Nonenzymatic Browning Kinetics in Low-Moisture Food Systems As Affected by Matrix Composition and Crystallization SONG MIAO and YRJO H

JFS E: Food Engineering and Physical Properties

Nonenzymatic Browning Kinetics in Low-moisture Food Systems as Affected by Matrix Composition and Crystallization SONG MIAO AND YRJO H. ROOS

ABSTRACT: Nonenzymatic browning (NEB) in lactose, trehalose, and lactose/trehalose food model systems containing L-lysine and D-xylose (5% w/w) as reactants was studied at 4 different relative vapor pressure (RVP) (33.2%, 44.1%, 54.5%, 65.6%) environments at room temperature. Sorption isotherms of model systems were determined gravimetrically, and data were modeled using the Brunauer-Emmett-Teller (BET) and Guggenheim-Anderson-deBoer

(GAB) models. Glass transition, Tg, was measured by differential scanning calorimetry (DSC). NEB was followed spectrophotometrically. Although the 3 model systems showed similar glass transition behavior, different crystallization properties were observed from loss of sorbed water. Mixtures of trehalose and lactose showed delayed crystallization of component sugars. The NEB rate was affected by sugar composition. At a low RVP (33.1%) environment, the NEB rate in trehalose system was higher than in the lactose/trehalose system, and the NEB rate in lactose system was the lowest. The NEB rate in different models seemed to be affected by component crystallization. The highest extent of browning in the trehalose matrix system seemed to be related to the formation of trehalose crystals in the system after crystallization at high RVP. The results indicated that the composition of the carbohydrate-based low- moisture real food systems should be considered in controlling NEB reaction. Keywords: crystallization, lactose, nonenzymatic browning, water content, trehalose

Introduction Bell 1996; Bell and others 1998; Lievonen and others 1998; Schebor tability of food and other biological systems containing sugars and others 1999; Lievonen and Roos 2002b). The kinetics of NEB Sand/or polymers is often related to the properties of their amor- have been studied as a function of many parameters (water activ- phous phases (O’Brien 1996). The physical state of amorphous ity, temperature, reactant concentration pH, and so forth) and even matrix greatly influences their stability and affects both physical related to physical aspects such as crystallization (Saltmarch and and chemical changes during processing and storage (Eichner others 1981; Vuataz 1988; Buera and Karel 1995) and viscosity 1981; Slade and Levine 1991; Karmas and others 1992; Buera and (Eichner 1981). The glass transition together with material compo- E: Food Engineering & Physical Properties Karel 1995). The glass transition controls the physical state and sition, water activity, temperature, and other factors affect the rate physical characteristics of amorphous food materials and has been of the reaction (Bell 1996; Roos and others 1996b; Lievonen and related to their chemical stability (Sperling 1986; Roos and Karel others 1998, 2002; White and Bell 1999; Lievonen and Roos 2002b; 1991; Slade and Levine 1991; Roos 1995; Roos and others 1996b; Miao and Roos 2004b). Champion and others 2000). The glass transition has been recog- Time-dependent crystallization of amorphous carbohydrates is nized as a possible factor affecting kinetics of enzymatic and non- known to result in serious adverse quality losses in dairy food pow- enzymatic changes in low-moisture foods (Slade and Levine 1991; ders during storage (Saltmarch and others 1981; Roos and Karel Karmas and others 1992; Karmas and Karel 1994; Schebor and oth- 1992; Roos and others 1996b; Biliaderis and others 2002). Crystal- ers 1999; Kouassi and Roos 2000, 2001). lization of component sugars may enhance the rate of an enzymatic Nonenzymatic browning (NEB) through the Maillard reaction is reaction as a result of release of sorbed water during the crystalliza- a major deteriorative factor in the storage of dehydrated dairy food tion process (Cardona and others 1997; Kouassi and Roos 2000). products (Labuza and Saltmarch 1981; Friedman 1996; Roos and The crystallization of amorphous lactose in dehydrated milk prod- others 1996b; van Boekel 2001). NEB may cause unacceptable ucts has also been observed to result in acceleration of NEB reac- nutritional and sensory effects in some stored food products and tions, loss of lysine as well as other deteriorative changes and cak- may be a limiting factor in the shelf life of some products (O’Brien ing (Labuza and Saltmarch 1981; Saltmarch and others 1981; Roos and Morrissey 1989; O’Brien 1996). NEB as a model of a possible 2001). The rate of crystallization depends on the composition of diffusion-controlled binary reaction (Buera and Karel 1995) be- food materials (Arvanitoyannis and Blandshard 1996; Shamblin tween an amino acid and a reducing sugar has probably been given and others 1996; Biliaderis and others 2002). Addition of a 2nd poly- the most attention in studies of relationships between reaction rates meric additive, such as maltodextrin, to an amorphous matrix de- and the physical state controlled by the glass transition (Karmas layed crystallization (Roos and Karel 1991; O’Brien 1996; Shamblin and others 1992; Karmas and Karel 1994; Roos and Himberg 1994; and others 1996; Gabarra and Hartel 1998; Biliaderis and others 2002); but in the food industry, because of quality and sensorial MS 20040489 Submitted 6/22/04, Revised 9/13/04, Accepted 9/24/04. Authors considerations, addition of polymeric materials is often replaced Miao and Roos are with Dept. of Food and Nutritional Sciences, Univ. Col- by addition of a material with similar characteristics, such as sugars lege Cork, Ireland. Direct inquiries to author Roos (E-mail: [email protected]). (Mazzobre and others 2001).

© 2005 Institute of Food Technologists Vol. 70, Nr. 2, 2005—JOURNAL OF FOOD SCIENCE E69 Further reproduction without permission is prohibited Published on Web 2/7/2005 NEB kinetics in food systems . . .

Trehalose has been reported to be the most effective cryopreser- samples were subsequently kept at 22 °C to 23 °C over saturated

vation sugar for stabilizing proteins and membranes against dam- solutions of LiCl, CH3COOK, MgCl2, K2CO3, Mg(NO3)2, NaNO2, and age caused by desiccation or freezing (Crowe and others 1984; Leslie NaCl with relative vapor pressures (RVPs) of 11.4, 23.1, 33.2, 44.1, and others 1995; Uritani and others 1995). Several studies have 54.5, 65.6, and 76.1%, respectively (Greenspan 1977; Labuza and reported stabilization of enzymes by incorporation of trehalose in others 1985) until the sample weights leveled off, indicating matrices (Cardona and others 1997; Mazzobre and others 1997a, steady-state water contents. The samples were weighed at 3, 6, 9, 1997b; Schebor and others 1997). As a nonreducing sugar, treha- 11, 21, 24 h, and then at 24 h intervals for RVPs Յ 44.1%; samples lose has also been used as an ingredient in the food industry (Roser were weighed every hour up to 6 h, then at 8, 10, 21, 24 h, and then 1991; Karmas 1994). Dairy-based powdered flavors can also be at 24-h intervals during storage for RVPs Ն 54.5%. All vials were improved by the addition of trehalose. The nonhygroscopic nature kept closed with caps after the vacuum was released in the desicca- of trehalose decreases moisture sensitivity and product clumping tors before weighing. while imparting a mild sweet rounded flavor. The unique glass- The Brunauer-Emmett-Teller (BET) (Brunauer and others 1938; forming properties combined with mild sweetness make it an ex- Roos 1993) and Guggenheim-Anderson-deBoer (GAB) (van den cellent carrier for flavors or other spray-dried product components Berg and Bruin 1981; Roos 1993) sorption isotherm models were (Roser 1991; Schiraldi and others 2002). In dairy products, lactose fitted into water sorption data using linear regression analysis and is a very important component dominating the physical and chem- quadratic analysis, as described by Jouppila and Roos (1994). The ical properties of dairy powders. The physical state of lactose is BET model is described as Eq. 1. important in maintaining food quality (Simatos and Karel 1988; Hartel and Shastry 1991; Roos and Karel 1992). Crystallization of lactose has been reported to be delayed by trehalose (Mazzobre (1) and others 2001). We presumed the physical properties of sugar- containing dairy products were affected by adding trehalose as a new ingredient, and thus the chemical stability might also be af- where mm is the BET monolayer value (g water/100 g solids), and K fected during storage. is a constant. Equation 1 can be simplified to a linear form as shown The objective of the present work was to study NEB kinetics in by Eq. 2: E: Food Engineering & Physical Properties lactose, trehalose, and lactose/trehalose matrices as affected by glass transition and water content changes, and to evaluate the (2) effects of matrix composition and component crystallization on NEB behavior. The GAB isotherm model is described by Eq. 3: Materials and Methods

Preparation of food models (3) Three amorphous food model systems were prepared. The models consisted of lactose (Sigma Chemical Co., St. Louis, Mo., U.S.A.), The GAB model can also be written to the form of the second-order trehalose (British Sugar Co., Peterborough, U.K.), and lactose/tre- polynomial: halose (1:1) as matrix materials, and they contained L-lysine and D-xylose (Sigma Chemical Co.) 1:1 as reactants. According to the results of previous studies (Miao and Roos 2004a) and pre-experi- (4) ments carried out in the present study, the exact amounts of the reactants were adjusted to 5% (w/w) of the total solids. Lactose, from which the constants C, KЈ and m can be calculated as de- lactose/trehalose (1:1), and trehalose were dissolved in distilled m scribed by Roos (1995). water under mild heating to obtain clear solutions. The reactants The relative percentage Root Mean Square (RMS-value) was were then added after cooling to about 20 °C and mixed (cooling calculated as an indication of the fit of the BET and GAB models was essential to minimize browning during mixing), final 20% (w/w (Bizot 1983; Lievonen and Roos 2002a). solids) clear solutions of the matrix materials, the reactants and distilled water were thus prepared. Solutions in 20- to 22-mL aliquots were frozen on Petri dishes immediately after preparation (24 h at –20 °C, followed by 24 h at –80 °C) and freeze-dried (>48 h, P < 0.1 mbar) (Lyovac GT 2; Amsco Finn-Aqua Gmbh, Hürth, Ger- (5) many) to produce amorphous glassy food models (Roos and Karel 1991). The freeze-dried materials were ground using a mortar and pestle immediately after freeze-drying. Ground powders were sub- where Wi is the experimental water content; Wi* is the calculated sequently stored in evacuated desiccators over P2O5 for a period of no less than 1 wk. These samples were taken as “zero” moisture con- water content; and N is the number of experimental points. tent (Buera and others 1992) before experiments. Differential scanning calorimetry Water sorption The glass transition temperatures for the 3 different models Sorption isotherms for the 3 food models were determined gravi- stored at different RVPs were determined using Differential Scan- metrically. Triplicate samples of 1 g of the freeze-dried food mod- ning Calorimetry (DSC) (Mettler Toledo 821e system with 821e TA e els, prepared in 20-mL glass vials, were stored in a vacuum desicca- processor, DCS 821e measuring cell, and STAR Thermal Analysis System version 6.0 software; Mettler-Toledo AG, Schwerzenbach, tor over P2O5 for 1 wk. After storage, the samples were considered “anhydrous” (Lievonen and Roos 1998). The dehydrated triplicate Switzerland). The instrument was calibrated for temperature and

E70 JOURNAL OF FOOD SCIENCE—Vol. 70, Nr. 2, 2005 URLs and E-mail addresses are active links at www.ift.org NEB kinetics in food systems . . . heat flow as reported by Miao and Roos (2004a). Samples (9 to 15 the coefficients of determination (R2) were calculated from zero- mg) were prepared in preweighed DSC aluminum pans (40 ␮L; order kinetics using a linear regression analysis (Labuza and Baisi- Mettler Toledo-2731, Switzerland) and stored in open pans for 120 er 1992). h at room temperature in evacuated desiccators over different saturated salt solutions: LiCl, CH3COOK, MgCl2, K2CO3 (Sigma Chem- Results and Discussion ical Co.) with respective RVPs of 11.4%, 23.1%, 33.2%, and 44.1%, giving a water activity, aw, of 0.01 × RVP at equilibrium (Labuza and Water sorption others 1985). After equilibration, the pans were hermetically sealed. The freeze-dried lactose/reactant, lactose/trehalose/reactant, Triplicate samples of each material were analyzed. An empty pan and trehalose/reactant model systems were very hygroscopic. was used as a reference. The samples were scanned 1st at 5 °C/min When exposed to the RVPs Յ 44.1% environments, steady-state over the glass transition range, then cooled at 10 °C/min to at least water contents were reached after 24 h storage at room temperature 30 °C below the glass transition, and an immediate 2nd heating (22 °C to 23 °C), as showed in Figure 1. When models were exposed scan at 5 °C/min was run to at least 30 °C above the glass transition to RVPs Ն 54.5%, the 3 systems showed different sorption behav- for each sample to eliminate relaxations and improve interpretation ior (Figure 2). Increasing sorption of water followed by a decrease in Ն of the thermograms. The glass transition temperature, Tg, was tak- sorbed water content was observed in 3 systems at RVPs 65.6%. en as the onset temperature of the glass transition temperature The loss of sorbed water at higher RVPs resulted from component range. An average of 3 replicate samples was used as the glass crystallization (Berlin and others 1968; Vuataz 1988; Lai and transition temperature. Schmidt 1990; Roos and Karel 1992; Jouppila and Roos 1994; Haque The Gordon-Taylor (Gordon and Taylor 1952) Eq. 6 was fitted to and Roos 2004). Compared with the lactose/reactant and trehalose/ the experimental glass transition temperature data: reactant systems, crystallization of component sugars in the lactose/trehalose/reactant system occurred more slowly (Mazzobre and others 2001; Miao and Roos 2004a). At 54.5% RVP, lactose/tre- (6) halose/reactant system did not show decreasing water contents within 250 h, whereas loss of sorbed water in the lactose/reactant system and trehalose/reactant system was observed after 48 h and where w1 and w2 are the weight fractions of the solute and water 43 h, respectively. As showed in Figure 2, at RVP 65.6% and 76.1%, respectively, Tg1 is the Tg of anhydrous solute, Tg2 is the Tg of amor- loss of sorbed water in the lactose/reactant system was observed phous water, –135 °C (Johari and others 1987) and k is a constant. after 21 h and 8 h of storage, in the trehalose/reactant system after The value for k was the mean derived from the experimental Tg and 5 h and 3 h, respectively, and in the lactose/trehalose/reactant water content data as described by Jouppila and Roos (1994). Equa- system, after 79 and 33 h, respectively. The loss of water during tion 9 was applied in the present study for predicting water plasti- sorption study was used by several researchers as the indicator for cization of the multicomponent model systems, considering that all studying the lactose crystallization in different systems (Berlin and solid components contributing to the observed glass transition others 1968; Linko and others 1982; Vuataz 1988; Lai and Schmidt were miscible and formed a single phase (Roos 1995). Because sub- 1990; Roos and Karel 1992; Jouppila and Roos 1994; Saleki-Gerhardt stantial browning occurred at high temperatures above 100 °C, and Zografi 1994; Haque and Roos 2004). Although the main com- water produced in the reaction plasticized the samples (Roos and ponent of the lactose/reactant system is lactose (90%), the sorption others 1996b). Hence, no reliable glass transition temperature (Tg) results in the present study showed that time-dependent crystal- could be measured for the anhydrous model systems. We assumed

lization of lactose in the lactose/reactant system was delayed com- E: Food Engineering & Physical Properties that the low concentration of reactants only slightly affected the Tg, pared with the pure lactose reported by other authors (Haque and and we used the Tg’s of the lactose, lactose/trehalose, and trehalose Roos 2004; Jouppila and Roos 1994; Miao and Roos 2004a). In our systems as the Tg’s for the anhydrous lactose/reactant, lactose/tre- previous study (Miao and Roos 2004a), crystallization of freeze- halose/reactant, and trehalose/reactant systems. dried pure lactose was observed from the loss of sorbed water after 3 h, 4 h, and 11 h of storage at 76.1%, 65.6%, and 54.5% RVPs, re- Nonenzymatic browning spectively. Jouppila and Roos (1994) reported pure lactose crystal- Aliquots of 1 g of the powdered freeze-dried model materials lized within 24 h at RVPs > 40%, whereas Haque and Roos (2004) were transferred into 20-mL glass vials, and the glass vials were observed that the loss of the sorbed water in spray-dried lactose stored in the desiccators at 4 RVP values ranging from 33% to 65% powder appeared after 2 h, 3 h, and 20 h of storage at 76.1%, 65.6%, at room temperature (22 °C to 23 °C). Four saturated salt solutions and 54.5% RVPs, respectively. This was probably because of the (MgCl2, K2CO3, Mg(NO3)2, and NaNO2) were used to achieve RVP addition of the reactants and the difference in materials and sorp- values of 33.2%, 44.1%, 54.5%, and 65.6%, respectively (Greenspan tion conditions. Iglesias and others (1997) studied the sorption iso- 1977; Labuza and others 1985). Triplicate samples were removed at therm of amorphous trehalose; time-dependent trehalose crystal- 10-h to 24-h intervals depending on the RVP, and subsequently lization observed in their study was similar to the present study for stored at –80 °C to avoid further browning before analysis. The ex- the trehalose/reactant system. In the present study, 3 model sys- tent of browning was determined spectrophotometrically (Perkin- tems sorbed different amounts of water before showing the loss of Elmer Lambda 2 UV-VIS spectrometer; Norwalk, Conn., U.S.A.) water when exposed to high RVP environments, especially at 65.6% from the optical density (OD) at 280 nm, which is considered to in- and 76.1% RVPs. The lactose/trehalose/reactant system sorbed 19.3 dicate formation of furfural compounds (Hodge and Osman 1976; and 25.2 g water/100 g dry solids, respectively; the trehalose/reac- Resnik and Chirife 1979) and is used to detect products of early tant system sorbed 14.4 and 16.8 g water/100 g solids; and the lac- stage of browning (Flink 1983), and at 420 nm for yellow and brown tose/reactant system sorbed 12.6 and 13.1 g/100 g solids before the pigments (Whistler and Daniel 1985). Samples were dissolved with loss of water occurred at 65.6% and 76.1% RVPs. This higher water 40 mL of distilled water at 5 °C to avoid further browning and dilut- content in the lactose/trehalose/reactant system probably related ed when necessary to obtain reliable absorbance readings. Rate to the interaction between the lactose and trehalose in the mix. In constants of the browning, k, and their 95% confidence limits, and fact, the lactose/trehalose/reactant system also took longer time to

URLs and E-mail addresses are active links at www.ift.org Vol. 70, Nr. 2, 2005—JOURNAL OF FOOD SCIENCE E71 NEB kinetics in food systems . . .

release the sorbed water during crystallization. At RVPs Ն 65.6%, the trehalose/reactant model. These water contents for trehalose/ loss of water after initial crystallization in 3 model systems leveled reactant systems in the present study agreed with the results of off after 150 h of storage (Figure 2b and 2c), crystallization of com- Karmas (1994), Iglesias and others (1997), and Cardona and others ponent sugars in the model systems was considered to proceed only (1997). The water contents in the lactose/reactant system after to a given extent. crystallization ranged from 1.7 to 5.7 g/100 g of dry solids depend- The experimental sorbed water contents of the 3 model systems ing on RVP, which were slightly higher than those for pure lactose stored for 144 h at various RVPs at room temperature are given in reported in previous studies (Jouppila and Roos 1994, Miao and Table 1. At lower RVPs (Յ 44.1%), leveled-off water contents in the Roos 2004a), indicating that reactant in the system might have 3 systems were close to each other. At higher RVPs (Ն 54.5%), after sorbed more water after lactose crystallization or there could be in- crystallization of components, 3 model systems had different wa- complete crystallization of lactose in presence of reactants. ter contents. The effect of crystallization on water content in a system containing crystalline and amorphous components was depen- Sorption isotherm dent on the type of crystals (Karmas 1994; Roos 1995; Cardona and The BET and the GAB sorption isotherms could be used to model

others 1997). Formation of anhydrous crystals (such as lactose and experimental sorption data of the 3 model systems over the aw sucrose) released water (Iglesias and Chirife 1978; Roos 1995), but range from 0.114 to 0.441, as shown in Figure 3. The isotherms had during formation of hydrated crystals (such as monohydrate lactose the sigmoid shape typical of most food materials (Biliaderis and and dihydrate trehalose), water was sorbed (Karmas 1994; Iglesias others 1999; Bell and Labuza 2000). The BET and GAB monolayer and others 1997). The amount of water in the trehalose dihydrate values and constants for lactose/reactant, lactose/trehalose/reac- was 10.5% (dry trehalose basis). This was close to the water content tant, and trehalose/reactant systems are given in Table 2. The GAB corresponding to the plateau in the sorption isotherm (Figure 2) of model showed better fit to the sorption data than the BET model, E: Food Engineering & Physical Properties

Figure 1—Water content for freeze-dried lactose/reactant (a), lactose/trehalose/reactant (b), and trehalose/reactant Figure 2—Water content for freeze-dried lactose/trehalose/ (c) systems as a function of storage time under various reactant, lactose/reactant, trehalose/reactant as function relative vapor pressures (RVPs) Յ 44.1% (× = 44.1%RVP; of storage time under various relative vapor pressures ᭝ = 33.2%RVP; ᭿ = 23.1%RVP; ᭛ = 11.4% RVP) at room (RVPs) (54.5%, 65.6%, and 76.1%) at room temperature temperature (22 °C to 23 °C). (22 °C to 23 °C).

E72 JOURNAL OF FOOD SCIENCE—Vol. 70, Nr. 2, 2005 URLs and E-mail addresses are active links at www.ift.org NEB kinetics in food systems . . .

Table 1—Water contents (mean ± SD) of model systems Table 2—Monolayer water contents (g/100 g solids), stored for 144 h at various relative vapor pressure (RVP) Brunauer-Emmett-Teller (BET) constant K, Guggenheim- at room temperature (22 °C to 23 °C) (n = 3, SD = standard Anderson-deBoer (GAB) constant KЈ and C, and R2 for lac- deviation)a tose/reactant, lactose/trehalose/reactant, and trehalose/ reactant model systemsa Water content (g/100 g dry solids) BET RVP (% 2 Salt of water) L/R L/T/R T/R mm KR%RMS LiCl 11.4 2.89 ± 0.08 3.31 ± 0.09 3.26 ± 0.15 Lactose/reactant 6.57 4.783 0.990 1.471 CH3COOK 23.1 5.03 ± 0.03 4.96 ± 0.14 5.37 ± 0.12 Lactose/trehalose/reactant 6.89 4.939 0.937 3.121 MgCl2 33.2 6.71 ± 0.03 7.04 ± 0.02 6.88 ± 0.06 Trehalose/reactant 6.68 5.480 0.965 0.804 K CO 44.1 9.48 ± 0.11 10.24 ± 0.17 10.10 ± 0.08 2 3 GAB Mg(NO3)2 54.5 3.57 ± 0.16 13.00 ± 0.08 9.77 ± 0.03 2 NaNO2 65.6 4.29 ± 0.05 9.61 ± 0.11 9.46 ± 0.15 mm K’ CK’ %RMS NaCl 76.1 5.97 ± 0.07 10.58 ± 0.04 10.22 ± 0.11 Lactose/reactant 5.20 1.184 5.969 0.981 0.633 aL/R = lactose/reactant; L/T/R = lactose/trehalose/reactant; T/R = trehalose/ Lactose/trehalose/reactant 3.95 1.439 11.649 0.986 0.598 reactant. Trehalose/reactant 4.56 1.309 9.048 0.962 0.457 a Ј2 ␣ 2 ␤ ␥ 2 R for the quadratic regression (aw/m) = aw + aw + . R for the linear regression [aw/m (1 - aw)] = b aw + c. RMS = Root Mean Square. as observed from a smaller RSM value (Bizot 1983; Lievonen and Roos 2002a; Miao and Roos 2004b). The GAB monolayer values were 5.20, 3.95, and 4.56 g water/100 g solids, whereas the BET monolayer values were 6.57, 6.89, and 6.68 g water/100 g solids for lactose/ aw, respectively, in agreement with the results for dehydrated milk reactant, lactose/trehalose/reactant, and trehalose/reactant sys- products reported by Lai and Schmidt (1990), Roos (1993), and tems, respectively. The corresponding aw for the GAB monolayer Jouppila and Roos (1994). Most deteriorative reactions, such as values were 0.24, 0.16, and 0.19 for lactose/reactant, lactose/treh- NEB, are suggested to be more related to molecular mobility above alose/reactant, and trehalose/reactant systems, respectively, and Tg, rather than the monolayer value (Slade and Levine 1991; Kar- the corresponding aw for the BET monolayer values were 0.31, 0.31, and 0.30, respectively.

Glass transition temperature

The glass transition temperatures, Tg, of 3 model systems stored at various RVPs for 144 h are given in Table 3. At RVPs Ն 54.5%, lactose and trehalose in lactose/reactant and trehalose/reactant systems crystallized during equilibration; no glass transition could be observed. The glass transition of the lactose/reactant system was close to that of pure lactose observed by Miao and Roos (2004a) and slightly lower than that of lactose found by Jouppila and Roos (1994). The glass transition temperatures of the trehalose/reactant

system at different RVPs were lower than those measured by Car- E: Food Engineering & Physical Properties dona and others (1997) for trehalose; this was probably due to the additional plasticization of the trehalose by xylose and lysine (Roos and Karel 1991; Buera and others 1992). The Gordon-Taylor (Gor- don and Taylor 1952) Eq. 9 was successfully fitted to the experimental glass transition temperatures, Tg, of the 3 model systems, as shown in Figure 4. Water plasticized the food models and caused a substantial decrease of the glass transition (Roos 1987; Slade and Levine 1991; Roos 1993). The constant, k, for the Gordon-Taylor equation was found to be 7.6 ± 0.8 for lactose/reactant systems, 7.2 ± 0.7 for lactose/trehalose/reactant systems, and 7.9 ± 0.9 for trehalose/reactant systems. The small difference between the k values and the close glass transition temperature (Table 3) indicated that the 3 model systems had similar glass transition behavior, which was typical of lactose-based dairy products (Miao and Roos 2004a). Glass transition is often related to the stability of amorphous foods (White and Cakebread 1966; Roos and Karel 1990; Slade and Le- vine 1991). The critical water content was defined as the water content that depressed Tg to ambient temperature (Roos 1993). The critical water contents at 23 °C in the present study obtained from

Tg data for lactose/reactant, lactose/trehalose/reactant, and trehalose/reactant systems were 7.0, 7.4, and 7.1 g/100 g of dry solids, respectively, which were higher than the GAB monolayer water con- Figure 3—Water sorption isotherm for the model systems (L/R = lactose/reactant system; L/T/R = lactose/trehalose/ tents given in Table 3. Consequently, the critical water activity cal- reactant system; T/R = trehalose/reactant system). culated from GAB model for the lactose/reactant, lactose/treha- Guggenheim-Anderson-deBoer (GAB) isotherm (a) and lose/reactant, and trehalose/reactant systems were 0.34, 0.35, 0.33 Brunauer-Emmett-Teller (BET) isotherm (b) are shown.

URLs and E-mail addresses are active links at www.ift.org Vol. 70, Nr. 2, 2005—JOURNAL OF FOOD SCIENCE E73 NEB kinetics in food systems . . .

mas and others 1992; Roos 1993). Critical water contents are useful Table 3—Glass transition temperatures (mean ± SD) of and important as criteria for processing and storage of amorphous, model systems stored for 144 h at various relative vapor pressure (RVP) at room temperature (22 °C to 23 °C) (n = 3, low-moisture foods. SD = standard deviation)a Glass transition temperature (°C) Nonenzymatic browning RVP (% NEB was observed to occur in the 3 model systems at all RVP Salt of water) L/R L/T/R T/R environments studied during storage. Plot of OD at 280 nm and LiCl 11.4 58.4 ± 0.3 59.7 ± 0.1 55.5 ± 0.1 420 nm against storage time is shown in Figure 5. The browning CH3COOK 23.1 39.4 ± 0.0 36.9 ± 0.2 37.1 ± 0.3 kinetics observed by 280 nm and 420 nm were very similar. The 3 MgCl2 33.2 28.8 ± 0.2 28.8 ± 0.3 26.6 ± 0.1 K CO 44.1 11.9 ± 0.6 9.6 ± 0.1 9.9 ± 0.3 model systems showed different NEB kinetics during storage at 2 3 Mg(NO ) 54.5 0.2 ± 0.4 different RVPs. At RVPs 33.2% and 44.1% (Figure 5a and 5b), optical 3 2 aL/R = lactose/reactant; L/T/R = lactose/trehalose/reactant; T/R = trehalose/ densities for the 3 food models increased linearly, indicating appar- reactant. ent zero-order kinetics (Labuza and Baisier 1992), but leveled at a plateau as the reaction proceeded, and agreed with the results of Lievonen and others (1998) and Lievonen and Roos (2002b). The rate constants of NEB were determined by linear regression of the given in Table 4. The NEB rate at 44.1% RVP was much higher than OD versus storage time over the linear region of the plots and are at 33.2%. Higher RVPs increased the NEB rate. The NEB rates in the trehalose/reactant system were higher than those in lactose/trehalose/reactant systems, and the rate in lactose/reactant system was the lowest of 3 model systems studied at 33.2% and 44.1% RVP.

In the present study, the Tg of the 3 model systems at 33.2% RVP ranged from 26.6 °C to 28.8 °C, which was higher than the storage temperature (22 °C to 23 °C). However, the NEB still occurred, although at a lower rate. This results gave further evidence for the statement by Simatos and others (1995), that glass transition could E: Food Engineering & Physical Properties not be used alone to characterize a material in relation to temperature/water conditions prevailing during storage and could not be considered as an absolute threshold of stability. Although NEB in low-moisture food materials was often suggested to be a diffusion- controlled reaction (Karel and Saguy 1991; Slade and Levine 1991; Karel and Buera 1994; Buera and Karel 1995; Roos 1995, 2001), the formation of glassy polymeric matrix results in significantly “frozen” translational molecular motion, and translational mobility and diffusion are considered to be virtually nonexistent (Slade and others 1989; Slade and Levine 1991, 1992). The heterogeneities of materials and preexisting pores due to defects and porosity in the matrix Figure 4—The relationship between the glass transition and structure (Kovarskii and others 1978; Hori and others 1986; Cardo- water contents for the model systems (L/R = lactose/re- na and others 1997; Schebor and others 1999) probably could ac- actant system; L/T/R = lactose/trehalose/reactant system; T/R = trehalose/reactant system). The Gordon-Taylor equa- count for the occurring of NEB at temperature lower than Tg. The

tion was applied for predicting glass transition (Tg).

Figure 5—Optical density development of the model systems at 280 nm as a function of storage time at room temperature (22 °C to 23 °C). The inset figure shows the optical density development at 420 nm. (L/R = lactose/reactant; L/T/R = lactose/trehalose/reactant; T/R = trehalose/reactant). Error bars represent standard deviation (n = 3).

E74 JOURNAL OF FOOD SCIENCE—Vol. 70, Nr. 2, 2005 URLs and E-mail addresses are active links at www.ift.org NEB kinetics in food systems . . .

Table 4—Rate constants (with 95% confidence limits) of nonenzymatic browning determined by the linear regression of the optical density (OD units/g solid) at 280 nm as a function of storage time at various relative vapor pressures (RVPs)a Rate constant, k ± 95%CL (OD units/g solid/h) RVP (% of water) L/R R2 L/T/R R2 T/R R2 33.2 0.06 ± 0.03 0.9943 0.11 ± 0.09 0.9888 0.19 ± 0.21 0.9895 44.1 0.45 ± 0.08 0.9810 0.78 ± 0.12 0.9862 2.04 ± 0.09 0.9950 54.5b 1.02 ± 0.11 0.9980 1.29 ± 0.07 0.9915 4.68 ± 0.08 0.9993 65.6b 4.69 ± 0.04 0.9841 1.90 ± 0.36 0.9822 5.88 ± 0.11 0.9998 aL/R = lactose/reactant; L/T/R = lactose/trehalose/reactant; T/R = trehalose/reactant. bRate constants were calculated by linear regression from the linear region of OD development plot.

reaction between the xylose and lysine may have occurred primary tion of the primary site of the NEB reaction (pores), the less mobile in the pores that remained after freeze-drying. The pores and cap- matrix with pore loss lowered the browning rate (Bell and others illaries in food materials can hold moisture (Labuza 1968; Karel 1998; White and Bell 1999). In agreement with the previous study 1973; Labuza 1975), which may also explain the reactions in glassy by Buera and Karel (1995) in the polyvinylpyrrolindone (PVP) sys- systems (Bell and others 1998). In fact, from the result of the sorp- tem, Buera and Karel (1995) showed that in the supercooled liquid tion study for the 3 model systems, water contents of the 3 model state (rubbery state) there was a decrease in the browning rate con- system at 33.2% RVP (shown in Table 1) were slightly higher than stant, attributable to increasing moisture content diluting the re- the BET and GAB monolayer water contents for corresponding actants; collapse also retarded the browning reaction. model systems. This might also partially account for the browning The 2nd stage started at about 50 h of storage; the OD in- at 33.2% RVP. It is also possible that materials contain local reactant- creased rapidly and linearly, which was followed by the 3rd pla- rich regions with lower Tg and higher water contents. Such regions teau stage. In agreement with the water sorption in Figure 2a, crys- can brown substantially faster than the continuous glassy matrix. tallization of component sugars in the lactose/reactant and The NEB extent in trehalose model is highest at 33.2% RVP, prob- trehalose/reactant systems observed from the loss of water con- ably because of the local heterogeneities and the special molecu- tents occurred during this 2nd stage. Crystallization of the com- lar structure of trehalose. Trehalose has a special molecular struc- ponent sugars in the model systems seemed to increase the ture: the disaccharide bond is formed between the C1 atoms of the browning rate, indicating that separation of a crystalline phase in pyranose rings, and trehalose has no direct internal hydrogen the material improved reactant mobility and enhanced NEB. Kar- bonds; this arrangement gives the molecule unusual flexibility mas (1994) reported similar results for the trehalose/xylose/lysine above the disaccharide bond (Roser 1991). This may allow treha- system; browning was maximum in samples humidified at RVPs lose to conform more closely to the irregular surface of macromol- in which crystal hydrate formation of trehalose occurred. When ecules than more rigid, directly hydrogen-bonded disaccharides component sugars crystallized, they essentially form a separated and probably leaves reactants closer in the trehalose/reactant sys- phase (Karmas 1994; Sun and Davidson 1998), resulting in phase tem than in the lactose/reactant system. In fact, during our prep- separation in the systems, and reactant molecules were assumed aration of the freeze-dried samples in Petri dishes, we found that to be repelled or released from the crystal structure. In this case,

trehalose models showed different extents of browning between the it was also assumed that the concentration of the reactants in- E: Food Engineering & Physical Properties top and bottom of the sample, and furthermore, that trehalose creased in the remaining noncrystalline matrix (Cardona and oth- samples were more adhesive to the Petri dish surface. All samples ers 1997) during the crystallization period. Meanwhile, during in the present study were prepared by freeze-drying, and the crystallization, sorbed water was released (Roos 1995), which also freeze-dried samples had higher porosity and surface area. Hence, accelerated the molecular mobility in the matrix (Levine and in the amorphous trehalose/reactant system, the trehalose mole- Slade 1992). All these accelerated the NEB in the model systems. cules were arranged more heterogeneously and were adhesive, In fact, during our experiments, different extents of browning in allowing more mobility and availability to react. Of course, the up- 1 sample at different locations were observed, which confirmed take of water at 33.2% and 44.1% RVP environments enhanced this the local heterogeneous and uneven redistribution of the reac- mobility (Sherwin and others 2002; Sherwin and Labuza 2003). Al- tants after crystallization. Comparing the browning rates in the though trehalose is a nonreducing sugar and lactose is a reducing linear region at this stage (Table 4), the rate for trehalose/reactant sugar, which might also participate in the NEB reaction as reactant, systems was higher than the other 2 systems. This was probably in present study, the reactant concentration in the systems was de- related to the formation of dihydrate trehalose crystals in the tre- signed to be 5% (w/w), which was very high. Participation of the halose/reactant system after crystallization. When trehalose takes matrix lactose in the NEB reaction could be very small in compari- water molecules from the environment to form the crystal, the son with the previous effects. reactants are concentrated in the noncrystalline phase, much At 54.5% and 65.6% RVPs, interesting NEB kinetics were ob- more concentrated than in systems with sugars that crystallize as tained, as shown in Figure 5c and 5d. When samples were stored at monohydrate or anhydrate and are less soluble, such as lactose. 54.5% RVP (Figure 5c), OD development was observed in the 3 food When the model samples were stored in a 65.6% RVP environ- models in 3 stages. In the 1st stage, browning developed at a low ment, OD of the 3 systems was observed at 11 h, 24 h, and then at rate, corresponding to the water sorption in Figure 2a, and loss of 24 h intervals up to 240 h, as shown in Figure 5d. The OD in the tre- water contents was not observed, indicating no crystallization of halose/reactant system increased linearly from 11 h, which was component sugars at this stage. During our experiment, we found followed by a leveled-off plateau. In lactose/reactant systems, from 3 model systems at this stage were in a transparent completely col- 11 h to 24 h, the rate constant was low, then it started to increase lapsed state, probably in the collapsed supercooled liquid state linearly and rapidly due to the crystallization of lactose, and finally (rubbery state). On collapse of the matrix, because of the elimina- ended at the plateau stage. In agreement with the water sorption

URLs and E-mail addresses are active links at www.ift.org Vol. 70, Nr. 2, 2005—JOURNAL OF FOOD SCIENCE E75 NEB kinetics in food systems . . .

results for the model systems at 65.6% RVP (Figure 2b), crystalliza- References tion of component sugars was observed from the loss of sorbed Arvanitoyannis I, Blanshard JMV. 1994. Rate of crystallization of dried lactose- sucrose mixtures. J Food Sci 59(1):197–205. water at 5 h, 21 h, and 56 h for trehalose/reactant, lactose/reactant, Bell LN. 1996. Kinetics of non-enzymatic browning in amorphous solid systems: and lactose/trehalose/reactant systems, respectively. The rapid distinguishing the effects of water activity and glass transition. Food Res Int 28:591–7. linear increasing region of OD in the trehalose/reactant and lac- Bell LN, Labuza TP. 2000. Determination of moisture sorption isotherms. In: tose/reactant systems seemed to be directly related to the crystal- Moisture sorption. Practical aspects of isotherm measurement and use. 2nd Ed., St. Paul, Minn.: American Assn. of Cereals Scientists. p 33–8. lization of the component sugars in the systems. The slow loss of Bell LN, Touma DE, White KL, Chen Y. 1998. Glycine loss and Maillard browning the sorbed water in the lactose/trehalose/reactant system probably as related to the glass transition in a model food system. J Food Sci 63:625–8. Berlin E, Anderson BA, Pallansch MJ. 1968. Water vapour sorption properties of explained the continuously linear increase in the OD before the OD various dried milks and whey. J Dairy Sci 51(9):1339–44. leveled off. It should be mentioned that the rate constant (shown Biliaderis CG, Lazaridou A, Arvanitoyannis I. 1999. Glass transition and phys- in Table 4) in the trehalose/reactant system at 65.6% RVP was close ical properties of polyol-plasticized pullulan-starch blends at low moisture. Carbohydr Polym 40:29–47. to that in lactose/reactant system at 65.6% RVP, and also only Biliaderis CG, Lazaridou A, Mavropoulos A, Barbayiannis N. 2002. Water plasti- slightly higher than that at 54.5% RVP. Typically, the maximum NEB cizing effects on crystallization behavior of lactose in a co-lyophilized amorphous polysaccharide matrix and its relevance to the glass transition. Int J rates in foods and model systems occur in a RVP range of 50% to 75% Food Prop 5(2):463–82. (Labuza and Saltmarch 1981; Buera and Karel 1993) depending on Bizot H. 1983. Using the GAB model to construct sorption isotherms. In: Jowitt R, Escher F, Hallström B, Meffert HFT, Spiess WEL, Vos G, editors. Physical proper- the composition of the systems; probably the rate constant in the ties of foods. London: Applied science Publishers. p 43–54. trehalose/reactant system at 65.6% RVP seemed to approach the Brunauer S, Emmett PH, Tellet E. 1938. Adsorption of gases in multimolecular layers. J Am Chem Soc 60:309–19. maximum value. Buera MP, Karel M. 1993. Application of the WLF equation to describe the com- Trehalose has been proved in many studies to be a excellent bined effects of moisture and temperature on nonenzymatic browning rate in protectant for stabilization of enzymes and other biomaterials food systems. J Food Proc Preserv 17:31–45. Buera MP, Karel M. 1995. Effect of physical changes on the rates of nonenzymatic (Crowe and others 1984; Leslie and others 1995; Scherbor and oth- browning and related reaction. Food Chem 52:167–73. ers 1996; Cardona and others 1997; Mazzobre and others 1997a, Buera MP, Levi G, Karel M. 1992. Glass transition in poly(vinylpyrrolidone): effect of molecular weight and dilutes. Biotechnol Prog 8:144–8. 1997b). Some studies (Mazzobre and others 1997a) suggested that Cardona MF, Schebor C, Buera MP, Karel M, Chirife J. 1997. Thermal stability of the trehalose-protecting effect cannot be based solely on its glass- invertase in reduced-moisture amorphous matrices in relation to glassy state and trehalose crystallization. J Food Sci 62(1):105–12. forming properties because other dried molecular weight polymers Champion D, Le Meste M, Simatos D. 2000. Toward an improved understanding E: Food Engineering & Physical Properties of glass transition and relaxations in foods: molecular mobility in the glass with higher Tg (such as, PVP 400000) than trehalose exerted less transition range. Trends Food Sci Technol 11:41–55. protective effect than trehalose, and a specific protective mecha- Crowe LM, Crowe JH, Chapman D. 1984. Preservation of membranes of anhydro- nism, related to the maintenance of the tertiary structure, was sug- biotiv organisms: the role of trehalose. Science 233:701–3. Eichner K. 1981. Antioxidative effect of Maillard reaction intermediates. Prog gested (Cardona and others 1997). From the results of the present Food Nutr Sci 5:441–51. study, the relatively higher NEB rate in trehalose/reactant system Flink JM. 1983. Nonenzymatic browning of freeze-dried sucrose. J Food Sci seemed to be more related to the composition, rather than the 48:539–42. Friedman M. 1996. Food browning and its prevention: an overview. J Agric Food glass transition properties. Chem 44:631–53. Gabarra P, Hartel W. 1998. Corn syrup solids and their saccharide fraction affect crystallization of amorphous sucrose. J Food Sci 63:523–8. Conclusions Gordon M, Taylor JS. 1952. Ideal copolymers and the second-order transitions of he present study compared the physical state, thermal behav- synthetic rubbers. I. Non-crystalline copolymers. J Appl Chem 2:493–500. Greenspan L. 1977. Humidity fixed points of binary saturated aqueous solution. Tior, and the NEB kinetics in freeze-dried lactose-, trehalose-, J Res Nat Bur Stand, Phys Chem 81A:89–96. and lactose/trehalose-based food model systems at different RVP Haque MDK, Roos YH. 2004. Water plasticization and crystallization of lactose in spray-dried lactose/protein mixtures. J Food Sci 69:23–9. environments. The results have shown that the 3 model systems Hartel RW, Shastry AV. 1991. Sugar crystallization in food products. Crit Rev Food had similar glass transition behavior but different crystallization Sci Nutr 30:49–112. Hodge JE, Osman EM. 1976. Carbohydrates. In: Fennema OR, editor. Principle of properties. Mixtures of lactose and trehalose showed delayed crys- food science. New York: Dekker. p 41–138. tallization of components. Addition of trehalose in lactose-based Hori T, Fujita I, Shimizu T. 1986. Diffusion of disperse dyes into Nylon 6 above dairy products may prevent crystallization during storage. The NEB and below the glass-transition temperature. J Society Dyers Colourists. 102:181–5. rate was accelerated by crystallization and affected by the matrix Iglesias HA, Chirif J. 1987. Delayed crystallization of amorphous sucrose in hu- composition. In the trehalose/reactant system, the NEB rate was midified free dried model systems. J Food Technol 13:137–44. Iglesias HA, Chirife J, Buera MP. 1997. Adsorption isotherm of amorphous treh- higher than in lactose/trehalose/reactant and trehalose/reactant alose. J Sci Food Agric 75:183–6. systems. As a nonreducing sugar, trehalose has been suggested to Johari GP, Hallbrucker A, Mayer E. 1987. The glass-liquid transition of hyper- quenched water. Nature 330:552–3. be used as a new ingredient in the food industry to prevent NEB, Jouppila K, Roos YH. 1994. Water sorption and time-dependent phenomena of especially in the dairy industry. We found that the NEB rate was the milk powders. J Dairy Sci 77:1798–808. Karel M. 1973. Recent research and development in the field of low-moisture highest in the trehalose/reactant system, and the reaction was ac- and intermediate moisture foods. CRC Crit Rev Food Technol 3:329–73. celerated by crystallization during storage. These results indicated Karel M, Buera MP. 1994. Glass transition and its potential effects on kinetics of condensation reactions and in particular on nonenzymatic browning. In: that NEB was related to the composition of the matrix. When treh- Labuza TP, Reineccius GA, Monnier VM, O’Brien J, Baynes JW, editors. Maillard alose was used in dairy products, especially those containing pro- reactions in chemistry, food and health. Cambridge, U.K.: Royal Society of teins and reducing sugars, which might likely cause browning, Chemistry. p 164–9. Karel M, Saguy I. 1991. Effects of water on diffusion in food system. In: Levine H, much attention should be paid in controlling the NEB reaction Slade L., editors. Water relationships in foods. New York: Plenum Press. p 157–74. during storage and processing. Because most of the dairy powders Karmas R. 1994. The effect of glass transition on non-enzymatic browning in dehydrated food systems [DPhil thesis]. New Brunswick, N.J.: Rutgers, the State are produced by spray-drying, to evaluate the NEB kinetics in real Univ. of New Jersey. 168 p. food products, further research is needed to confirm the effects of Karmas R, Buera MP, Karel M. 1992. Effect of glass transition on rates of nonenzymatic browning in food systems. J Agric Food Chem 40(5):873–9. matrix composition and crystallization on NEB kinetics in spray- Karmas R, Karel M. 1994. The effect of glass transition on Maillard browning in dried systems. food models. In: Labuza TP, Reineccius GA, Monnier VM, O’Brien J, Baynes JW, editors. Maillard reactions in chemistry, food and health. Cambridge U.K.: Royal Society of Chemistry. p 182–7. Acknowledgments Kouassi K, Roos YH. 2000. Glass transition and water effects on sucrose inver- sion by invertase in a lactose-sucrose system. J Agric Food Chem 48:2461–6. The study was carried out with the financial support of PRTL1 Cycle Kouassi K, Roos YH. 2001. Glass transition and water effects on sucrose inver- 3 of the Higher Education Authority of Ireland. sion on noncrystalline carbohydrate food systems. Food Res Int 34:895–901.

E76 JOURNAL OF FOOD SCIENCE—Vol. 70, Nr. 2, 2005 URLs and E-mail addresses are active links at www.ift.org NEB kinetics in food systems . . .

Kovarskii AL, Placek J, Szocs F. 1978. Study of rotational mobility of stable nitrox- rides and sugars. Biotechnol Prog 7:49–53. ide radicals in solid polymers. Polymer 19:1137–41. Roos YH, Karel M. 1990. Differential scanning calorimetry study of phase tran- Labuza TP. 1968. Sorption phenomena in foods. Food Technol 22:263–72. sition affecting the quality of dehydrated materials. Biotechnol Prog 6:159–63. Labuza TP. 1975. Sorption phenomena in foods: theoretical and practical aspects. Roos YH, Karel M. 1992. Crystallisation of amorphous lactose. J Food Sci 57:775– In: Rha C, editor. Theory determination and control of physical properties of 7. food materials. The Netherlands, Dordecht: D. Reidel Publishing. p 197–219. Roos YH, Karel M, Kokini JL. 1996b. Glass transition in low moisture and frozen Labuza TP, Baisier WM. 1992. The kinetics of nonenzymatic browning. In: foods: Effects on shelf life and quality. Food Technol 50(11):95–108. Schwartzberg HG, Hartel RW, editors. Physical chemistry of foods. New York: Roser B. 1991. Trehalose, a new approach to premium dried foods. Trends Food Marcel Dekker. p 595–649. Sci Technol 2:166–9. Labuza TP, Kaanane A, Chen JY. 1985. Effect of temperature on the moisture sorp- Saleki-Gerhardt A, Zografi G. 1994. Non-isothermal and isothermal crystallization isotherms and water activity shift of two dehydrate foods. J Food Sci 50:385– tion of sucrose from the amorphous state. Pharm Res 11(8):1166–73. 91. Saltmarch M, Vagnini-Ferrari M, Labuza TP. 1981. Theoretical basis and applica- Labuza TP, Saltmarch M. 1981. The nonenzymatic browning reaction as affected tion of kinetics to browning in spray-dried whey food systems. Prog Food Nutr by water in foods. In: Rockland LB, Stewart GF, editors. Water activity: influenc- Sci 5:331–44. es on food quality. New York: Academic Press. p 605–50. Schebor C, Buera MP, Chirife J. 1996. Glassy state in relation to the thermal in- Lai H, Schmidt SJ. 1990. Lactose crystallization in skim milk powder observed by activation of enzyme invertase in amorphous dried matrices of trehalose, hydrodynamic equilibria, scanning electron microscopy and 2H nuclear mag- maltodextrin and PVP. J Food Eng 30:269–82. netic resonance. J Food Sci 55(4):994–9. Schebor C, Buera MP, Karel M, Chirife J. 1999. Color formation due to non-enzy- Leslie SB, Israeli E, Lighthart B, Crowe JH, Crowe LM. 1995. Trehalose and su- matic browning in amorphous, glassy, anhydrous model systems. Food Chem crose protect both membranes and protein in intact bacteria during drying. 65(4):427–32. Appl Environ Microbiol 61:3592–7. Schebor C, Burin L, Buera MP, Aguilera JM, Chirife J. 1997. Glassy state and ther- Levine H, Slade L. 1992. Glass transition in foods. In: Schwartzberg HG, Hartel mal inactivation of invertase and lactase in dried amorphous matrices. Bio- RW, editors. Physical chemistry of foods. New York: Marcel Dekker. p 83–221. technol Prog 12:857–63. Lievonen SM, Laaksonen TJ, Roos YH. 1998. Glass transition and reaction rates: Schiraldi C, di Lernia I, de Rosa M. 2002. Trehalose production : exploiting novel nonenzymatic browning in glass and liquid systems. J Agric Food Chem approaches. Trends Biotechnol 20(10):420–5. 46:2778–84. Shamblin SL, Huang EY, Zografi G. 1996. The effects of co-lyophilized polymeric Lievonen SM, Laaksonen TJ, Roos YH. 2002. Non-enzymatic browning in food additives on the glass transition temperature and crystallization of amor- models in the vicinity of glass transition: effects of fructose, glucose, and xy- phous sucrose. J Thermal Anal 47:1567–79. lose as a reducing sugar. J Agric Food Chem 50(24):7034–41. Sherwin CP, Labuza TP. 2003. Role of moisture in Maillard browning reaction Lievonen SM, Roos YH. 2002a. Water sorption of food models for studies of glass rate in intermediate moisture foods: Comparing solvent phase and matrix transition and reaction kinetics. J Food Sci 67:1758–66. properties. J Food Sci 68:588–94. Lievonen SM, Roos YH. 2002b. Non-enzymatic browning in amorphous food Sherwin CP, Labuza TP, McCormick A, Chen B. 2002. Cross-polarization/magic models: effects of glass transition and water. J Food Sci 67(6):2100–6. angle spinning NMR to study glucose mobility in a model intermediate-mois- Linko P, Pollari M, Heikonen M. 1982. Water sorption properties and the effects of ture food system. J Agric Food Chem 50:7677–83. moisture on structure of dried milk products. Lebensm-Wiss Technol 15:26–30. Simatos D, Blond G, Pérez J. 1995. Basic physical aspects of glass transitions. In: Mazzobre MF, Buera MP, Chirife J. 1997a. Protective role of trehalose on thermal Barbosa-Cánovas GV, Welti-Chanes J, Eds. Food preservation by moisture con- stability of lactase in relation to its glass and crystal forming properties and trol Lancaster, Pa.: Technomic. p 3–31. effect of delay crystallization. Lebensm-Wiss Technol 30:324–29. Simatos D, Karel M. 1988. Characterization of the condition of water in foods Mazzobre MF, Buera MP, Chirife J. 1997b. Glass transition and thermal stability physiochemical aspects. In: Seow CC, editor. Food preservation by moisture of lactase in low-moisture amorphous polymeric matrices. Biotechnol Prog control. London: Elsevier Applied Science. p 1–39. 13:195–9. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an Mazzobre MF, Soto G, Aguilera JM, Buera MP. 2001. Crystallization kinetics of alternative approach to the assessment of food quality and safety. Crit Rev lactose in systems co-lyofilized with trehalose. Analysis by differential scan- Food Sci Nutr 30(2/3):115–359. ning calorimetry. Food Res Int 34:903–11. Slade L, Levine H, Finley JW. 1989. Protein-water interactions: water as a plas- Miao S, Roos YH. 2004a. Nonenzymatic browning kinetics of a carbohydrate- ticizer of gluten and other protein polymers. In: Phillips RD, Finley JW, editors. based low –moisture food system at temperatures applicable to spray drying. Protein quality and the effects of processing. New York: Marcel Dekker. p 9– J Agric Food Chem 52:5250–7. 124. Miao S, Roos YH. 2004b. Comparison of nonenzymatic browning kinetics in Sperling LH. 1992. Introduction to physical polymer science. New York: John spray-dried and freeze-dried carbohydrate-based food model systems. J Food Wiley & Sons. 594 p. Sci 69:322–31. Sun WQ, Davidson P. 1998. Protein inactivation in amorphous sucrose and treh- O’Brien J. 1996. Stability of trehalose, sucrose and glucose to nonenzymatic alose matrices: effects of phase separation and crystallization. Biochim Bio- browning in model systems. J Food Sci 61:679–82. phys Acta 1425:235–44. O’Brien J, Morrissey PA. 1989. Nutritional and toxicological aspects of the Mail- Uritani M, Takai M, Yoshinaga K. 1995. Protective effect of disaccharides on re- lard browning reaction in foods. Cri Rev Food Sci Nutr 28:211–48. striction endonucleases during drying under vacuum. J Biochem 117:774–9. Resnik S, Chirife J. 1979. Effects of moisture and temperature on some aspects of Van Boekel MAJS. 2001. Kinetic aspects of the Maillard reaction: a critical re- E: Food Engineering & Physical Properties nonenzymatic browning in dehydrated apple. J Food Sci 44:601–5. view. Nahrung 45:150–9. Roos YH. 1987. Effect of moisture on the thermal behavior of strawberries studied Van den Berg C, Bruin S. 1981. Water activity and its estimation in food systems: using differential scanning calorimetry. J Food Sci 52:146–9. Theoritical aspects. In: Rockland LB, Stewart GF, editors. Water activity: influ- Roos YH. 1993. Water activity and physical state effects on amorphous food sta- ences on food quality. New York: Academic Press. p 1–61. bility. J Food Process Preserv 16:433–47. Vuataz G. 1988. Preservation of skim milk powders: role of water activity and Roos YH. 1995. Phase transitions in foods. New York: Academic Press. 360 p. temperature in lactose crystallization and lysine loss. In: Seow CC, editor. Roos YH. 2001. Water activity and plasticization. In: Eskin NAM, Robinson DS, Food preservation by moisture control. London: Elsevier Applied Science. p editors. Food shelf life stability. New York: CRC Press. p 3–36. 73–101. Roos YH, Himberg MJ. 1994. Non-enzymatic browning behavior, as related to Whistler RL, Daniel JR. 1985. Carbohydrates. In: Fennema OR, editor, Food chem- glass transition of a food model at chilling temperatures. J Agric Food Chem istry. 2nd ed. New York: Dekker. p 69–137. 42(4):893–8. White GW, Cakebread SH. 1966. The glassy state in certain sugar-containing Roos YH, Jouppila K, Zielasko B. 1996a. Non-enzymatic browning-induced water food products. J Food Technol 1:73–82. plasticization. Glass transition temperature depression and reaction kinetics White KL, Bell LN. 1999. Glucose loss and Maillard browning in solids as affected determination using DSC. J Thermal Anal 47(5):1437–50. by porosity and collapse. J Food Sci 64:1010–4. Roos Y, Karel M. 1991, Phase transition of mixtures of amorphous polysaccha-

URLs and E-mail addresses are active links at www.ift.org Vol. 70, Nr. 2, 2005—JOURNAL OF FOOD SCIENCE E77