WATER ACTIVITY OF SKIMMED MILK POWDER IN THE TEMPERATURE RANGE OF 20 – 45 ° C

·tencl J . : Water Activity of Skimmed Milk Powder in the Temperature Range of 20 – 45 °C. Acta Vet. Brno 1999, 68: 209–215. Water activity data for adsorption and desorption of moisture from skimmed milk powder were investigated at temperatures in the range of 20 – 45 oC and moisture content of the material tested from 3.2 to 20 % (wet basis). The experimental procedure used was a gravimetric dynamic method with continuous registration of sample weight changes. Four mathematical models of sorption isotherms (Chung-Pfost, Halsey, Henderson, and Oswin) were evaluated to determine the best fit for the experimental data. The modified Oswin equation was a good model for moisture adsorption and desorption of skimmed milk powder. Water sorption capacity decreased as temperature increased. The critical value of equilibrium moisture content of milk powder tested, corresponding to the water activity equal 0.6, was 11 % (wet basis) at the temperature of 20 oC. Repeated rehydration of the material brought an increase in the original equilibrium moisture content 3.2 % (wet basis) to 6.3 – 8.1 % (wet basis) in relation to the temperature. It was also demonstrated that an increase in equilibrium moisture content was very small (about 4 % wet basis) in the range of water activity 0.1 to 0.9. Higher levels of water activity than 0.9 resulted in a marked increase of equilibrium moisture content and susceptibility to spoilage by microorganisms. The hysteresis effect between moisture adsorption and desorption was insignificant. Adsorption, desorption, microbial food stability, milk powder, modeling, water activity The moisture sorption isotherm is an extremely valuable tool for food scientists and technologists because it can be used to predict potential changes in food stability; it can be used for storing method determination, packaging selection and ingredient selection. The moisture sorption isotherms of foodstuffs show usually the equilibrium relationship between water activity (aw) and moisture content (MC) of the food at constant temperatures and pressures. A critical aw also exists below which no microorganisms can grow (Beuchat 1981). For most foodstuffs, this is in the range of 0.6-0.7 aw. In general, dehydrated foods have aw’s less than 0.6; semi-moist foods, such as cereal grains, raisins, dates, syrups, and intermediate-moisture pet foods usually have aw between 0.62 and 0.92. Cheeses, jams, jellies, meat, fish etc. have aw’s greater than 0.92. Thus, with knowledge of the moisture sorption isotherm, we can predict the maximum moisture that the food can be allowed to gain during storage (Kiesl ingerová and Bart l 1993). Of course, higher aw’s can be allowed if other factors such as pH, salt, antimicrobial agents, and temperature are taken into consideration. Table 1 lists minimum aw values for growth and toxin production by pathogens (Beuchat 1981). Generally, temperature has important influence on aw. Investigations of water adsorption/desorption isotherms have been the subject of study for numerous products due to the development of modern techniques for their processing and storage. Procedures for obtaining water sorption isotherms in foods were described in detail by ACTA VET. BRNO 1999, 68: 209–215 Address for correspondence: Doc.Ing. Jifií ·tencl, CSc. Department of Postharvest Technology and Engineering Mendel University of Agriculture and Forestry Zemûdûlská 1, 613 00 Brno, Czech Republic Phone: ++420 5 45132116 Fax: ++420 5 45212044 E-mail: stencl@mendelu.cz http://www.vfu.cz/acta-vet/actavet.htm

The moisture sorption isotherm is an extremely valuable tool for food scientists and technologists because it can be used to predict potential changes in food stability; it can be used for storing method determination, packaging selection and ingredient selection.The moisture sorption isotherms of foodstuffs show usually the equilibrium relationship between water activity (a w ) and moisture content (MC) of the food at constant temperatures and pressures.A critical a w also exists below which no microorganisms can grow (Beuchat 1981).For most foodstuffs, this is in the range of 0.6-0.7 a w .In general, dehydrated foods have a w 's less than 0.6; semi-moist foods, such as cereal grains, raisins, dates, syrups, and intermediate-moisture pet foods usually have a w between 0.62 and 0.92.Cheeses, jams, jellies, meat, fish etc. have a w 's greater than 0.92.Thus, with knowledge of the moisture sorption isotherm, we can predict the maximum moisture that the food can be allowed to gain during storage (Kieslingerová and Bartl 1993).Of course, higher a w 's can be allowed if other factors such as pH, salt, antimicrobial agents, and temperature are taken into consideration.Table 1 lists minimum a w values for growth and toxin production by pathogens (Beuchat 1981).
Generally, temperature has important influence on a w .Investigations of water adsorption/desorption isotherms have been the subject of study for numerous products due to the development of modern techniques for their processing and storage.
Procedures for obtaining water sorption isotherms in foods were described in detail by Wolf et al. (1990), T roller andChristian (1978) and Gal (1975).The principal methods are gravimetric, manometric and hygrometric.The gravimetric method is the most common type of sorption test.It is possible to obtain MC changes of samples continuously or periodically using a static system (usually a closed jar containing saturated salt solutions or sulphuric acid solutions which give a certain equilibrium relative humidity) or a dynamic system (circulated air with a constant flow rate).The dynamic system with continuous registration of weight changes is technically more complicated than the static one but the flow of air around the sample makes the wetting and drying processes faster.Moreover, this system gives better results in cases of layered materials (Labuza et al. 1976).Numerous models for predicting the relationship between equilibrium moisture content (EMC), a w and temperature have been developed.Iglesias and Chirife (1976) reviewed several equations for modelling equilibrium MC and reported that some models were adequate to characterize the sorption behaviour of particular foods for the given range of temperature and a w or relative humidity (r.h.).Chen and Morey (1989a) evaluated four models (Chung-Pfost Halsey, Henderson and Oswin) for their ability to fit data from 18 grain and seed crops.The modified Henderson and Chung-Pfost equations were good for fibrous and starchy materials while the modified Halsey fitted well for high oil and  Many other agricultural products have been investigated from the aspect of MC/a w (or equilibrium relative humidity, ERH) relationship, e.g.rice (Banaszek andSiebenmorgen 1990 (Pixton andWarburton 1975), onion (Mazza and Le Maguer 1978), figs (Pixton and Warburton 1976), apple (Resnik and Chirife 1979), walnuts (Vaidya et al. 1977), malt (Pixton and Henderson 1981).The Chung-Pfost, Halsey, Henderson, and Oswin models are commonly used to describe the sorption behaviour of a wide range of biological materials.
The objective of this study is to determine the effect of temperature on the moisture adsorption and desorption isotherms of skimmed milk powder in the temperature range of 20 -45 o C, to analyze four sorption isotherm equations available in the literature and to determine a model corresponding to the isotherms measured.Furthermore to establish critical a w and to carry out microbial analysis of the material tested.

Materials and Methods
A fully computerized laboratory drying device with special control software was developed for the purpose of sorption tests (•tencl et al. 1995).The apparatus consists of two main function parts: an air duct with electronically controlled temperature, velocity and relative humidity, and an electronic balance.
Tested samples of skimmed milk powder were taken directly from a spray dryer; the quality was as follows (Table 2): Samples were without coliform organisms, Staphylococcus, Salmonella, neutralizing agents and free from antibiotics.
Moisture equilibrium data for adsorption and desorption of water from skimmed milk powder were investigated at temperatures in the range of 20 -45 o C in 5 o C steps and a w ranging from 0.4 to 1.0 in 0.1 steps.The procedure of each of the tests was as follows: after reaching the equilibrium moisture content (EMC) of the sample at a certain a w (at a constant air temperature, velocity and pressure), the a w was automatically increased (adsorption) or decreased (desorption) and a new equilibrium was obtained under these conditions.Each test was repeated three times with material of the same sampling.
The experimental EMC data were processed using the specially developed software and analyzed using the nonlinear regression procedure of UNISTAT (1995).Four equations (Madamba et al. 1994;Chen and Morey 1989a) describing relation between EMC and ERH were evaluated for their ability to fit data for skimmed milk powder:

Results
Equations ( 1) -( 4) to model the dependence of EMC of skimmed milk powder on a w in the temperature range of 20 -45 o C were investigated and reviewed.Analysis of residuals and goodness of fit tests were carried out after parameter determination.The comparison of Chung-Pfost, Halsey, Henderson and Oswin models is given in Table 3.
The following statistics were compared: SEE -standard error of estimate, P -mean relative percentage deviation, d -Durbin-Watson statistic (significant points d L = 1.63; (d U = 1.72) (Draper and Smith 1981), chi-square (Hanke and Reitsch 1991) and R 2 -coefficient of determination.Best results have been received with the Oswin's equation ( 4), Table 3.
Parameter estimation for the Oswin's model of EMC for skimmed milk powder both for adsorption and desorption is indicated in Table 4.
Figures 1 and 2 were generated to illustrate adsorption and desorption models of skimmed milk powder given by equation (4) using fitted parameters in Table 4. Significant quality changes in the samples tested have not been found after realization of sorption tests.An increase of MC to the level 8.1% (wet basis) at 20 o C was important if compared with the original material (Table 2).Further quality parameters were equal.The value of a w was approximately 0.1 at the final moisture content, i.e. a risk of growth of molds and yeast does not exist.The critical value of EMC was about 11% (a w = 0.6) at the temperature 20 o C.

Discussion
Most biological products follow a sigmoid curve representing the type II isotherm BET classification (Labuza 1984).The resulting curve is caused by the additive effects of Raoult's law, capillary effects, and surface water interactions.Two bends are noted, one around an a w of 0.03 and other at 0.95.These are the result of changes in the magnitude of the separate physical-chemical effects.
Part of sorption isotherms measured shows the type II BET classification shape.An increase in temperature causes an increase in water activity for the same MC and, if a w is kept constant, an increase in temperature causes a decrease in the amount of absorbed water.This indicates that the material becomes less hygroscopic at higher temperatures.These effects are considerable especially up to water activity (0.95 both for adsorption and desorption (Figs 1 and 2).The diagrams show that the repeated rehydration of the material tested brought an increase of original EMC 3.2% (wet basis) on the level of 6.3 -8.1% (wet basis) in relation to the temperature.This fact is important especially from the aspect of packaging and long-term storing of the skimmed milk powder.It has been further demonstrated that an increase of EMC was very small (about 4% MC wet basis) in the range of a w from 0.1 to 0.9.Higher levels of a w than 0.9 resulted in increase of EMC and susceptibility to spoilage by microorganisms.As mentioned in Materials and Methods, the moisture sorption curves have been generated from adsorption and desorption processes (Figs. 1 and 2) but the hysteresis effect was insignificant.
protein materials.A study by M a z z a and Jayas (1991) revealed that the Guggenheim-Anderson-De Boer (GAB) model was superior to three other models (Chung-Pfost, Halsey and Henderson) in characterizing the sorption behaviour of sunflower seed, hulls and kernels.A recent study by •tencl et al. (1998) concluded that the modified Henderson equation characterized very well the moisture adsorption and desorption of dried blood flour in the temperature range of 20-50 o C.

Chung
, C = constants for the particular equation.

Table 3
Comparison of Chung-Pfost, Halsey, Henderson and Oswin models for adsorption and desorption of skimmed milk powder