This is shown for a 68w% waxy corn starch sample in Fig. 3.

The total heat flow is separated into a non-reversing heat flow and a heat capacity signal. Tg calculated out of the Cp signal for this water content is -21°C. This determination would not be possible using only the total heat flow signal (see Fig. 3). These samples are in a non-equilibrium state during cooling and phase separation will occur during the subsequent heating. The 'cold-crystallisation', Tc, of the water-rich phase is observed in the total heat flow curve during

Temperature I Cl

Fig. 3 Total heat flow and heat capacity measured during cooling and subsequent heating of a 68w% starch-water system. Heating 1: cold crystallisation during heating. Heating 2: reheating after cold crystallisation. Tg' is at -5°C

Temperature I Cl

Fig. 3 Total heat flow and heat capacity measured during cooling and subsequent heating of a 68w% starch-water system. Heating 1: cold crystallisation during heating. Heating 2: reheating after cold crystallisation. Tg' is at -5°C

heating just above Tg (see Fig. 3: heating 1). The melting of this phase, at Tm, which is also seen in the heat capacity curve, follows immediately afterwards. The same effect is described for maltohexaose [46]. Note that, after phase separation, Tg of the starch-rich phase has increased in relation to the increased dry substance of this phase. By reheating the sample after cold crystallisation, Tg is shifted to Tg', which overlaps with the melting of the water-rich phase (see Fig. 3: heating 2).

A minimum value for Tg as low as -25°C is measured for a starch concentration of 66 w%. This means that even though a separate water-rich phase should be created, Tg continues to decrease, indicating that part of the additional water is still acting as a plasticizer, as long as the phase separation did not occur.

TG AS A FUNCTION OF MOLECULAR MASS

Tg of low molecular mass carbohydrates, dextrose (DPI) up to maltoheptaose (DP7), has been determined by (M)DSC analysis. In contrast with the large polydispersity of starch, the molecular weight of these maltooligosaccharides is monodisperse. The narrow distribution of the molecular mass, together with the increased ACp at Tg, increase the accuracy of the measurements.

Tg of these maltooligosaccharides was measured at different concentrations (Fig. 4). For the clarity of the graph, only maltotriose (DP3), maltotetraose (DP4) and maltoheptaose (DP7) are compared with the results for starch. The de crease in Tg with increasing water content is similar as for starch.

In literature large differences are found in the reported Tg values for dry maltooligosaccharides [21, 30]. This discrepancy is probably caused by the fact that some samples were not completely dry or that degradation occurred. For this last reason, Tg of dry maltooligosaccharides with Mw above DP5 (Tg = 174°C) is

Fig. 4 Tg of maltotriose (■), maltotetraose (a), maltoheptaose (♦) and starch (+) as a function of ds

difficult to measure without simultaneous degradation of the product. Therefore, these experiments were also done with conventional DSC at 10°Cmin-1 in this work, to minimise the time spent at high temperatures.

It is commonly found that the variation of Tg with degree of polymerisation (DP) can be described by an equation of the form:

where Tgœ is the high molecular mass limit of Tg and A is a constant.

Based on our measurements for dry maltooligosaccharides the extrapolated value to infinite molecular weight is 240°C. This value is comparable with the value of 250°C, obtained via extrapolation of Tg values of plasticized starch to 0 w% water.

TG AS A FUNCTION OF THE STARCH TYPE

The plasticizing properties of water on different kinds of starches (corn, waxy corn and potato) were investigated. The results are shown in Fig. 5. The decrease in Tg with the addition of water is similar for all starches studied, indicating that the difference in the amylose/amylopectin ratio (degree of branching) does not affect the position of Tg to a large extent. This finding is in agreement with literature [30]. Others found, however, that amylose has an increased Tg [27, 34].

60 70 80 90 100

Dty siibilaiikic (w%)

Fig. 5 Comparison of Tg of different types of starch as a function ofds: x waxy corn starch; • corn starch; a potato starch

60 70 80 90 100

Dty siibilaiikic (w%)

Fig. 5 Comparison of Tg of different types of starch as a function ofds: x waxy corn starch; • corn starch; a potato starch

SLOW CRYSTALLISATION OF STARCH: IN SITU MEASUREMENT WITH MODULATED DSC

Amorphous starchy materials crystallise very slowly and the crystallisation rate is strongly dependent on the starch concentration. Therefore attention has to be paid to prevent the evaporation of water (or other solvents) during the entire crystallisation. Moreover, the crystallisation process is characterised by a small exothermicity. These starch characteristics make it difficult to choose an appropriate method to measure the crystallisation behaviour of starchy systems.

The MDSC procedure used here is based on the fact that during crystallisation the heat capacity (Cp) of a material decreases [50]. With MDSC this (negative) change in heat capacity (ACpcryst) can be measured continuously, even in (quasi) isothermal conditions. From Fig. 6, it is clear that the decrease in heat capacity occurs on the same time scale as the increase of the heat of fusion. The heat of fusion was measured after partial crystallisation at the same temperature. It is obvious that analogous but continuous information on the same time scale is available from the (quasi) isothermal MDSC heat capacity signal. It was shown previously that the crystallinity evolution seen in the heat capacity signal coincides with the one measured with X-ray, DMA and Raman spectroscopy [19].

Crystallisation time (hi

Fig. 6 Crystallisation at 60°C of 69w% pregelatinised waxy corn starch, measured as a function of crystallisation time (h). The MDSC heat capacity evolution (in arbitrary units) is compared to the heat of fusion after partial crystallisation

Crystallisation time (hi

Fig. 6 Crystallisation at 60°C of 69w% pregelatinised waxy corn starch, measured as a function of crystallisation time (h). The MDSC heat capacity evolution (in arbitrary units) is compared to the heat of fusion after partial crystallisation

This reproducibility in a time span of 40h (and more) can only be achieved if several experimental conditions are fulfilled. The most important ones are; the fully amorphous nature, the exact concentration and homogeneous distribution of water in the starch sample before crystallisation, the constancy of temperature and water content in the starch system throughout (quasi) isothermal crystallisation. MDSC experiments with high-pressure stainless steel pans meet these experimental constraints. No water loss was noticed for all thermal treatments applied, and the deviation from the average temperature during (quasi) isothermal crystallisation was ± 0.01°C for the total time interval studied.

The MDSC method enables a quantification of the starch crystallisation process. The half conversion time (t1/2) is defined as the time to reach half of the decrease in Cp (1/2 ACpcryst; see Fig. 6). The reciprocal of this time (1/t1/2) can be used as a measure for the rate of isothermal crystallisation.

The heat flow phase signal of MDSC might also be interesting in the study of starch crystallisation, and reflects the change in heat capacity (ACpcryst)[51]. However, in the experimental conditions used, the change in the heat flow phase is not always reproducible and a small effect of the crystallisation process might be superimposed [52].

The tot al heat flow signal (not shown), equivalent with the conventional DSC signal cannot be used to determine the crystallisation behaviour of starch. The heat released is so small and is spread out over such a long crystallisation time, that the exothermal signal is no longer reliable due to baseline drift and noise. This technique, however, is very useful to follow the crystallisation behaviour of systems that crystallise fast, like the low molecular weight components lactose and sucrose [53].

MDSC is further used to study the crystallisation rate as a function of the crystallisation temperature, concentration, starch type and the effects of crystallisation on the thermal transitions (Tg and Tm).

INFLUENCE OF THE CRYSTALLISATION ON Tg

After a first heating to 170°C in HPS pans, the initial glass transition temperature (Tg0) of 76w% pregelatinised starch samples was measured with MDSC. These amorphous samples were then crystallised in MDSC at temperatures between 45°C and 100°C. After the (quasi) isothermal step, the samples were immediately

Fig. 7 Glass transition region (conventional DSC at 5°Cmin-1) for 76w%

pregelatinised waxy corn starch after crystallisation at 60°C for different crystallisation times t (from bottom to top: t = 0, 8, 12, 15, 24 and 48h)

Fig. 7 Glass transition region (conventional DSC at 5°Cmin-1) for 76w%

pregelatinised waxy corn starch after crystallisation at 60°C for different crystallisation times t (from bottom to top: t = 0, 8, 12, 15, 24 and 48h)

cooled below Tg. In a consecutive heating, the thermal properties of the semi-crystalline material were measured (Tg, the temperature range of melting and the enthalpy of fusion).

The glass transition region of 76w% starch was measured before and after (partial) crystallisation (Fig. 7).

Tg decreases continuously as a function of crystallisation time, from 9°C for 76w% amorphous starch to -4°C for semi-crystalline starch. This is explained by the fact that water is expelled from the crystals. The remaining amorphous phase of the semi-crystalline sample will, therefore, contain more water than initially in the fully amorphous sample. A final Tg of - 4°C corresponds to about 72 wt% starch. Note that for the highest crystallisation times, a small endothermic melting peak of ice is superimposed on the glass transition signal. In these later stages of crystallisation, a water-rich phase is segregating from the plasticized starch-rich amorphous phase. In literature, this phenomenon of water expulsion from the crystals, syneresis, is only described for diluted starch systems [20, 54]. Note that for the diluted systems the accompanying decrease in Tg could never be measured, because Tg0 already reached the minimum value for those diluted systems.

INFLUENCE OF THE ISOTHERMAL CRYSTALLISATION TEMPERATURE ON THE CRYSTALLISATION RATE

Both the melting temperature (region) and the glass transition temperature (region) of starch are important parameters controlling the rate of crystallisation. Crystallisation can only take place at temperatures between Tg and Tm [55]. Both thermal transitions are influenced by the concentration of water in the starch sample.

In Fig. 8, the half conversion time, ty2 defined as the time to reach half the decrease in Cp (see Fig. 6), measured with MDSC, is depicted as a function of the crystallisation temperature, giving rise to a bell-shaped curve. The bell-shaped

f 30

0 20 40 60 80 100

0 20 40 60 80 100

Fig. 8 Crystallisation rate as a function of the crystallisation temperature for different concentrations (x 78%, o 76%, ♦ 70%, ■ 60%). The lines are drawn as a guide to the eye

Fig. 8 Crystallisation rate as a function of the crystallisation temperature for different concentrations (x 78%, o 76%, ♦ 70%, ■ 60%). The lines are drawn as a guide to the eye curve is well-known for many polymers [21, 56]. For concentrated starch-water systems, however, experimental data are scarce, but predictions were made by modelling [9, 55]. Our data are consistent with these predictions.

For 76w% waxy corn starch, the bell-shaped curve goes through a minimum at 75°C (Tcmax). At Tcmax, approximately 65°C above Tg0, the maximum crystallisation rate is obtained. At crystallisation temperatures below Tcmax, the crystallisation rate decreases (t1/2 and tmax increase). Due to the higher viscosity at lower temperatures, transport of starch chains to the boundary of the starch crystal is restricted, and the crystallisation rate gets diffusion controlled. At temperatures above Tcmax, the crystallisation rate decreases as well (t1/2 and tmax increase again), since the thermodynamic driving force for crystallisation (primary and sec ond ary nu cle ation) de creases.

INFLUENCE OF THE CONCENTRATION

In Fig. 8 the bell-shaped curves for the different waxy corn starch concentrations (60, 70, 76 and 78%) are shown.

The crystallisation rate for 60w% samples was only measured at temperatures above 0°C. Therefore only part of the bell-shaped crystallisation rate curve was measured. The difference between 76w% and 78w% is small and difficult to measure accurately. The problem is that only the low end of the bell-shaped curve can be measured for the 78w% samples, since degradation of the material occurs if samples are kept for a long time (days) at temperatures above 100°C.

The temperature of maximum crystallisation rate (minimum in the bell-shaped curve), Tcmax, decreases with decreasing starch concentration. For 60w% samples, Tcmax is at about 25°C, for 70w% samples at about 60°C, and for 76w% samples at about 75°C. The scatter on the data for 60w% samples is very large. It is shown in literature that the retrogradation process could be monitored using X-ray diffraction and it was modelled by a physical formulation developed by Lauritzen and Hofmann [16, 17]. For 60 and 70w% samples, Tcmax values of 65°C and 80°C, respectively, were reported. No data were given for higher concentrations. These literature results are not in agreement with the findings of this work. Especially for the lowest starch concentration, a much lower Tcmax value is obtained.

Figure 8 also shows that the starch concentration affects the value of the maximum crystallisation rate at Tcmax, vmax (determined as t1/2 at Tcmax). The scatter on the data points, however, makes it difficult to accurately establish vmax. The value of vmax is highest at a concentration of 70w%; a value for t1/2 of about 330min is measured. By decreasing the concentration to 60w%, t1/2 is doubled (rate of crystallisation reduced to 50%). This could be explained either by the increasing difficulty to form stable nuclei at increased solvent concentrations or by the decreasing probability for chains to meet. For a starch-water system of 76w% starch, t1/2 increases to about 500min. This decrease in vmax is probably due to increasing viscosity.

Expulsion of water from the crystals during crystallisation has an effect on the crystallisation rate. For a sample with an initial concentration of 76%, Tg decreases from 9°C to -5°C thus the final concentration is approximately 73%. From Fig. 8 can be deduced that the crystallisation rate decreases with increasing water content for crystallisation temperatures well above Tcmax. For example from 76% to 70% at 90°C the rate decreases by a factor of about two. Since the change in concentration during crystallisation is less and since it occurs gradu -ally, the decrease in rate will also be somewhat less than a factor of two, but the influence will anyhow be important.

In a similar way can be seen that for the crystallisation temperatures below Tcmax the decreasing Tg during crystallisation will increase the crystallisation rate. This auto-catalytic effect can explain why the crystallisation rate around t1/2 is relatively high for low crystallisation temperatures, whereas for high crystallisation temperatures, the crystallisation retards itself. The effect might still be enhanced by an inhomogeneous water distribution [57].

UNIVERSAL CRYSTALLISATION RATE CURVE

An attempt was made to combine the results of Fig. 8 in order to obtain a universal crystallisation rate curve. All t1/2 values were normalised against the corresponding t1/2 at Tcmax (= vmax). These normalised rates were plotted as a function of (Tc-Tg)/(Te-Tg), with Tg, the glass transition temperature and Te, the end-temperature of melting (see page 64). In this way, the crystallisation temperature Tc is normalised too, ranging between 0 (Tc = Tg) and 1 (Tc = Te). This makes sense, as both Tg and Te are limiting the temperature range for crystallisation. The same procedure was applied for lactose and sucrose samples [58].

Fig. 9 Universal crystallisation rate curve for different concentrations of pregelatinised waxy corn starch (■ 60w%, ♦ 70w%, o 76w%): t1/2, normalised against t1/2 at Tcmax, as a function of (Tc-rg)/(re-rg)

Fig. 9 Universal crystallisation rate curve for different concentrations of pregelatinised waxy corn starch (■ 60w%, ♦ 70w%, o 76w%): t1/2, normalised against t1/2 at Tcmax, as a function of (Tc-rg)/(re-rg)

Tg values for totally miscible starch-water systems were taken calculated from equation 1, even for concentrations where phase separation might occur. In this case, the crystallisation rate is related to the calculated Tg for an ideal, homogeneous starch-water system without phase separation for the whole concentration range [59]. Figure 9 clearly indicates that with this procedure a universal bell-shaped crystallisation rate curve for starch can be obtained. The value of Te was taken as the melting temperature of the most stable crystals or the end-set temperature. The values for Tg and Te used in both approaches are listed in Table 1.

Te 115 136 149

Figure 9 clearly shows that a universal crystallisation rate curve for starch, independent of the starch concentration, can be obtained. Differences in Tcmax for different concentrations are explained by changes in Tg and Te, which in turn are explained by differences in water content. The normalised value of Tcmax for all concentrations is at ca. 0.5 on the normalised temperature scale, meaning a Tcmax value almost equal to (Tg + Te)/2.

Figure 10 shows the starch-water state diagram. Te and Tcmax decrease almost linearly in the concentration range studied.

A linear extrapolation of these results enables an estimation of Tcmax for other concentrations. Extrapolation shows that for concentrations below 50w% (food

Fig. 10 State diagram showing temperature of maximum crystallisation rate, Tcmax, end -temperature of melting, Te, and glass transition temperature, Tg, as a function of the starch concentration. Tg is calculated according to eq.1. Tg = -5°C at C ' = 73w%

systems), the calculated Tcmaxis below Tg'. As explained before, phase separation of a water-rich phase and a starch-rich phase occurs during the crystallisation of the water-rich phase upon cooling (or reheating). Due to this interfering phase separation, the highest crystallisation rates for starch concentrations below 50w% will be found around 0°C, since crystallisation below the glass transition of the starch-rich phase (Tg' = -5°C) is prohibited by diffusion limitations.

INFLUENCE OF THE STARCH-TYPE [57]

Next to pregellatinised waxy corn starch, the crystallisation of pregelatinised corn starch and pregelatinised amylose extender waxy starch (aewx) was also performed in MDSC. The comparison between waxy corn starch, aewx starch and corn starch was made for 76w% starch samples.

The crystallisation rate of aewx and corn starch is much higher than the crystallisation rate of waxy corn starch. t1/2 at Tcmax is less than 45min for aewx and 60min for corn compared to 500min for waxy corn.

MELTING OF STARCH [60]

The samples crystallised in MDSC conditions are crystallised quasi-isothermally, but no difference can be observed in their melting profile when compared to iso-thermally crystallised samples. The melting profiles of starch samples prepared at different concentrations are shown in Fig. 11. All melting profiles are very broad and at least bimodal. The broad melting range indicates a great heterogeneity of starch crystals.

70 90 110 130 150 170 Temperature (°C)

Fig. 11 Influence of Tc (indicated) on the melting endotherms for 78 and 60w% pregelatinised waxy corn starch (conventional DSC at 5°C/min)

The start of melting, To, shifts with the isothermal crystallisation temperature, Tc. Melting starts about 16°C above Tc for low values of Tc, whereas at high values of Tc this interval is reduced to about 5°C. This finding is independent of the concentration.

The end temperature, Te, of melting is only slightly influenced by Tc. For 78w% samples the increase is only about 5°C for an increase in Tc of 40°C. The fact that Te remains almost constant whereas To increases substantially with Tc results in a narrowing of the melting range with increasing Tc.

Te changes largely with the concentration, from 115°C for 60w% to 150°C for 78w% starch.

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