J. Dairy Sci.

88:1636–1645
© American Dairy Science Association, 2005.

Rate of Maillard Browning in Sweet Whey Powder
R. Sithole, M. R. McDaniel, and L. Meunier Goddik
Department of Food Science and Technology, Oregon State University, Corvallis 97331

ABSTRACT

INTRODUCTION

The objective of this study was to evaluate the rate
of Maillard browning in 3 commercial sweet whey powders (WC1, WC2, and MW1), under accelerated shelflife testing (ASLT) and under normal storage conditions (21°C and 35% RH). Rate of brown pigment formation (k) obtained from short-term ASLT of whey
powder was compared with actual findings obtained
from the long-term shelf-life testing under normal
conditions. Deterioration by Maillard browning, measured by spectrophotometer, was compared with
changes in color (Hunter Laboratory), free moisture,
titratable acidity, and sensory attributes. Results suggest that estimated k (from ASLT) was comparable
with the observed rate (obtained at ambient temperature) for 2 producers (WC1, MW1). The actual k values
observed for samples WC1, WC2, and MW1, stored
under normal conditions, were 0.0031, 0.0080, and
0.0148 color units/g of solid per mo, respectively. The
estimated values of k for samples WC1, WC2, and
MW1 were 1.12, 4.90, and 1.35 times more than the
observed values, respectively. The Q10 values (increase in reaction rate for a 10°C temperature increase) ranged from 1.77 to 4.14, and the activation
energies ranged from 15.9 to 28.4 kcal/mol. Hunter
Laboratory values L* and a* appeared most sensitive
to changes during storage. Free moisture content, and
acidity increased significantly with storage. However,
no significant changes were detected by the sensory
panel in the attributes considered.
(Key words: Cheddar whey powder, storage stability,
Maillard browning, accelerated shelf-life testing)

Nonenzymatic browning via the Maillard reaction
is an important mode of deterioration in dried milk
and whey powders, which limits shelf life (Saltmarch
and Labuza, 1980). Whey powders contain relatively
high concentrations of lactose (approximately 73%)
and protein (approximately 12%) high in lysine content. In the presence of moisture, these components
readily participate in Maillard reactions. The Maillard reaction is affected by the concentration of the
initial reactant species, pH, water content, and presence of substances such as humectants and bisulfite
(Franzen et al., 1990). Some physical factors, such
as processing and storage temperature, atmospheric
oxygen, and packaging during storage can also affect
the Maillard reaction in foods. The deleterious effects
of nonenzymatic browning include: decreased nutritional value from protein loss, off-flavor development,
undesirable color, decreased solubility, texture
changes, destruction of vitamins, and increased acidity (Saltmarch and Labuza, 1981; Ford et al., 1983;
Villota and Hawkes, 1983).
Brown pigment formation has been used as an indicator of the Maillard reaction in food (Waletzko and
Labuza, 1976; Saltmarch and Labuza, 1981). Franzen
et al. (1990) confirmed that the Maillard reaction is
described by zero-order reaction in the sense that the
concentration of the brown pigments is negligible compared with the concentration of reactants present. Optimum conditions for the Maillard reaction in whey
powders as determined by brown pigment formation
have been studied (Choi et al., 1949; Labuza and Saltmarch, 1981).
In their method development for accelerated shelf
life testing (ASLT) of whey powder, Labuza and Saltmarch (1981) stored samples in an open system. However, Kim et al. (1981) investigated the effects of storing the samples in open vs. sealed pouches. They observed that the nonenzymatic browning reaction is
greatly increased in closed samples compared with
samples stored open to the environment, and cautioned that care be taken in using data from ASLT to
predict kinetics of deteriorative reactions during
shelf-life tests of sealed food systems.
The purpose of this work was to use sealed containers to estimate the rate of deterioration of 3 commer-

Abbreviation key: a* = Hunter Lab red-green parameter, ASLT = accelerated shelf-life testing, b* =
Hunter Lab blue-yellow parameter, Ea = energy of
activation, k = rate constant, L* = Hunter Lab
lightness-darkness parameter, MW = Midwest processor, WC = West coast processor.

Received October 25, 2004.
Accepted December 28, 2004.
Corresponding author: L. Meunier Goddik; e-mail: Lisbeth.goddik
@oregonstate.edu.

1636

37. 12.05). 985. method no. Titratable acidity as percentage lactic acid and pH were determined according to the methods outlined by the American Dairy Products Institute (1991). 1985). Palo Alto. 17. 935. Hunter Associates Laboratory Inc. method no. Sulfite was quantified and reported as sulfur dioxide (AOAC. VA). 2000.MAILLARD BROWNING IN SWEET WHEY POWDER cial whey powders under accelerated storage conditions through the use of the Arrhenius equation with extrapolation. and 21 d of storage. 88. and scorched particles were analyzed according to the method outlined by the American Dairy Products Institute (1991). and 55°C. method no. 2000. Chymotrypsin type II from bovine (C4129-1G). 24. and samples stored at 55°C were withdrawn after 5. with analyses performed in triplicate. 10. The ASLT data were compared with Maillard browning occurring under normal storage conditions (21°C. 10° observer value. sweet. Samples stored at 45°C were withdrawn after 9. 984. 45. The operating conditions were: illuminant C.15). The storage times were chosen based on the study by Presa-Owens et al. 20. and MW1). This impermeable system enabled Maillard reaction in high-temperature/low-humidity cabinets without moisture loss. 40. 2000. 9. Louis. 91. WC2. and 174 d of storage. Analysis was by the tristimulus reflectance colorimeter. 19. Fat content was determined by the Mojonnier method (AOAC. 20. 35% RH). The rate of deterioration by browning of the 3 commercial Cheddar cheese whey powder samples was compared with changes in microbiological. Color measurements were determined according to the Hunter Laboratory method of measuring loose powder (Nielsen et al.20). Reston.. method no. Kyoto. 19. and 19 mo.. Fisher Scientific. Duplicate analyses were carried out for all analyses. Calcium was determined by atomic absorption spectrophotometry (AOAC. 991.5-g whey powder sample in 100 mL of distilled water. Whey powders from WC1. and 46 d of storage. 2000. 1995. and MW1 were each packaged into 100-mL glass bottles and equilibrated to 0. and reflectance mode and 45/0 sensor. # 206-67001-92. Duplicate analyses were carried out for all analyses.. Five to 6 bottled whey powder samples from each processor were incubated at each temperature. Samples stored at 50°C were withdrawn after 11. (1995). 50. 26. Odor perception and descriptive identification were assessed by having the MATERIALS AND METHODS Storage Conditions for Control Twelve bags from the same lot of whey powder from 2 West coast processors and 1 Midwest processor (WC1. 1997). 1985.35). physicochemical. 1999) (Bio Spec – 1601 DNA/Protein Analyzer cat. Free moisture was determined by oven drying according to the standard methods for the examination of dairy products (Richardson. CA).44 water activity (aw) according to the method outlined by Labuza and Saltmarch (1981). Samples stored at 35°C were withdrawn after 7. 64. and sensory quality during storage to establish the keeping quality of the whey powders and to validate the use of Maillard browning in determining shelf life. Analytical Methods Four incubators were maintained at 35. Samples for sensory analyses were packed in 1-L amber glass bottles supplied with polytetrafluoroethylene-sealed screw caps and stored at −37°C until analysis. Japan). High Temperature Storage Journal of Dairy Science Vol. and lactose was determined using an enzymatic method (AOAC. Fairlawn. The CIE LAB values L*. and bitter). 2005 . 5. and b* were measured. 1637 IX-5 T0303-1G). Evaluation of Brown Pigment Formation Sensory Evaluation Brown pigment formation was analyzed by the modified method for milk powders of Choi et al. No. MO). and peptidase from porcine intestinal mucosa (P7500-10UN) were purchased from Sigma (St. The bottles were then sealed with polytetrafluoroethylenesealed screw caps. Samples were withdrawn for analysis after storage for 1. 5. sour. 60. 989. The pH was measured using a pH meter (Accumet Research AR25 dual Channel pH/Lon Meter. (1949) as outlined by Labuza and Saltmarch (1981). were stored at 21°C and 35% RH. NJ) on a 6.29.43). WC2. trypsin (type Prospective panelists were screened for their ability to recognize and rate the intensity of the 4 basic tastes (salty. and 96 d of storage. using a Beckman model TJ-6 centrifuge (Beckman. Absorbance was measured at 420 nm (Ca¨mmerer et al. 20:123–125). method no. method no. a*. Shimadzu. 930. Salt was determined by indirect Volhart method (AOAC. The solubility index was determined by centrifugation according to A/S Niro atomizer (1978). Hunter Laboratory CT 1100 Color Quest (SNC49038. Samples were withdrawn from the incubators after the specified storage period and immediately stored at −37°C. and protein content determined by the Kjeldahl method (AOAC.

1993) was used. 2 ppb in skim milk.2967a 1. pasteurized skim milk heated to 85°C for 55* min. 2001.4-decadienal (12117CB. caramelized.5967b 1. Sigma) Aroma and flavor associated with cooked milk. metallic. and barny via duplicated triangle tests.8967b 11.7a 12. Basic taste associated with sucrose. Results of acuity and rating tests were used to make the final selection of 13 panelists (12 women and 1 man).530b 488. and burnt. caramel. 5. slightly sulfuric aroma/taste acquired by a product that has been submitted to heat treatment.6800b 1. Aldrich) Papery/cardboard aroma Cardboard Sweet Sucrose Salty NaCl Sour Citric acid Aroma/flavor associated with old oil.0a 72. Basic taste associated with acid. 88.05%* wt/vol citric acid solution. description terms. Aroma associated with cardboard.2367a 1.1638 SITHOLE ET AL. 0.7c 403. 1%* wt/vol sucrose solution. 2001). Descriptors used are shown in Table 1.8a 6.2%* wt/vol NaCl solution. Means from triplicate analysis of composite sample.1533a 6. Table 1.197a 2. production. 0. (Panelists created the preliminary lexicon after tasting the whole range of whey powders. No. Milk autoclaved at 121°C for 30 min.8700a 12. 1999) was used because it uses scientific terms. Oxidized aroma and flavor (E. The Quantitative Descriptive method (Stone and Sidel. This characteristic goes with the series biscuit. Testing was a complete randomized block design with three coded samples presented Table 2. 10* ppm in skim milk. Processor Salt (%) Calcium (%) Protein (%) Free moisture (%) Fat (%) Lactose (%) pH Sulfite (ppm) WC1 WC2 MW1 2. panelists take short sniffs of the attributes cooked.2a 71. paint. which relate to product composition. a-c 1 Journal of Dairy Science Vol. taste suggestive of soft caramel due to Maillard reaction.2300a 1. Aroma and flavor associated with diacetyl. 2 ND = Not detected. Training was carried out during nineteen 1-h sessions.. Additional descriptors and standards were adopted from Karagu¨l-Yu¨ceer et al. Descriptor Reference Description and preparation Cooked aroma and flavor Heated skim milk Caramelized aroma and flavor Autoclaved skim milk Sweet aromatic/cake mix aroma and flavor Barny/animal-like aroma and flavor Pillsbury Moist Supreme classic white premium cake mix p-cresol (C7525.6067b 71.187a 2. sensation provoked by defective packaging or by a slight oxidation of product.2900a 6. Descriptors. (2001). and preparation of the reference materials for descriptive sensory evaluation of whey powder (adopted from Karagu¨l-Yu¨ceer et al.05). Proximate composition1 with separation of means of commercial Cheddar whey powder samples from processors from the West coast (WC1 and WC2) and the Midwest (MW1).E)-2. 2005 .2667a ND2 ND ND Means within columns with different letters are statistically different (P < 0. pieces of cardboard paper (3% wt/vol) soaked in skim milk overnight.6000b 1. Basic taste associated with salt.0b 559. and development. rather than consumer terms. A modified Spectrum method (Lawless and Heymann. making it easier to relate to the science behind the product. *Signify standards modified from Karagu¨l-Yu¨ceer.

9982 = = = = = = = = = = = = = = = 0.01 mL of color (McCormick red food color to prevent the appearance of whey from influencing a panelist’s decision) were suspended in 100 mL of spring water at 40°C and mixed by electric mixer at 200 rpm for 2 min.2007 0. For the compositional and microbiological Figure 1.8431 0. 1999) to allow for release of volatile aroma into the head space. Processors (WC1. All methods used were from the American Dairy Products Institute (1991). dotted line separated on PC1. After evaluation of aroma.8647 0.2129 0. The microbiological media were supplied by VWR International (West Chester.9803 0.3829 0.8811 0.9831 0.0073x + 0. Equations and R2 values for the Arrhenius plots (log k vs. First 2 principal components (PC1.0116x + 0. Panelists constituted a block.1592 0. Microbiological Analysis The microbiological analyses carried out on the whey sample were standard plate count. Samples (70 mL) were served in clear wine glasses covered with plastic lids.. Journal of Dairy Science Vol. No. MW = processor from Midwest. and were evaluated at 25 ± 2°C after standing for 2 h (Kamath et al. coliform bacteria. to the panelist at one time.0018x + 0. Statistics Multivariate and univariate ANOVA were performed.0028x + 0. storage time.9327 0. flavor was evaluated.0199x + 0.8684 0.9395 0.891 3468.9542 0.9945 0. and yeast and molds. Storage temperature (°C) Processor1 Equation R2 35 45 50 21 35 50 55 21 35 45 50 21 Arrhenius plot Arrhenius plot Arrhenius plot WC1 WC1 WC1 WC1 WC2 WC2 WC2 WC2 MW1 MW1 MW1 MW1 WC1 WC2 MW1 y y y y y y y y y y y y y y y 0.0006x + 0. 10 g of whey powder and 0. MW1) in same circle have no significant difference at 95% confidence interval.8247 0.3046 0. Bold line separated on PC2.3181 0. 2005 . WC2.2772 0.1639 MAILLARD BROWNING IN SWEET WHEY POWDER Table 3.188 1 WC = Processors from West coast.2387 0.7555 5770.3076 0.00204x + 0. 1/K) and the plots of color units/g of solid vs.000 0.2x − 16.2447 0.8923 1.118 6206.4x − 16. mesophilic and psychrophilic spores. 88.0057x + 0. PC2) of compositional analysis of commercial whey.0148x + 0. 5.0031x + 0. For aroma and flavor evaluation.0034x + 0. Each panelist judged each sample twice.0084x + 0.3x − 8. PA).988 0.2895 0.

and presence of trace metals and other catalysts.. Rate constants at 21. B. Samples MW1 and WC2 were significantly different from WC1 based on principal component 2. C. 2005 Figure 2. ANOVA with means separation was performed using SPSS version 11. 88. Rate of brown pigment formation for sweet whey powder stored at elevated temperatures: 35°C (♦). IL) at 95% confidence. 5. No. Ea = activation energy (kcal/mol).. Most parameters that influence browning were kept constant except for composition. Principal component 2 separates the samples based on moisture.0 (SPSS Inc. Rate of Brown Pigment Formation Rate constants were obtained for brown pigment formation assuming a zero-order rate relationship based on the work of Labuza and Saltmarch (1981) on kinetics of browning and protein quality loss in whey powders. simultaneous confidence interval for differences on the 5% level were calculated using Tukey LSD.303R. storage time (Figures 2A and 2B). Inc. and 50°C were obtained from the plots of color units/g of solid vs. Chicago. Principal component analysis (Figure 1) demonstrated that the 3 samples were significantly different based on principal component 1. RESULTS AND DISCUSSION Composition The compositional data of the whey powders is given in Table 2. and R (ideal gas constant) = 1. . Principal component 1 separates the samples based on calcium.98722 cal/K per mol.1640 SITHOLE ET AL. and 50°C (䊏). Differences between whey powders were recognized by principal component analysis. history of the raw material. whose slope = −Ea/2. 45°C (▲). and protein. variation in the rate of brown pigment formation would likely be a result of one or a combination of these factors. coreactant level. Plots for determining rate of brown pigment formation under accelerated shelf life testing (ASLT). The Q10 factor is defined as the rate of reaction a temperature (T + 10) divided by the Journal of Dairy Science Vol. Therefore. where k = rate constant for deteriorative reaction at temperature T. Arrhenius plot for the rate of brown pigment formation. For the sensory data. The data presented in Figure 2 were based on sample MW1. 35% RH. A. analyses. The factors storage time. panelist. salt. NC). and fat. Cary. lactose. 35. The temperature dependence of browning was determined using the typical Arrhenius (Figure 2C) relationship to obtain activation energies (Ea) and Q10 values (increase in reaction rate for a 10°C increase in temperature). The Arrhenius equation is derived from a plot of log k vs. Based on the results of the ANOVA. ambient conditions and the Arrhenius plot for whey powder sample from Midwest. pH. 3-way ANOVA for each descriptor was performed using multivariate ANOVA using SAS version 8 (SAS Institute. Rate of brown pigment formation at 21°C. 1/T(K). and replication were taken into consideration. 45.

The higher free moisture contents in WC2 and MW1 likely contributed to their higher rates of brown pigment formation.40 15. which should result in increased Maillard reaction rate.872 26. Actual k values were obtained from samples stored at 21°C and 35% RH. the increase in rate of brown pigment formation was the smallest. All 3 samples darkened with storage (Table 5). 2005 . 1/T(K) (Figure 2 C). Although only results from MW1 were plotted. other samples exhibited similar behavior. R2 between 0. Although WC2 was most prone to Maillard reactions. MW = processor from Midwest.77. whereas WC2 and MW1 gained in redness. However. 4 Factor by which the estimated value is bigger than the actual k. Sample WC2 had the lowest Ea and the highest initial pH. Actual k was determined from whey powders stored at 21°C and 35% RH.8 3. Samples WC1 and MW1 fell within this range. Plots of color units/g of solid vs. 2 rate of reaction at temperature (T). unlike WC2.90. Storage time had a highly significant effect on moisture. No. One possible cause of the lower than expected rate of Maillard reaction in WC2 might be the presence of an additive that could inhibit or retard the Maillard reaction. hence increasing the reaction rate. The estimated values from accelerated testing of k at room temperature were determined by extrapolation from the Arrhenius plots of log k vs.90 1. (2000) reported an increased Maillard reaction with increased salt content in anhydrous systems. The salt content of MW1 was significantly higher than for the other samples. Processor1 Actual2 k × 103 Estimated3 k × 103 Estimated4 k/actual k Q10 Ea5 WC1 WC2 MW1 3.1641 MAILLARD BROWNING IN SWEET WHEY POWDER Table 4. The most commonly implicated Maillard reaction inhibitor is sulfur dioxide. k (Table 4).00.35 3. moisture content increased by 32% for MW1. An increase in a* represents an increase in red. respectively. 88. 5. No sulfite was detected in the sample (Table 2). It is well known that certain salts and buffers have an accelerating effect on Maillard reactions in solution (Saunders and Jervis. Q10.1 1. Rate of brown pigment formation (k).0 to 8.12 4. it had the smallest Q10 implying that for the same increase in temperature. 1981). Activation energy values of 20 to 40 kcal/mol have been reported by Labuza and Saltmarch (1981) in food systems.0 reported for brown pigment formation for a number of food systems (Labuza and Saltmarch. a decrease in L* represents an increase in darkness (Hutchings. 3 Rate of brown pigment formation (k) determined from accelerated shelf-life testing.14 28.5 39. storage time for the 3 whey samples gave straight lines as supported by a high correlation to zero order. WC2 had a Q10 of 1. 1981) The higher color units/g of solid in WC2 under high temperature storage could be due to its high mineral content. which may explain the higher level of browning for this sample when stored under normal storage conditions (Table 4). which is below the range of 2.82 and 1.0 14. The estimated k values were 1. and 1.12. WC2. 45% for WC1. but this does not exclude the possibility of sulfur dioxide addition. and MW. 4. Bell.41 1 WC = Processors from West coast. Sample WC1 did not change in a* parameter. During the 19-mo storage period. WC2. MW1). 1994). Sample WC1 was the most stable whey powder with the highest Ea and the lowest rate of deterioration as measured by the rate of Maillard reaction.2 20. as it is not readily recovered following reaction (McWeeny.77 4. WC2 had the highest calcium content although the salt content was similar to WC1 (Table 2). and Journal of Dairy Science Vol. 1966.35 times bigger than the actual k values for WC1. Burin et al.1 8.00 1. 1997). and activation energy (Ea) for 3 sweet whey powders (WC1. 5 In kcal/mol. Sample WC1 also had the lowest free moisture content (Table 1). Color units/g of solid was calculated from the standard curve. Labuza and Saltmarch (1981) found that an increase in water content above the monolayer value dissolves and mobilizes reactant species for the Maillard reaction. Physicochemical and Microbiological Analyses Principal component analysis of the microbiological and physicochemical properties of sample MW1 stored under normal storage conditions (Figure 3) showed that Hunter Laboratory values L* (lightness-darkness parameter) and a* (redness-greenness parameter) provided optimal sensitivity for detecting changes in whey powders during storage. Equations and respective R2 values at the different temperatures for the various processors as determined from the graphs (Figure 2) are summarized in Table 3 and confirm that the Maillard reaction follows zero order in whey powder.

20a 0.55bc +22.11 0.34c +24.12a 0.65a 1.45e 6. 88.57a +22.84c +22.11a −0.72b −0.80a +1. 2 Titratable acidity determined as % lactic acid.90d 90.10b 6.15b 0. 2005 a b* SPC5 (cfu/g) MSC6 (cfu/g) Y&M/ Coli7 (cfu/g) +21.31 1.32b 1.34b 1.27a −0.2b 0.73b −0. Table 5. 6 MSC = Mesophilic spore count.12b +21.11a 0.1 <0.21c −0.99ab +22.1a 0. SPC = standard plate count.12b 5.25b 0.31c <10 <10 <10 <10 <10 400b 200a 260a 230a 150a 220a 1910c 2480d 4060e 1070b 180a ND8 ND ND ND ND 70a 85a 90ab 150b 90ab 85a ND ND ND ND ND <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 <10/<10 .89b +22.15b 6.2a 6.50b +21. 3 SI = Solubility index.81b 90.06a 6. dotted line separated on PC1. No.35c 6.10c 0.10a 0. PC2) on physicochemical and microbiological data for Midwest whey processor (MW1).42b 90.3c 0.15a 91.27b −0. b* = Hunter Lab blue-yellow parameter.69a 90.11 +1.01ab +21.12a SI3 (mL) pH b 6.12a 0.44e 6.08a 6.27b 0.80a 2.66c 90.1642 SITHOLE ET AL.22c 91.66b −0.31b 6.18b 6. molds.3c 0. Microbiological and physicochemical properties of whey powder stored for up to 19 mo at 21°C and 35% RH.64a 1. L* = Hunter Lab lightness-darkness parameter.25a +1.40c Means within columns with different letters are statistically different (P < 0.67c 90. 8 ND = Not detected.26bc 1.09ab 0.69a +22.13 6.00a 6.02ab 2.61a 1.25b 0.46a −0.09b 0. First 2 principal components (PC1.13a 0.25b 0. 1 Processor WC1 WC2 MW1 Storage time (mo) 1 5 9 10 19 1 4 5 7 10 19 1 5 9 11 19 TA2 (%LA) a 0.35d 90.17a +1. M = mo of storage at 21°C and 35% RH.94b +22.14a 1.12a 0.53b 89.87c +20.59b −0.05). 5.08ab 0.15d 90. WC = Processors from West coast. There was no significant difference in storage time in the same circle at 95% confidence interval.08a 0.26b 0. Figure 3.93b 2.86ab +22.97a 6.42a −0. 5 SPC = Standard plate count.63 91.92c 90.75a +24.20a 0.04c 1.86a 2.24a a % MC b 1. a-c 1 Journal of Dairy Science Vol.12a 0. Bold line separated on PC2.38d 91. 7 Y&M/Coli = Yeast.66a 1. and coliform bacteria. MW = processor from Midwest.90c 1.02b +21.82d 90.37c 4 L* a* e 91.09b 0.4d 6. a* = Hunter Lab red-green parameter.28b 0.43a −0. 4 MC = Moisture content (%).11a 0.09a <0.12a 0.22a +1.

95) 3.80) 2.13 (1.53) 1.23 (1.31 (1.46) 2.00 (1.90 (1.92 (1.44 (1.87) 2.77 (1. No.94) 3.49 (1.90 (2.07) 1.28 (1.21 (1.51 (1.41 (1.26 (1.04) 1.40) 2.542) 1.18 (2.91) 2.72 (1.48) 1.32) 1.43) 2.10 (1.92 (1.54 (1.03 (1.60) 3.92 (1.94) 1.60) 1.81) 4.59 (2.59 (1.46) 1.92) 3.26) 1.56 (1.36) 0.54) 1.11) 0.54 (2.42) 2.15 (1.74) 1.10) 1.46 (1.38 (1.91) 1.33 (1.91) 4.73) 2.38 (1.48) 1.87 (2.74 (1.56) 2.32) 0.64) 3.38 (1.10 (2.53) 2.44 (1.530) 2.41 (1.64 (1.05 (1.92 (1.89) 4.93) 2.10 (1.10) 2.25) 2.72 (1.86) 2.20) 3.30) 1.77 (1.41) 1.90) 3.60) 4.03 (2.77 (1.56 (1.28) 1.03 (1.00) 4.92) 2.33 (1.31) 0.22) 1.52) 1.90 (1.746) 1.38) 1.31 (1.12) 1.67 (2.20) 1.06) 4.97 (1.12) 2.44 (1.20) 3.46 (1.85 (1.73) 4.03) 3.72 (2.98) 2.10) 1.53 (2.05 (1.04) 2.49 (2.97 (1.35) 1.56) 1.46 (1.55) 3.07) 1.51) 1.46 (2.19) 1.28 (2.31 (1.62 (1.26) 2.97 (1.21 (1.08) 0.72 (1.97 (1.79 (1.64 (1.97 (2.48) 1.54 (1.34) 2.08) 3.99) 3.67) 1.15 (1.81) 1.959) 4.75) 4.61) 2.49 (1.64 (2.69) 2.16 (2.00 (1.98) 4.54) 2.72 (1. Storage time (mo) WC1 1 5 12 19 WC2 1 5 10 12 19 MW1 1 9 12 19 Cooked flavor Cardboard flavor Oxidized flavor Barny Sweet aromatic aroma Caramelized aroma Cooked aroma Cardboard aroma Oxidized aroma Sweet flavor Sour flavor Salty flavor 1.42) 2.15 (2.82 (1.59 (1.25) 3.18 (1.95) 2.95 (1.09) 3.10 (1.33 (2.34) 1.49 (1.38 (1.18) 1.23 (1.89) 4.18 (1.67 (1.55) 3.59) 4.46) 2.73) 1.67 (2.83) 1.77 (1.59) 1.31 (1.31 (1.31 (1.72 (1.79 (1.74 (1.41 (1. and 7 = moderate.32) 2.56 (2. 2 1643 Journal of Dairy Science Vol.72 (1.83) 4.82 (1.30) 1.21 (1.92) 2.51 (1.30) 2.17) 1.55) 2.51 (2.70) 1.33) 1.44 (1.29) 1.13 (1.92 (2.63) 1.92 (1.44) 1.13) 1.21 (2.Table 6.80) 4.20) 2.03 (1.41 (1.15) 1.38 (1.40) 2.76) 1.15 (1.82 (1.97 (1. 2005 5 Caramelized flavor MAILLARD BROWNING IN SWEET WHEY POWDER 9 Sweet aromatic flavor .91) 4.08 (1.36 (1.76) 3.36 (1.38 (1.64 (1.55) 1.65) 1.60) 3.59 (1.56 (1.59) 2.53) 1. rating scale was a 15-point scale.44 (2.10 (2.14) 2.10 (1.64 (1.90 (1.56) 4.36 (2.71) 1.31) 3.79 (1. MW1)2 stored for up to 19 mo at 21°C and 35% RH.39) 2.59 (1.18 (1.15 (2.51) 1 Mean of duplicate testing by 13 panelists.85 (1. WC = Processors from West coast.14) 1.58) 1.44 (1.61) 2.40) 3.36 (1.59 (1.01) 4.61) 1.17) 2.56) 1.23 (1.59) 1.34) 1.95 (2.67 (1.35) 1. where 3 = slight.31 (1.18 (2.30) 1.26) 1.15 (2.65) 1.87 (1.54) 1.36) 1.85 (1.65) 3.28 (1. Mean rating1 (± standard deviation) for sensory descriptors tested of sweet whey powder (WC1.23 (1.52) 1.18) 4.87 (1.03 (1.54) 1.49) 1.87 (2.85 (1.54) 2.79 (1.45) 2.74 (2.10 (2.29) 1.15 (2.79 (1.41) 1.00 (2.84) 1.90 (1.41 (1.54) 1.45) 2.21 (1.28) 4.03 (1.85) 3.41 (1.72 (2.00 (1.05) 4.59) 1.31 (1.67) 3.69 (1.49 (1.41 (1.05 (2.26 (1.69 (1.37) 4.79) 3.48) 1.11 (2.74 (1. MW = processor from Midwest.62 (1.10 (1.22) 1.58) 1.82 (1.62 (1.98) 4.15 (1.13) 1.85 (2. WC2.52) 1.25) 0.20) 4.87) 1.12) 2.69 (1.51 (1.44) 1.41) 1.54 (1.07) 3.70) 3.00 (1.21 (1.79 (2.85) 2.10 (1.02) 2.59 (2.79) 2.95 (1.72 (1.14) 1.97 (1.14) 4.41) 1. 5.32) 2.13) 2.10 (1.95 (1.27) 1.17) 1.40) 0.69) 2.67) 1.26 (1.44 (1.52) 2.67) 1.44 (1.10) 3.08) 4.48) 3.26) 1.95 (1.78) 4.56) 1.63) 1.48) 1.08 (2.97) 3.42) 2.74 (1.38) 4.41 (1.41) 1.31 (2.33) 1.10) 1.23 (1.89) 1.23) 1.54 (1.03 (1.49 (2.54 (1.40) 2.23 (1. 88.11) 1.85 (1.21 (1.00 (1.

Ca¨mmerer. Agric. Rajalakshmi.1644 SITHOLE ET AL. 88. 1995.. Karagu¨l-Yu¨ceer. Piecky. O’Malley. Nonenzymatic browning reactions of retro-adol degradation products of carbohydrates. Copenhagen. Food Process. 1997. Gaithersburg. and L. 1949. K. R. and M. 1999. Y. 2000. L. such as hygroscopicity and caking. The Maillard reaction has been reported to lead to flavor and aroma changes during storage. F. Official Methods of Analysis. P. The mean rating and standard deviation for each of the descriptors tested are given in Table 6. which is the typical shelf life reported by commercial suppliers. ed.. Singh. Choi. 49:2948–2953. and P. J. 5:49–57. De Forenede Trykerier A/S. Jouppila. 1991. J. B. Buera. L. Hutchings. E.. 1999. 17th ed. (1990).. P. L. 5. 1983. Gaithersburg. as no standard plate count or mesophilic spore count was detected. K. molds. J. Because free moisture content changed significantly with storage at ambient temperature.. the flavor and aroma of all whey powder samples remained constant. ACKNOWLEDGMENTS Financial support was provided by Tillamook County Creamery Association. Furthermore. Burin. even the most rapidly deteriorating whey powder. L. Nutr. 4th ed. A proposed method for the determination of color of dry products of milk. B. Variations in moisture gain might be attributed to differences in water vapor transmission rate of the packaging material and to differences in hygroscopicity. Nonenzymatic nonoxidative browning in hydrolyzed shelf-stable concentrated cheese whey. J. N. and S. Standard plate count decreased during storage for WC2 and MW1. Influence of the content of water-soluble vitamins. and M. The deterioration rates of the sweet whey powders from the 3 processors were different. Hurrel. A/S Niro Atomizer. 16th ed. Food Res. 1994. parameters related to moisture content. and D. AOAC. Decreases in whey pH as browning progresses were reported by Pilar Buera et al. 49:355–364.. 32:580–586. No yeast. Kinetics of nonenzymatic browning in dried skim milk. and WC2 had a 2-fold decrease (Table 5). 2005 at elevated temperatures. Determination of moisture. REFERENCES American Dairy Products Institute. Sample MW1 exhibited a 10-fold decrease. Food Chem. 14:303–319. J. Franzen. 2001. Official Methods of Analysis. Pages 10–11 in Analytical Methods for Dry Milk Products. 2. Pages 7–19 in Whey and Whey Products: Definitions. M. Sens. M. M. Blackie Academic & Professional. MD. Resnik. J. Ravi. as determined by the Maillard reaction. It appears likely that whey powders deteriorate because of loss of functional properties such as flowability and dispersibility before any changes in taste and aroma are observed. No information was available regarding differences in processing parameters at the 3 manufacturing plants. Kansikas. Other causes for variation in deterioration rate of the whey powders might be related to differences in processing. J. The sample with the fastest rate of deterioration. as determined by the Maillard reaction. A. Technol. Association of Official Analytical Chemists International. 59:143–147. The microbiological quality of WC1 was excellent. the only sample that contained mesophilic spores. Sensory profiling and positioning of commercial samples of milk powder. W. Kim. Sensory Evaluation Even though the sensory panel was highly trained. Y. J. I. No. R. and V. American Dairy Products Inst. UK. J. was found to deteriorate the fastest as determined by the physicochemical parameter (Hunter Laboratory color parameter L*). Glasgow. Kamath. M. 1981. the rate at ambient conditions for the 3 samples was too low to cause any significant difference during storage up to 19 mo. Drake. However. 11:225–239. Ford. 1978. Finot. Br. Food Sci. R. ed. 1990.. J. Fairbanks. and T. 24% for WC2. Haugaard Sorensen. Maillard reaction as influenced by buffer type and concentration. Cadwallader. should be explored in shelf-life studies of whey powder. 55:697–700. . However.. 48:5263–5268. Chirife. and B. CONCLUSION Our data suggest that whey powder has a shelf life longer than 12 mo. J. Saltmarch. Krag. no significant differences were found in odor and flavor of the whey powders during 19 mo of storage. Labuza. Association of Official Analytical Chemists International. R. Bell. Westergaard. Denmark. Bulletin W-16. Dairy Sci. Standard Methods of Analysis. Nonenzymatic browning of hygroscopic whey powders in open versus sealed pouches. K. Color formation in dehydrated modified whey powder systems as affected by compression and Tg. differences in the gain in free moisture could be a result of the lactose crystal formed as bound water can be released during the transformation among amorphous and crystal lactose as well as the different lactose crystalline forms. and K. Page 221 in Instrumental Specification. Kroh. Chicago. there was no significant change in the mesophilic spore count with storage for WC2. J. Stud. Storage of milk powders under adverse conditions. 209:261–265. J. AOAC.. C. Agric. B. J. A. R. M. Preserv. However. B. Composition. in this study. Eur. Wedzicha. 1990. IL. del Pilar Buera. Roos. Food Color and Appearance. R. Food Chem. Food Eng. and the Eckelman Foundation. J. Aroma active components of nonfat dry milk. Although the Maillard reaction plays a major role in determining the rate of deterioration of whey powders Journal of Dairy Science Vol. still had flavor and aroma perceived to be similar to that after 1 mo of storage. Okos. Food Chem.. MD. A decreasing trend in pH was observed for all 3 samples. 2000. W. or coliform bacteria were detected in any of the samples. Hutchings.

Int.. and F. Prog. Accelerated shelf life testing of an intermediate moisture food in air and in an oxygen-free atmosphere. J. 1983. D. Electron Microsc... Saltmarch. 1995. H. G. 1981. 17:245–249. T. L. Rivero-Urgell. and M. MD. H. and T.. Soc. Lawless. B. McWeeny. Jervis. Stapelfeldt. J. Boston... 15th ed.. 47:92–96. Academic Press. T. H. 1981. Dairy J. P. H. St. Kinetics of browning and protein quality loss in whey powders during steady state and non steady state storage conditions. Richardson. Joseph. Inc. 1981. No. Saunders. Food Sci.. SEM investigation of the effect of lactose crystallization on the storage properties of spray dried whey. Skibsted. R. M. 5. Effect of processing on kinetics of nutrients and organoleptic changes in foods. H. P. and H. 7:341–348. and T. Lo´pez-Sabater. Waletzko. Shelf-life prediction of an infant formula using an accelerated stability test (Rancimat). P. Influence of relative humidity on the physicochemical state of lactose in spray-dried sweet whey powders. and M. DC. Agric Engr. J. and T.. MA. 1999. Food Nutr. Food Chem. and L. MI. Heymann.. 1985. 41:1338–1344. Labuza.. Food Sci. J.MAILLARD BROWNING IN SWEET WHEY POWDER Labuza. Hawkes. Principles and Practices. Labuza. 1980. Washington. Agric. Stone. Labuza. The role of buffer salts in nonenzymatic browning. M. J. 3:659–665. and J. 88. Standard Methods for the Examination of Dairy Products. Food Agric. 43:2879–2882. Sci. Villota. 45:1231–1236. Journal of Dairy Science Vol. S. Food Sci. Nielsen. Paper presented at the Winter Meeting of the Am. 1976. Sensory evaluation practices.. Sci. 1966. G. 1993. Presa-Owens. J. American Public Health Association. 1997. Frederick. Inc. 5:395–404. 1645 Saltmarch.. J. Sulfur dioxide and the Maillard reaction in food. M. Pages 358–362 in Sensory Evaluation of Food. Scan. Early prediction of the shelf-life of medium-heat whole milk powders using stepwise multiple regression and principal component analysis. P. Sidel. Saltmarch. and J. R. Aspen Publishers. 2005 .