Thermal evaluation of sucrose-maltodextrin-sodium citrate bioglass: Glass transition temperature

Food Hydrocolloids 60 (2016) 589e597

Contents lists available at ScienceDirect

Food Hydrocolloids journal homepage: www.elsevier.com/locate/foodhyd

Thermal evaluation of sucrose-maltodextrin-sodium citrate bioglass: Glass transition temperature Eakasit Sritham, Sundaram Gunasekaran* Department of Biological Systems Engineering, University of WisconsineMadison, 460 Henry Mall, Madison, WI 53706, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 February 2016 Received in revised form 18 April 2016 Accepted 20 April 2016 Available online 21 April 2016

Sucrose-maltodextrin-sodium citrate (SC-MD-NaCit) bioglass model systems were studied using modulated differential scanning calorimetry (DSC) to test the hypothesis that addition of salt to a complex amorphous carbohydrate system at low level of moisture content will significantly affect its glass transition properties. Samples were formulated with different SC/MD ratios (7:3, 5:5 and 3:7, by mass) and NaCit/SC ratios (0, 0.1 and 0.2, by mole) and two levels of residual moisture content, low (0.27 e0.35%wb) and high (2.83e4.40%wb). The glass transition characteristics of these systems were strongly dependent on the moisture content and other constituents. On average, glass transition temperature (Tg) of the 7:3, 5:5 and 3:7 SC/MD systems were approximately 77, 84 and 101
C, respectively. The Tg values tended to increase when NaCit was added, with a noticeable increase occurring in the 7:3 SC/MD system at low moisture content. The increase in moisture content from low to high level had a significant plasticization effect on the bioglass as elucidated based on the decrease in the Tg from approximately 106 to 67
C. The DSC thermograms suggested that water molecules may interfere with intermolecular interactions between the glass-forming molecules causing changes in the molecular mobility of the bioglass matrix. The findings reveal that addition of NaCit can enhance the stability of low-moisture bioglass by primarily interacting with SC and forming large, less-mobile clusters, which helps to improve the Tg and restrict matrix mobility. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Bioglass Differential scanning calorimetry Glass transition Maltodextrin Sodium citrate Sucrose

1. Introduction When a liquid is rapidly cooled the equilibrium crystalline state may be avoided and an amorphous material is obtained. This phenomenon occurs due to a tremendous increase in viscosity during cooling that prevents molecules to rearrange themselves into a crystal lattice. The amorphous material exists as either a highly viscous supercooled liquid or rubbery material or glassy solid, depending on temperature (Allen, 1993; Elliott, Rao, & Thomas, 1986). The temperature at which a material undergoes the transformation from rubbery to glassy state is known as the ‘glass-transition temperature (Tg)’ (Young & Lovell, 2011). Such a process to convert a material into a glassy solid without any crystalline structure is also known as vitrification. A significantly reduced molecular mobility in the glassy state leads to its relative stability against crystallization while the rubbery state is considerably physically unstable (Roos, 1995).

* Corresponding author. E-mail address: [email protected] (S. Gunasekaran). http://dx.doi.org/10.1016/j.foodhyd.2016.04.030 0268-005X/© 2016 Elsevier Ltd. All rights reserved.

Food material like sugars can be readily vitrified upon cooling their concentrated solutions. Amorphous sugar systems are capable of preserving seeds, spores, and living organisms under extreme conditions (Buitink & Leprince, 2004; Crowe et al., 2001). Sugars and some other carbohydrates in the glassy state also help modulate the stability of dry and frozen food and pharmaceutical products (Roos, 1995). A number of common food processes, such as drying, freezing, grinding, extrusion etc. yield products in the amorphous or partially amorphous states (Liu, Bhandari, & Zhou, 2007). These products usually face stability issues. In general, water activity of glassy products is considerably low and hence their microbial stability is not a significant concern; however, their chemical and structural stabilities are (Poirier-Brulez, Roudaut, Champion, Tanguy, & Simatos, 2006). The Tg is an important property of amorphous materials; it is related to a number of physiochemical changes in dried food products, for instance crystallization, collapse, stickiness etc. Accordingly, the Tg has become a good indicator of quality changes in food products during storage (Bhandari & Howes, 1999; Le Meste, Champion, Roudaut, Blond, & Simatos, 2002). However, in some cases, temperature and moisture

590

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597

content related changes in food and biomaterial properties are not satisfactorily explained using solely the concept of glass transition (Champion, Le Meste, & Simatos, 2000). Like many other amorphous systems, carbohydrate glasses possess excellent protective property which makes them attractive for numerous food applications. However, materials in the amorphous state are thermodynamically out-of-equilibrium. Their properties change toward equilibrium with time, generally known as ‘aging.’ Food technologists are then challenged as how to formulate amorphous carbohydrates with desirable functional properties for the designated applications, and substantial degree of stability. Of course, the knowledge about characteristics and parameters that could affect the formulated system is fundamentally indispensable. The use of Tg as a reference temperature for predicting the rate of diffusion-controlled process was found to be relevant (Champion et al., 2000), even though not exclusively since certain reactions can still take place at temperatures below Tg (Hancock & Zografi, 1997). Among glass-forming food materials, sucrose (SC) can be readily prepared as amorphous matrix. SC is able to form hydrogen bonds with polymers to a greater extent than other sugars having a higher Tg. This ability of SC could be of prime importance for the stabilization of polymers in amorphous SC matrix (Davidson & Sun, 2001; Taylor & Zografi, 1998; Wolkers, van Kilsdonk, & Hoekstra, 1998). The utilization of sugar glasses as an encapsulation matrix could be enhanced by mixing with larger biomolecules such as maltodextrin (MD); this combination could also contribute to certain desirable characteristics. In spray drying, for example, an ideal carrier should have a high degree of solubility, limited viscosity at 30e45% solid content, emulsifying characteristics, good drying properties, bland taste, and non-reactivity. MD serves well for these purposes (Desobry, Netto, & Labuza, 1999; Galmarini, Zamora, Baby, Chirife, & Mesina, 2008). In addition, MD has proven to inhibit crystallization in amorphous SC matrix and provide good product stability to dry powder products (Bhandari & Hartel, 2005). It was also found that glasses which is formulated from the mixtures of SC and MD exhibit stronger hydrogen bonding than pure SC glasses (Oldenhof, Wolkers, Fonseca, Passot, & Marin, 2005). Other interesting functional properties of MD are the ability to bind flavors and fat, and to serve as oxygen barrier (Chronakis, 1998). From economic stand point, both SC and MD are considerably inexpensive and substantially available. Other than mixing with larger biomolecules, the stability of an amorphous sugar matrix could also be enhanced by the inclusion of small amount of salt. Salts can inhibit crystallization by altering molecular interactions and reducing molecular mobility (Izutsu & Aoyagi, 2005). The result from a SC crystallization study suggested that the presence of salts constrained number of configurations for crystal growth. This could be attributed to the effects of salt on the nucleation mechanism of ion-induced microheterogeneities in the supercooled solutions. It was also found that the crystallization was delayed without affecting the Tg of the system (Longinotti, Mazzobre, Buera, & Corti, 2002). As a result, it was proposed that the salt effect on SC crystallization occurs at a molecular level, but in such a dynamic way that the cooperative relaxations that give rise to glass transitions are not affected (Buera, Schebor, & Elizalde, 2005). Some recent studies reported the potential of salts to enhance the stability of sugar glasses when residual moisture was less than 2%. It was suggested that the mechanisms that help improve matrix stability are interactions between either ions (You & Ludescher, 2008) or the functional groups (Kets, Ijpelaar, Hoekstra, & Vromans, 2004) of salts and sugar molecules. Sodium citrate (NaCit) is among the salts that is of much of interest from researchers. This type of salts has been widely used in food products

for a number of purposes. Based on these evidences found in the literature we hypothesis that addition of salt to a complex amorphous carbohydrate system at low level of moisture content will significantly affect its glass transition properties. We consider SC, MD, and NaCit are good candidates to utilize as amorphous matrix components for various applications in foods. However, efficient formulation and process control would be made possible only after the characteristics and behaviors of SC-MD-NaCit system have been thoroughly investigated. Thus, our objective was to investigate the dependence of calorimetric properties and behaviors of the SC-MD bioglass systems as a function of SC, MD, NaCit and moisture content. 2. Materials and methods 2.1. Materials SC (crystalline, C12H22O11; molecular weight (Mw), 342.2965 g/ mol; melting point, 190
C; Chemical Abstracts Service (CAS) number, 57-50-1) and NaCit (trisodium citrate dehydrate, C6H5Na3O7.2H2O; Mw, 294.099568 g/mol; melting point, >300
C; CAS number, 6132-04-3) were purchased from Fisher Scientific (Fair Lawn, NJ). MD (powder, H(C6H10O5)2-OH; Mw, 342.29648 g/ mol; dextrose equivalent (DE) maximum 20%; melting point, 240
C; CAS number, 9050-36-6) was purchased from Spectrum Chemical Manufacturing Corporation (Gardena, CA). All ingredients were used as received without further purification. 2.2. Sample preparation The experimental factors and their levels studied were as follows: SC/MD ratio: 7:3, 5:5, and 3:7 (by mass); NaCit/SC ratio: 0, 0.1, 0.2 (by mole) and equilibrating relative humidity (RH): low (LRH) and high (HRH). The actual mass values of ingredients used and the experimental design and treatments are listed in Table 1. Two replicates were prepared for each treatment, and the order of experiment was completely randomized. Ingredients were weighed as necessary and dissolved in deionized (DI) water to make a total volume of ~150 mL. These solutions were stirred at 150 rpm using a magnetic stirrer for 3 h at room temperature. Then the sample solutions were held in a glass beaker and continuously stirred while heating over a hot plate while monitoring the solution temperature with a thermometer; heating was increased gradually until the solution temperature reached 175
C and held there to bring the moisture content decreased to approximately 4.5%e9%, which was the lowest attainable. The progress of moisture reduction was monitored by periodic weighing. The final moisture content of all samples was measured via the Karl Fischer titration technique using the Volumetric Karl Fischer titrator, model 795 KFT Titrino (Metrohm Ltd., Herisau, Switzerland) at 50
C as recommended for most confectionary products (Hoffmann & Felgner, 2011). Once moisture content reached the level mentioned above, the molten mixtures were poured over a flat surface of a Teflon sheet and spread with a spatula to a thickness of 2e3 mm to rapidly cool them down to room temperature and then transferred to a desiccator to maintain the moisture content. Glass samples were subsequently ground using a mortar and pestle and sieved through a 170-mesh screen (WS Tyler Sieve 800 FH-SS-SS-US-170, Sepor Inc., Wilmington, CA), corresponding to a 90 mm opening, to obtain samples as fine powder. Samples were stored in jars containing phosphorus pentoxide (P2O5) powder for low relative humidity (RH) of ~1% and lithium chloride (LiCl) solution for high RH of ~11% for at least three months to allow for equilibration.

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597

591

Table 1 Sample composition and sample preparation conditions used in the experiments. Equilibrating atmosphere

Na citrate/ Sucrose

Sucrose: Maltodextrin

Sucrose (g)

Maltodextrin (g)

Na citrate (g)

Low RH (~1%)

0 0 0 0.1 0.1 0.1 0.2 0.2 0.2 0 0 0 0.1 0.1 0.1 0.2 0.2 0.2

7:3 5:5 3:7 7:3 5:5 3:7 7:3 5:5 3:7 7:3 5:5 3:7 7:3 5:5 3:7 7:3 5:5 3:7

70 50 30 66.78 48.33 29.39 63.84 46.77 28.81 70 50 30 66.78 48.33 29.39 63.84 46.77 28.81

30 50 70 28.62 48.33 68.58 27.36 46.77 67.22 30 50 70 28.62 48.33 68.58 27.36 46.77 67.22

0 0 0 4.61 3.33 2.03 8.81 6.45 3.97 0 0 0 4.61 3.33 2.03 8.81 6.45 3.97

High RH (~11%)

2.3. Differential scanning calorimetry measurements

2.4. Statistical analysis

The calorimetric measurements were carried out using a modulated differential scanning calorimeter (MDSC) equipped with a refrigerated cooling system (Model 2920, TA Instruments, New Castle, DE). Nitrogen was used as purge gas to prevent condensation and control the atmosphere within DSC cell at the flow rate of 35 mL/min. It was also used as a cooling medium flowing through the cooling unit at the rate of 150 mL/min. A wide range of temperature scan rates, from 0.1 to 20
C/min are often used; however, since calorimetric Tg is sensitive to the scan rate, both during heating and cooling, Angell (1997) suggested a rate of 10
C/min to allow more meaningful comparison with the results obtained from spectroscopy techniques such as dielectric or dynamic mechanical analysis. This heating rate has also been suggested to minimize the impact of experimental conditions on enthalpy relaxation determination (Liu, Bhandari, & Zhou, 2006). Hence, a heating rate of 10
C/min was selected along with the cooling rate of 15
C/min for all experiments. The temperature and heat flow calibrations were performed on the MDSC system using indium (melting point of 156.6
C), tin (melting point of 231.93
C), and distilled water (melting point of 0.0
C). Temperature modulation amplitude was set as ±1
C with a period of 60 s. Hermetic aluminum pans and lids were weighed prior to filling with powder samples inside a glove box flushed with nitrogen gas. The RH inside the glove box was brought down to ~5% before samples were taken out from the storage bottles. Samples were filled into the pans, closed with the pre-paired lid, and tightly sealed using a sample press. The sealed sample pans were then weighed and the sample mass was calculated, which were in the range of 10e17 mg. A sealed empty pan was used as a reference placing it inside the DSC cell along with the pan containing the measurement glass sample. Samples were subjected to two cycles of heating and cooling temperature scans in the range of approximately 40
C below the onset Tg up to approximately 40
C above the offset Tg of each sample. The first temperature scan was to erase thermal history of the samples. The Tg and the onset, the midpoint, and the offset, as well as the change in heat capacity at glass transition (DCp,GT) values were determined from the second temperature scan. Software from the DSC manufacturer (TA Instruments, New castle, DE) was used for obtaining data from the thermograms.

The results were statistically analyzed using the IBM SPSS Statistics 19 computer package. The analysis of variance (ANOVA) Ftest was first conducted using the general linear model (GLM). A post-hoc analysis was then performed when any significant effect was detected. Comparisons using the Least Significant Difference (LSD) for one pair comparison or the Tukey's range test for simultaneous multiple comparisons were conducted at a significance level (a) of 0.05. The Tukey test is preferable since it helps to control the experimental-wise or family error rate e the overall error rate e at the selected a level (Montgomery, 2005). 3. Results and discussion 3.1. Moisture content The moisture content of samples equilibrated in low- and highRH atmosphere were in the range of 0.27e0.35%wb (mean value ¼ 0.31 ± 0.04%wb) and 2.83e4.4%wb (mean value ¼ 3.64 ± 0.52%wb), respectively (Table 1). Further analyses indicated that the moisture contents of samples equilibrated in low RH atmosphere were not significantly different (p ¼ 0.906) while this was not the case for those samples equilibrated in high RH atmosphere (p < 0.001). In large part, the residual moisture content of the samples equilibrated in high RH tended to increase with increasing MD or NaCit concentration. The average values of residual moisture content for the samples with SC/MD of 7:3, 5:5 and 3:7 were approximately 3.10, 3.64 and 4.20%wb, respectively. For samples with NaCit/SC of 0, 0.1 and 0.2, the average values were around 3.39, 3.69 and 3.85%wb, respectively. At very low level of moisture content, the samples with high MD content became highly viscous, as observed by the slowing of the rotating stirrer bar, and moisture evaporation was extremely slow. High solution viscosity also made it difficult for the magnetic stirrer to function, virtually stopping it at the very end of the heat evaporation process. Note that we limited the heating step to 175
C to avoid decomposition of SC, which reportedly occurs near the vicinity of melting of many sugars (Jiang, Liu, Bhandari, & Zhou, 2008). Melting point of SC we used is 190
C. The final moisture content of samples with high MD concentration after heat evaporation did not reach the pre-determined moisture content of 4.5%wb. MD consists of glucose polymers

592

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597

with a broad spectrum of chain lengths and molecular weights (Chronakis, 1998; McPherson & Seib, 1997). These bulky molecules lead to an increased viscosity of bioglass systems. Also, the threedimensional network of glucose polymers in MD possesses a high degree of water holding capacity (Chronakis, 1998). Accordingly, high MD concentration may adversely affect a bioglass system. Firstly, high viscosity would cause a systematic difficulty in certain processes such as extrusion cooking. Secondly, the water held inside the matrix of glucose polymers could in turn plasticize the system and hence lower the Tg.

3.2. Glass transition temperature The extent of relaxation is known to affect the Tg value of amorphous materials, particularly when measured with conventional DSC (Surana, Pyne, Rani, & Suryanarayanan, 2005; Wungtanagorn & Schmidt, 2001). Accordingly, representing the Tg values of samples stored at different times and temperatures is inappropriate. Thus the Tg values reported here are those collected from samples before aging. In addition, two measuring techniques were applied to obtain more precise values of Tg. First, each sample was heated to around 40
C above its Tg as suggested in the literature (Liu et al., 2006). Second, the MDSC technique we used allows to decompose the total heat flow into reversing and non-reversing components. This way, the reversible transitions (glass transition, melting) can be observed separately from the non-reversible transitions (enthalpy recovery, evaporation), and the Tg values could be determined from the reversing heat flow signal without interferences from the endothermic peak associated with enthalpy recovery, or other peaks induced by thermal history. A large number of Tg measurements were performed according to the experimental design (Table 1). For low-moisture samples there were 648 measurements (36 per treatment  9 treatments  2 replicates), and for high-moisture samples there were 54 measurements (3 per treatment  9 treatments  2 replicates). The Tg values reported in Table 2 represent mean ± standard deviation from two replicates; the value of each replicate was obtained by averaging over a number of measurements mentioned above. Fig. 1 is a typical of DSC heat flow and the locations of onset-Tg, midpoint-Tg, and offset-Tg on a reversing heat

Fig. 1. Representative MDSC thermogram obtained during the glass transition of a bioglass sample (SC/MD ¼ 5:5, NaCit/SC ¼ 0.1, low-level moisture content). The endothermic peak as a result of enthalpy relaxation can be seen in non-reversing heat flow signal. Locations of Tg including the onset-, midpoint-, and the offset-Tg values determined from the reversing heat flow signal are illustrated.

flow signal. However, we used midpoint-Tg for our discussion. The ANOVA revealed that all the main effects from three factors e RH of equilibrating atmosphere (i.e., level of moisture content) and NaCit/SC and SC/MD ratios e and all two-way and three-way interaction effects on Tg values were statistically significant (p < 0.001). From further analyses, by considering them according to the levels of residual moisture content, the interaction effect between NaCit/SC and SC/MD was significant (p < 0.001) only for high-moisture samples; the effects of NaCit/SC and SC/MD on Tg

Table 2 Residual moisture content, glass transition temperature (Tg) and width of glass transition zone values of samples (equilibrated at low and high relative humidity (RH)) used in the experiments. SC:MD

NaCit/SC

Residual moisture content (%wb)

Tg (
C)

Width of glass transition (
C)

Onset Low RH (~1% RH) atmosphere (over P2O5 powder) 7:3 0.0 0.28 ± 0.01 7:3 0.1 0.31 ± 0.06 7:3 0.2 0.33 ± 0.01 5:5 0 0.34 ± 0.03 5:5 0.1 0.27 ± 0.03 5:5 0.2 0.31 ± 0.01 3:7 0.0 0.34 ± 0.11 3:7 0.1 0.35 ± 0.01 3:7 0.2 0.32 ± 0.01 High RH (~11%) atmosphere (over saturated LiCl solution) 7:3 0.0 2.83 ± 0.11 7:3 0.1 3.07 ± 0.02 7:3 0.2 3.40 ± 0.01 5:5 0.0 3.29 ± 0.02 5:5 0.1 3.86 ± 0.16 5:5 0.2 3.76 ± 0.04 3:7 0.0 4.04 ± 0.03 3:7 0.1 4.15 ± 0.17 3:7 0.2 4.40 ± 0.00

Midpoint

Offset

79.3 82.6 87.8 89.1 87.6 95.9 123.4 106.8 112.3

± ± ± ± ± ± ± ± ±

0.3 0.4 0.7 1.6 0.1 0.8 1.2 9.9 0.1

86.6 91.4 98.4 99.2 95.2 108.1 136.9 119.0 123.3

± ± ± ± ± ± ± ± ±

0.3 0.3 1.0 1.8 0.3 0.3 0.8 3.3 0.1

93.5 97.9 104.8 110.5 104.0 115.8 146.4 133.8 134.4

± ± ± ± ± ± ± ± ±

0.4 0.8 0.8 1.1 0.1 0.3 0.4 1.6 0.1

14.2 15.3 17.0 21.3 16.4 19.9 22.9 27.0 22.2

± ± ± ± ± ± ± ± ±

0.7 0.5 0.1 0.6 0.2 0.5 0.8 8.2 0.0

53.2 55.9 57.7 55.8 58.0 59.4 64.4 70.4 69.2

± ± ± ± ± ± ± ± ±

0.1 0.1 0.2 0.3 0.1 0.1 6.6 1.9 1.3

58.7 61.4 63.2 61.4 63.7 65.5 72.3 77.0 77.2

± ± ± ± ± ± ± ± ±

0.2 0.2 0.2 0.2 0.0 0.1 5.3 1.9 0.4

62.5 64.6 67.2 65.2 67.9 69.8 77.0 82.3 81.8

± ± ± ± ± ± ± ± ±

0.1 0.2 0.4 0.1 0.3 0.2 7.0 2.3 2.4

9.3 8.7 9.5 9.5 9.9 10.4 12.6 11.9 12.6

± ± ± ± ± ± ± ± ±

0.2 0.1 0.2 0.2 0.2 0.1 0.4 0.4 1.1

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597

values were independent. The Tg values of samples with different levels of residual moisture content were significantly different. On average, increasing the residual moisture content from low (0.27e0.35%wb) to high (2.83e4.4%wb) level resulted in the reduction of Tg from approximately 106 to 67
C, which indicates a strong and well-known plasticization effect of water. At high moisture content, the mean Tg values are significantly higher with NaCit, 67.4e68.6
C compared to 64.2
C for the samples without NaCit. However, the Tg values of samples with different NaCit/SC did not differ significantly. As SC/MD decreased, or the concentration of MD increased, the Tg increased significantly at every level of increase in MD, which is expected given its large molecular weight (Le Meste et al., 2002). In low-moisture samples, the effects of NaCit/SC and SC/MD could not be considered independently. There was a significant positive correlation between the addition of NaCit and Tg, which was clearly evidenced in 7:3 SC/MD samples. Without NaCit, the Tg was ~87
C. The Tg increased about 5 and 12
C when NaCit/SC values were 0.1 and 0.2, respectively. For the 5:5 SC/MD samples, a significant increase in Tg was observed only when NaCit/SC was 0.2. However, the effect of NaCit was different in the 3:7 SC/MD samples; the Tg decreased when NaCit was added. Interestingly, when NaCit/SC increased from 0.1 to 0.2, the Tg increased from 119.0 to 123.3
C; but it was still lower than that of the system without NaCit, which is 136.9
C. The effect of NaCit on the Tg value of low-moisture samples depended on SC/MD. The positive dependence of Tg on NaCit was less pronounced as MD concentration increased, and the correlation even reversed when MD was at the highest level. The result suggests that NaCit might primarily interact with SC rather than MD. Accordingly, in the system dominated by MD when small amount of NaCit (NaCit/SC ¼ 0.1) was added some NaCit molecules or their ions might interact with SC while some might simply plasticize the system and result in reducing the Tg. When NaCit concentration was increased (NaCit/SC ¼ 0.2), the increased interaction between NaCit and SC which firmed up the large, lessmobile clusters resulting in less plasticization, and hence the Tg was lowered to a smaller extent. It has been reported that the Tg value of SC-NaCit mixture with low moisture content dramatically increased with the addition of NaCit up to NaCit/SC ¼ 2.5 by mass, or 2.9 by mole (Kets et al., 2004), which was much higher than what we used. The interaction between salt and hydroxyl group of sugar could be in the form of ionedipole interactions (You & Ludescher, 2008). This scenario may be similar to what is generally found in amorphous polymer systems, where side groups could raise the Tg if they are polar or large, and bulky (Young & Lovell, 2011). The greater tendency for NaCit to interact with SC is likely due to the difference in the size and structure between SC and MD molecules. Oldenhof et al. reported that SC provides better protection for protein during drying by preventing dehydration-induced conformational changes as it interacted with protein more effectively than MD (DE ¼ 5e8) (Oldenhof et al., 2005). Koster et al. reported that only a small molecule MD could stabilize dry membranes (Koster, Maddocks, & Bryant, 2003). They reported an apparent size-exclusion limit for MDs between dehydrated phosphatidylcholine membranes between Mw 1000 and 5000. Thus, in general, the effectiveness of carbohydrate to stabilize protein during drying decreases with its increasing size (Wolkers et al., 1998). The effect of salts on Tg value of a system is also dependent on the type of salt (Izutsu, Fujimaki, Kuwabara, & Aoyagi, 2005). From a study on amorphous indomethacin salts, it is known that Tg value decreases with the increasing ionic radius of alkaline cation. The decrease in charge density of cation, as the ionic radius increases, may lower electrostatic interaction between the ions causing Tg depression (Tong, Taylor, & Zografi, 2002). In our preliminary

593

experiments (data not shown), adding NaCl into SC-MD mixture caused Tg depression. Farahnaky et al. also reported the Tg depression in amorphous potato and cassava starches upon adding NaCl up to 6% dry basis (Farahnaky, Farhat, Mitchell, & Hill, 2009). Salt renders starch system more hygroscopic, and so the decrease of Tg is due to the plasticization effects of both NaCl itself and the absorbed water, in which the latter showed the greater effect. At constant moisture content, the reduction of Tg due to NaCl at low moisture levels was more evident than that at high moisture level. However, the mechanism of Tg depression by NaCl is still ambiguous. Possibly, NaCl may not dissociate to ions, but simply acts as a plasticizer like other small molecules, or the dissociated atom of NaCl may reduce hydrogen in starch molecule by occupying a number of available sites (Farahnaky et al., 2009). In this study, Tg of high-moisture samples generally increased with the increasing NaCit concentration. However, the most dramatic effect of NaCit was found in the samples with high SC content and at low moisture level. The result suggests that SC and NaCit ions could efficiently interact only at very low moisture environment. Kets et al. reported a SC-NaCit mixture with a very high Tg of approximately 105
C when residual moisture was removed (Kets et al., 2004). This enhanced Tg value even exceeds both individual Tg for pure SC and NaCit. The observation led to a hypothetical explanation that removal of water allows SC and NaCit to restructure through hydrogen bonding more efficiently (Kets et al., 2004). In other words, water may interfere with hydrogen bonding between SC and NaCit. Water which usually prevails in food systems is known for its plasticization effect and the effect on molecular mobility of the systems; though the mechanisms are poorly understood (Liu et al., 2006). Kilburn et al. proposed two complex mechanisms potentially involved in the plasticization effect of water on carbohydrate systems (Kilburn et al., 2004). These are the formation and disruption of intermolecular hydrogen bonds among carbohydrates, and the changes of free volume in the matrix. In the absence of water, hydrogen bonding among carbohydrate molecules gives rise to the formation of less-mobile large clusters. Once absorbed into glassy carbohydrate matrix, water may either fill up the small voids or disrupt the intermolecular hydrogen bonding, which results in a higher degree of freedom for carbohydrate molecules (Liu et al., 2006). On the whole, it was clear that the Tg values of our bioglass samples primarily depended on SC/MD ratio; the higher the MD concentration, the higher the Tg value. Such dependence of Tg on molecular weight and structure of glass-forming molecules in an amorphous system is well documented (Orford, Parker, & Ring, 1990). 3.3. Width of glass transition zone Glass transition is a kinetic process, depending on the cooling rate, and takes place over a range of temperature. Fast cooling rate allows the transition to occur at a higher temperature range than does the slower cooling rate (Roudaut, Simatos, Champion, Contreras-Lopez, & Le Meste, 2004). Food products commonly contain a number of ingredients rendering them chemically heterogeneous, which dictates the width of glass transition zone (WGT). That is, the greater extent of heterogeneity, the wider the WGT (Ferry, 1980). The WGT, defined as the temperature range from the onset-Tg to the offset-Tg, for the bioglass systems studied are listed in Table 2. The ranges of WGT for the systems with low and high levels of residual moisture content were 14e22
C and 9e13
C, respectively. These ranges are comparable to the WGT of food saccharides reported in the literature, which is generally around 5e20
C (Liu et al., 2006, 2007). The moisture content and SC/MD ratio significantly affected the

594

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597

WGT (p < 0.001); these factors also exhibited a significant interaction effect (p < 0.05). No significant effect of NaCit was detected. The plots of WGT constructed according to the significant factors are given in Fig. 2. Note that, in this case, each data point represents an average of 216 measurements, 2 replicates  3 levels of NaCit/ SC  36 measurements at each NaCit/SC. The WGT is noticeably greater when the residual moisture content of samples is at low level. Also, the WGT tended to increase with increasing MD concentration; a slightly steeper change was found in low-moisture samples. The follow-up analyses showed that WGT of samples with different SC/MD were noticeably different for low-moisture samples. The WGT significantly increased from 15.5
C to 19.2
C, and to 24.1
C when SC/MD levels were 7:3, 5:5 and 3:7, respectively. At high level of residual moisture content, the WGT of samples with SC/MD of 7:3 (9.2
C) did not significantly differ from that of the samples with SC/MD of 5:5 (9.9
C), but differed from that of from the SC/MD of 3:7 samples (12.4
C). However, the WGT for samples with SC/MD of 5:5 and 3:7 were not significantly different. At every SC/MD level, the WGT obtained from the samples having different levels of residual moisture content significantly differed. As the residual moisture content increased from low to high level, the WGT decreased by approximately 50%, regardless of SC/MD level. In this case, the decrease in WGT should not be interpreted as the decrease of heterogeneity due to the increasing moisture content since there was no substantial chemical alteration of the glass matrix. Instead, the result suggests that water did accelerate the glass transition process. The decrease of WGT with the presence of water was also reported for amorphous amylopectin (Borde, Bizot, Vigier, & Buleon, 2002). There were other interesting cases where multi-component matrices exhibited a small WGT since some component(s) could accelerate the vitrification process (Kasapis, Al-Marhoobi, & Giannouli, 1999). This effect of water on WGT of carbohydrate systems is, however, opposite to that of the plasticizer commonly added to synthetic polymers (Bair, 1997). As could be gleaned from the overall analysis, the heterogeneity of SC-MD-salt bioglass was mainly influenced by MD. This is readily understandable since MD encompasses glucose polymers with variety of 30 Low MC level 25

High MC level

Width of GT zone (°C)

20

15

10

5

0 7:3

5:5

3:7

Sucrose:Maltodextrin (mass)

Fig. 2. The widths of glass-transition (GT) zone of bioglass samples at two levels of moisture content. Data are means of two replicates, and error bars reflect standard deviation.

chain lengths and molecular weights. The onset of translational motions or glass transition of these glucose polymers would occur at slightly different temperatures rendering a broad glass transition zone. A similar result was reported for amorphous mixtures containing SC, where the WGT was broader for SC mixtures than for SC alone. Accordingly, it was suggested that the midpoint-Tg represents an “average” value of Tg (Shamblin & Zografi, 1998). Even though the effect of NaCit on the heterogeneity of a bioglass was not discernable from DSC measurements, Ludescher and coworkers using a triplet erythrosine B probe revealed the effect of NaCl salt on dynamic site heterogeneity in SC glass with low residual moisture content, within approximately 2% (You & Ludescher, 2008). Erythrosine B phosphorescence is among the luminescence spectroscopic techniques that could provide direct information about molecular structure and mobility (Ludescher, Shah, McCaul, & Simon, 2001). Due to several advantages, e.g., high phosphorescence quantum yield, long life time, site selectivity, and high sensitivity (You & Ludescher, 2008), erythrosine B phosphorescence is extensively used to study the molecular motion in amorphous SC (Ludescher et al., 2001; Pravinata, You, & Ludescher, 2005; Shahs & Ludescher, 1995). Accordingly, there is a possibility that the effect of salt on the heterogeneity of SC-MD glass did exist but the sensitivity of DSC measurement was insufficient to detect. From the previous studies using a number of spectroscopic techniques, supercooled liquids and amorphous polymers were found to be dynamically heterogeneous through space and time (Ediger, 2000; Richert, 2002). The erythrosine B phosphorescence has provided evidence supporting the existence of spectral heterogeneity in amorphous sugars and alcohols (Pravinata et al., 2005; Shirke & Ludescher, 2005; Shirke, You, & Ludescher, 2006), and proteins (Lukasik & Ludescher, 2006; Nack & Ludescher, 2006; Sundaresan & Ludescher, 2008). This information suggests the dynamic site heterogeneity as a characteristic feature of amorphous biomaterials (You & Ludescher, 2008). 3.4. Change of specific heat capacity at glass transition During glass transition several physical properties undergo abrupt changes. As the temperature of a glassy material is raised through the glass transition region, its entropy, specific heat capacity (Cp), and diffusivity increase; while density, viscosity, and rigidity decrease (Champion et al., 2000). Of these, the change in specific heat capacity at glass transition (DCp,GT) can be readily observed from a DSC thermogram (Fig. 3). Gibbs and DiMarzio suggested that DCp,GT relates to the degree of organization in the glass matrix (Gibbs & DiMarzio, 1958). Accordingly, a glass at the physical state closer to being crystalline would exhibit greater DCp,GT, and hence the magnitude of DCp,GT should decrease with increasing molecular weight. In other words, the quality of hydrogen bonds among glass-forming molecules is manifested in terms of Cp (Avaltroni, Bouquerand, & Normand, 2004). The extent of DCp,GT also reflects the magnitude of changes in molecular mobility of amorphous matrix (Borde et al., 2002). Since glass transition is a kinetic process, the glass-forming condition could strongly affect the DCp,GT; a large standard deviation of the DCp,GT values may be expected (Orford, Parker, Ring, & Smith, 1989). Several studies on sugar glasses using FTIR spectroscopy reported that the magnitude of DCp,GT is associated with the strength of hydrogen bonding network in the glass matrix. Sugar glass with higher Tg has a weaker hydrogen bonding network in its structure (Wolkers, Oliver, Tablin, & Crowe, 2004b), and a smaller DCp,GT (Orford et al., 1990; Wolkers et al., 2004b) than the glass with lower Tg. The plots of DCp,GT at glass transition as a function of SC/MD for

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597

Fig. 3. Schematic representation of how DCp,GT is determined (distance between the onset Tg and offset Tg through the midpoint-Tg) from a DSC thermogram.

the systems with low and high levels of residual moisture content are shown in Fig. 4A and B, respectively. The DCp,GT values tended to decrease with the increasing concentration of MD, and so the increasing Tg, regardless of the residual moisture content level, as reported in the literature (Orford et al., 1990; Wolkers et al., 2004b). The ANOVA results show that all factors e moisture content, NaCit/ SC, and SC/MD e affected DCp,GT significantly (p < 0.001). Although the three-way interaction effect was not significant (p ¼ 0.091), all two-way interaction effects were significant (p < 0.05), which

595

made the interpretation of the results rather complicated. By considering the result from the ANOVA, the residual moisture content and SC/MD have likely had the most significant effect on the DCp,GT. Follow-up analyses and comparison tests were conducted to further elucidate these effects. In general, the DCp,GT value increased with NaCit concentration. The mean values of DCp,GT for bioglass systems with NaCit/SC of 0, 0.1, and 0.2 were approximately 0.55, 0.61 and 0.68 kJ/kg
C, respectively. The effect of NaCit on DCp,GT was significant only when the residual moisture content was high, where DCp increased at every increasing level of NaCit. The DCp,GT was positively dependent on the concentration of NaCit, but only when SC/MD levels were 7:3 or 5:5. Such dependence was absent in the system dominated by MD, i.e., when SC/MD ¼ 3:7. This result suggests the tendency of interaction between NaCit and SC, rather than MD, which then led to stronger hydrogen bonding network e reflected in the DCp,GT values e in system with high SC concentration as compared to MD. In most samples, the increasing residual moisture content significantly increased the DCp,GT, which are approximately 0.30 and 0.93 J/g
C for low- and high-moisture samples, respectively. Similar increase of DCp,GT with moisture content in amorphous amylopectin has also been reported (Borde et al., 2002). At every level of NaCit/SC, DCp,GT value of high-moisture systems was significantly greater than that of low-moisture systems. Further, across the range of SC/MD levels, the dependence of DCp,GT on residual moisture content was found only in systems with SC concentration equal to or higher than that of MD (i.e., SC/MD of 7:3 and 5:5). The DCp,GT values reported in the literature for SC is in the range of 0.57e0.65 J/g
C (Kawai, Hagiwara, Takai, & Suzuki, 2005; Liu et al., 2007; Mao, Chamarthy, & Pinal, 2006; Shamblin, Taylor, & Zografi, 1998), and for MD (DE19 to DE20) 0.44e0.45 J/g
C (Avaltroni et al., 2004; Roos & Karel, 1991). However, the value of DCp,GT for pure water is still a matter of debate (Borde et al., 2002): 0.092 J/g
C for liquid quenched water, 1.94 J/g
C for vapor deposited

Fig. 4. Effect of SC:MD ratio on the change of DCp,GT on samples equilibrated in the (A) low RH (~1%) and (B) high RH (~11%) atmospheres. Data are means of two replicates, and the error bars are standard deviations.

596

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597

water. In a mixture, the DCp,GT value of water was found in a range of 0.3e1.83 J/g
C. It appears that the value of DCp,GT for water is not much different from that for the glass-forming molecules in the SCMD system. Considering the small difference in average moisture contents between the low- and high-moisture systems, which were respectively around 0.31 and 3.64, and the large increase in DCp,GT mentioned above, the simple additivity laws may not be legitimate. Also, since water modulated the shape of DSC thermogram, specifically narrowing down the glass transition, the introduction of water to SC-MD system would be of a non-ideal type mixing, which suggests the existence of molecular interaction between water and carbohydrate molecules. A similar result was also reported (Borde et al., 2002). Since the moisture content of our samples was rather low and there were numerous hydroxyl groups on carbohydrate molecules, it was likely that water was bound to carbohydrates through hydrogen bonding. Such interference by water molecules would then increase the molecular mobility in the matrix and eventually affect the thermal properties. The mean value of DCp,GT for the samples with SC/MD of 7:3, 5:5, and 3:7 were 0.92, 0.70, and 0.24 J/g
C, respectively, suggesting a positive dependence of DCp,GT on SC concentration. That is, the DCp,GT decreased with the increasing MD concentration e the average molecular weight of glass-forming molecules was increased. This result is in good agreement with that reported in the literature (Avaltroni et al., 2004). At every level of decreasing SC concentration, the DCp,GT value decreased significantly, regardless of NaCit/SC level. The decrease of DCp,GT with decreasing SC concentration was also observed at both low and high levels of residual moisture content. 4. Conclusions Moisture contributed a significant plasticization effect on the SC-MD-NaCit bioglass system. A slight difference in residual moisture content, between low (0.27e0.35%wb) and high (2.83e4.40%wb) levels, resulted in substantial reduction of Tg from ~106 to 67
C. The Tg increased with increasing MD concentration. The average values of Tg for the systems with the SC/MD ratios of 7:3, 5:5 and 3:7 were 61, 64 and 76
C at high moisture content, and were 92, 101 and 126
C at low moisture content, respectively. When NaCit was added, the Tg value increased, and regardless of the SC/MD level, the Tg of high moisture systems increased slightly. These suggest a tendency of interaction between NaCit and SC, rather than between NaCit and MD and hence the creation of large, less-mobile large clusters of SC and NaCit. The WGT was affected by the SC/MD ratio and residual moisture content. For systems with different SC/MD, water accelerated the glass transition so that the WGT narrowed by approximately 50% as moisture content increased from low to high level. In general, DCp,GT tended to increase with the increasing amounts of moisture, SC and NaCit. The increase of DCp,GT with SC and NaCit concentrations suggests a greater degree of organization of the glass matrix, i.e., a more compact system. The increase of DCp,GT with moisture content suggests the interference of water on the intermolecular interactions among the glass-forming molecules. References Allen, G. (1993). A history of glassy state. In J. M. V. Blanshard, & P. J. Lillford (Eds.), The glassy state in foods (pp. 1e12). Loughborough: Nottingham University Press. Angell, C. A. (1997). Why C1 ¼ 16-17 in the WLF equation is physicaleand the fragility of polymers. Polymer, 38(26), 6261e6266. Avaltroni, F., Bouquerand, P. E., & Normand, V. (2004). Maltodextrin molecular

weight distribution influence on the glass transition temperature and viscosity in aqueous solutions. Carbohydrate Polymers, 58(3), 323e334. Bair, H. E. (1997). Thermal analysis of additives in polymers. In E. A. Turi (Ed.), Thermal characterization of polymeric materials (pp. 205e482). San Diego: Academic Press. Bhandari, B. R., & Hartel, R. W. (2005). Phase transitions during food powder production and powder stability. In C. Onwulata, & R. P. Konstance (Eds.), Encapsulated and food powder (pp. 261e292). New York: Marcel Dekker. Bhandari, B. R., & Howes, T. (1999). Implication of glass transition for the drying and stability of dried foods. Journal of Food Engineering, 40(1e2), 71e79. Borde, B., Bizot, H., Vigier, G., & Buleon, A. (2002). Calorimetric analysis of the structural relaxation in partially hydrated amorphous polysaccharides. I. Glass transition and fragility. Carbohydrate Polymers, 48(1), 83e96. Buera, P., Schebor, C., & Elizalde, B. (2005). Effects of carbohydrate crystallization on stability of dehydrated foods and ingredient formulations. Journal of Food Engineering, 67(1e2), 157e165. Buitink, J., & Leprince, O. (2004). Glass formation in plant anhydrobiotes: survival in the dry state. Cryobiology, 48(3), 215e228. Champion, D., Le Meste, M., & Simatos, D. (2000). Towards an improved understanding of glass transition and relaxations in foods: molecular mobility in the glass transition range. Trends in Food Science & Technology, 11(2), 41e55. Chronakis, I. S. (1998). On the molecular characteristics, compositional properties, and structural-functional mechanisms of maltodextrins: a review. Critical Reviews in Food Science and Nutrition, 38(7), 599e637. Crowe, J. H., Crowe, L. M., Oliver, A. E., Tsvetkova, N., Wolkers, W., & Tablin, F. (2001). The trehalose myth revisited: introduction to a symposium on stabilization of cells in the dry state. Cryobiology, 43(2), 89e105. Davidson, P., & Sun, W. Q. (2001). Effect of sucrose/raffinose mass ratios on the stability of co-lyophilized protein during storage above the Tg. Pharmaceutical Research, 18(4), 474e479. Desobry, S. A., Netto, F. M., & Labuza, T. P. (1999). Influence of maltodextrin systems at an equivalent 25DE on encapsulated b-carotene loss during storage. Journal of Food Processing and Preservation, 23(1), 39e55. Ediger, M. D. (2000). Spatially heterogeneous dynamics in supercooled liquids. Annual Review of Physical Chemistry, 51, 99e128. Elliott, S. R., Rao, C. N. R., & Thomas, J. M. (1986). The chemistry of the noncrystalline state. Angewandte Chemie International Edition in English, 25(1), 31e46. Farahnaky, A., Farhat, I. A., Mitchell, J. R., & Hill, S. E. (2009). The effect of sodium chloride on the glass transition of potato and cassava starches at low moisture contents. Food Hydrocolloids, 23(6), 1483e1487. Ferry, J. D. (1980). Viscoelastic properties of polymers. New York: John Wiley & Sons. Galmarini, M. V., Zamora, M. C., Baby, R., Chirife, J., & Mesina, V. (2008). Aromatic profiles of spray-dried encapsulated orange flavours: influence of matrix composition on the aroma retention evaluated by sensory analysis and electronic nose techniques. International Journal of Food Science & Technology, 43(9), 1569e1576. Gibbs, J. H., & DiMarzio. (1958). Nature of glass transition and the glassy state. The Journal of Chemical Physics, 28, 373e383. Hancock, B. C., & Zografi, G. (1997). Characteristics and significance of the amorphous state in pharmaceutical systems. Journal of Pharmaceutical Sciences, 86(1), 1e12. Hoffmann, H., & Felgner, A. (2011). Determination of water content in candies. In AnalytiX (Sigma-Aldrich), 10. Izutsu, K.-i., & Aoyagi, N. (2005). Effect of inorganic salts on crystallization of poly(ethylene glycol) in frozen solutions. International Journal of Pharmaceutics, 288(1), 101e108. Izutsu, K.-I., Fujimaki, Y., Kuwabara, A., & Aoyagi, N. (2005). Effect of counterions on the physical properties of l-arginine in frozen solutions and freeze-dried solids. International Journal of Pharmaceutics, 301(1e2), 161e169. Jiang, B., Liu, Y., Bhandari, B., & Zhou, W. (2008). Impact of caramelization on the glass transition temperature of several caramelized sugars. Part I: chemical analyses. Journal of Agricultural and Food Chemistry, 56(13), 5138e5147. Kasapis, S., Al-Marhoobi, I. M. A., & Giannouli, P. (1999). Molecular order versus vitrification in high-sugar blends of gelatin and k-carrageenan. Journal of Agricultural and Food Chemistry, 47(12), 4944e4949. Kawai, K., Hagiwara, T., Takai, R., & Suzuki, T. (2005). Comparative investigation by two analytical approaches of enthalpy relaxation for glassy glucose, sucrose, maltose, and trehalose. Pharmaceutical Research, 22(3), 490e495. Kets, E. P. W., Ijpelaar, P. J., Hoekstra, F. A., & Vromans, H. (2004). Citrate increases glass transition temperature of vitrified sucrose preparations. Cryobiology, 48(1), 46e54. Kilburn, D., Claude, J., Mezzenga, R., Dlubek, G., Alam, A., & Ubbink, J. (2004). Water in glassy carbohydrates: opening it up at the nanolevel. The Journal of Physical Chemistry B, 108(33), 12436e12441. Koster, K., Maddocks, K., & Bryant, G. (2003). Exclusion of maltodextrins from phosphatidylcholine multilayers during dehydration: effects on membrane phase behaviour. European Biophysics Journal, 32(2), 96e105. Le Meste, M., Champion, D., Roudaut, G., Blond, G., & Simatos, D. (2002). Glass transition and food technology: a critical appraisal. Journal of Food Science, 67(7), 2444e2458. Liu, Y., Bhandari, B., & Zhou, W. (2006). Glass transition and enthalpy relaxation of amorphous food saccharides: a review. Journal of Agricultural and Food Chemistry, 54(16), 5701e5717. Liu, Y., Bhandari, B., & Zhou, W. (2007). Study of glass transition and enthalpy relaxation of mixtures of amorphous sucrose and amorphous tapioca starch

E. Sritham, S. Gunasekaran / Food Hydrocolloids 60 (2016) 589e597 syrup solid by differential scanning calorimetry (DSC). Journal of Food Engineering, 81(3), 599e610. Longinotti, M. P., Mazzobre, M. F., Buera, M. P., & Corti, H. R. (2002). Effect of salts on the properties of aqueous sugar systems in relation to biomaterial stabilization Part 2. Sugar crystallization rate and electrical conductivity behavior. Physical Chemistry Chemical Physics, 4, 533e540. Ludescher, R. D., Shah, N. K., McCaul, C. P., & Simon, K. V. (2001). Beyond Tg: optical luminescence measurements of molecular mobility in amorphous solid foods. Food Hydrocolloids, 15(4e6), 331e339. Lukasik, K. V., & Ludescher, R. D. (2006). Effect of plasticizer on dynamic site heterogeneity in cold-cast gelatin films. Food Hydrocolloids, 20(1), 88e95. Mao, C., Chamarthy, S. P., & Pinal, R. (2006). Time-dependence of molecular mobility during structural relaxation and its impact on organic amorphous solids: an investigation based on a calorimetric approach. Pharmaceutical Research, 23(8), 1906e1917. McPherson, A. E., & Seib, P. A. (1997). Preparation and properties of wheat and corn starch maltodextrins with a low dextrose equivalent. Cereal Chemistry, 74(4), 424e430. Montgomery, D. C. (2005). Design and analysis of experiments (6th ed.). New Jersey: John Wiley & Sons, Inc. Nack, T. J., & Ludescher, R. D. (2006). Molecular mobility and oxygen permeability in amorphous bovine serum albumin films. Food Biophysics, 1(3), 151e162. Oldenhof, H., Wolkers, W. F., Fonseca, F., Passot, S., & Marin, M. (2005). Effect of sucrose and maltodextrin on the physical properties and survival of air-dried Lactobacillus bulgaricus: an in situ Fourier transform infrared spectroscopy study. Biotechnology Progress, 21, 885e892. Orford, P. D., Parker, R., & Ring, S. G. (1990). Aspects of the glass transition behaviour of mixtures of carbohydrates of low molecular weight. Carbohydrate Research, 196, 11e18. Orford, P. D., Parker, R., Ring, S. G., & Smith, A. C. (1989). Effect of water as a diluent on the glass transition behavior of malto-oligosaccharides, amylose and amylopectin. International Journal of Biological Macromolecules, 11, 91e96. Poirier-Brulez, F., Roudaut, G., Champion, D., Tanguy, M., & Simatos, D. (2006). Influence of sucrose and water content on molecular mobility in starch-based glasses as assessed through structure and secondary relaxation. Biopolymers, 81(2), 63e73. Pravinata, L. C., You, Y., & Ludescher, R. D. (2005). Erythrosin B phosphorescence monitors molecular mobility and dynamic site heterogeneity in amorphous sucrose. Biophysical Journal, 88(5), 3551e3561. Richert, R. (2002). Heterogeneous dynamics in liquids: fluctuations in space and time. Journal of Physics: Condensed Matter, 14(23), R703eR738. Roos, Y. (1995). Phase transitions in foods. San Diego: Academic Press. Roos, Y., & Karel, M. (1991). Water and molecular weight effects on glass transitions in amorphous carbohydrates and carbohydrate solutions. Journal of Food

597

Science, 56(6), 1676e1681. Roudaut, G., Simatos, D., Champion, D., Contreras-Lopez, E., & Le Meste, M. (2004). Molecular mobility around the glass transition temperature: a mini review. Innovative Food Science & Emerging Technologies, 5(2), 127e134. Shahs, N. K., & Ludescher, R. D. (1995). Phosphorescence probes of the glassy state in amorphous sucrose. Biotechnology Progress, 11(5), 540e544. Shamblin, S. L., Taylor, L. S., & Zografi, G. (1998). Mixing behavior of colyophilized binary systems. Journal of Pharmaceutical Sciences, 87(6), 694e701. Shamblin, S. L., & Zografi, G. (1998). Enthalpy relaxation in binary amorphous mixtures containing sucrose. Pharmaceutical Research, 15(12), 1828e1834. Shirke, S., & Ludescher, R. D. (2005). Dynamic site heterogeneity in amorphous maltose and maltitol from spectral heterogeneity in erythrosin B phosphorescence. Carbohydrate Research, 340, 2661e2669. Shirke, S., You, Y., & Ludescher, R. D. (2006). Molecular mobility and dynamic site heterogeneity in amorphous lactose and lactitol from erythrosin B phosphorescence. Biophysical Chemistry, 123(2e3), 122e133. Sundaresan, K. V., & Ludescher, R. D. (2008). Molecular mobility and oxygen permeability in amorphous b-lactoglobulin films. Food Hydrocolloids, 22(3), 403e413. Surana, R., Pyne, A., Rani, M., & Suryanarayanan, R. (2005). Measurement of enthalpic relaxation by differential scanning calorimetry e effect of experimental conditions. Thermochimica Acta, 433(1e2), 173e182. Taylor, L. S., & Zografi, G. (1998). Sugar-polymer hydrogen bond interactions in lyophilized amorphous mixtures. Journal of Pharmaceutical Sciences, 87(12), 1615e1621. Tong, P., Taylor, L. S., & Zografi, G. (2002). Influence of alkali metal counterions on the glass transition temperature of amorphous indomethacin salts. Pharmaceutical Research, 19(5), 649e654. Wolkers, W. F., Oliver, A. E., Tablin, F., & Crowe, J. H. (2004b). A Fourier-transform infrared spectroscopy study of sugar glasses. Carbohydrate Research, 339(6), 1077e1085. Wolkers, W. F., van Kilsdonk, M. G., & Hoekstra, F. A. (1998). Dehydration-induced conformational changes of poly-l-lysine as influenced by drying rate and carbohydrates. Biochimica et Biophysica Acta (BBA) e General Subjects, 1425(1), 127e136. Wungtanagorn, R., & Schmidt, S. J. (2001). Thermodynamic properties and kinetics of the physical ageing of amorphous glucose, fructose, and their mixture. Journal of Thermal Analysis and Calorimetry, 65(1), 9e35. You, Y., & Ludescher, R. D. (2008). The effect of sodium chloride on molecular mobility in amorphous sucrose detected by phosphorescence from the triplet probe erythrosin B. Carbohydrate Research, 343(2), 350e363. Young, R. J., & Lovell, P. A. (2011). Introduction to polymers (3 ed.). Boca Raton, FL: CRC Press.