JournalofFoodEngineering25 (1995) 531-552 Copyright Q 1995 Elsevier Science Limited Printed in Great Britain. All rights reserved 0260-X774/95/$9.

50 ELSEVIER 0260-8774(94)00033-6

A Critical Review of Some Non-equilibrium Situations and Glass Transitions on Water Activity Values of Foods in the Microbiological Growth Range
Jorge Chirife & Maria P. Buera*
Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad Buenos Aires, 1428 Buenos Aires, Argentina 1993; revised version received 20 April 1994: accepted 18 June 1994) de

(Received 24 November

ABSTRACT Some facts concerning the effect of non-equilibrium situations and glass transitions on the water activity (A3 values of foods in the microbiological growth range are reviewed. Analysis of literature data suggests that although some amorphous intermediate moisture foods are non-equilibrium systems (e.g. very slow rate of attainment of equilibrium), the stability (time dependency) of measured water sorption isotherms, or the prediction of water activity from equilibrium data corresponding to aqueous solutions of main food solutes (NaCl, sugars), are little affected. However, crystallization of amorphous components (sugars), which is also a consequence of non-equilibrium situations, may lead to significant modifications of A, values in foods. As far as the prediction of microbial growth in foods is concerned, changes in non-equilibrium semi-moist foods may be in many situations slow (within the timeframe of the food shelf-life) and/or so small that they do not seriously affect the application of A, concept as a predictor of microbial growth. However, the use of A, should always consider the possible influence of some non-equilibrium situations.

INTRODUCTION The solids of many foods (either biopolymers or low molecular weight carbohydrates) are often in an amorphous state which is sensitive to changes in temperature and moisture content (Slade & Levine, 1987; Roos & Karel, *Member of CONICET,
Argentina.

531

532

J. Chirife, M. P. Buera

199 1a). This amorphous matrix may exist either as a very viscous glass or as a more liquid-like rubbery structure, and the change from the glassy state to the rubbery state occurs at the glass transition temperature ( T ), which is specific for each material (Slade & Levine, 1987, 1991a, b; Roos & Karel, 1991 a, b). The glass transition temperature is decreased with increasing moisture content, because water plasticizes the amorphous structure (Slade & Levine, 1987; Roos & Karel, 1991 b) of food systems. Over the last decade, it has been suggested that certain food products (mainly of the low and intermediate moisture types) are non-equilibrium systems (Slade & Levine, 1987, 199 1 b; Franks, 1991 a, 6) and that during processing, such products can be maintained or brought into a state of thermodynamic metastability, or better designed pseudostability, since this condition may last longer than the products normal lifetime (van den Berg, 1991). This approach focuses increasing attention on the dynamic aspects of foods, rather than on equilibrium properties (Slade & Levine, 1991 b). Franks ( 1991 a) stated that, in many product situations, equilibrium thermodynamic descriptions are inappropriate, because the measured physical properties are time dependent; thus, water activity (A,) must not be used to describe the attributes of these systems. The measured A, relies on the existence of a liquid/vapor dynamic equilibrium, which is assumed to be re-established rapidly when one of the experimental variables (temperature, composition) is changed. This situation may not occur when the substrate is a quasi-solid dispersion such as a low moisture or even a semi-moist food product. In the latter situation, the high viscosity of the food system is related to a much slower rate of attainment of equilibrium; what is measured in such cases is a non-equilibrium relative vapor pressure (RVP) and not A, except for systems which are subject to rapid diffusion. van den Berg ( 1991) pictured such foods as multi-component, multi-phase systems which, at high enough moisture levels, are partially dissolved and in a gel condition. With the undissolved macromolecules constituting the solid phase and holding a liquid or semi-liquid phase, dispersed or continuous. He suggested that, in these systems, true equilibrium may not exist, but rather some pseudo equilibrium which is not clearly definable, but which may remain for months in that condition. As a consequence, the thermodynamic water activity cannot be defined throughout the system, but rather an empirical pseudo water activity which is reflected in the system properties. In spite of the above criticisms, it has been widely recognized that the concept of water activity has been of great importance in food preservation, because measured values generally correlated well with the potential for growth and metabolic activity of microorganisms (Gould, 1985, 1988; Gould & Christian, 1988; van den Berg, 1991). Its measurement has been a valuable tool for predicting the microbial stability (and safety) of foods (Leistner & Rodel, 1975; Leistner et al., 1981; Troller & Christian, 1978; Silverman et al., 1983; Dodds, 1989; Glass & Doyle, 1991). The question that may arise is why, if the concept of A, has been misleading, this parameter (A,) has been such a helpful tool in predicting microbial food stability. Franks (1991~) acknowledged that a measured relative vapor pressure, although not usually equal to A,, can serve as a useful but not the sole indicator of microbial safety. For these reasons and regardless of the rigorous theoretical meaning of any measured water activity in

Wateractivityvuluesoffoods

533

foods (e.g. a thermodynamic activity, a non-equilibrium RVP, or an empirical pseudo water activity), it is important to analyse the practical consequences of these non-equilibrium situations on water activity values of foods in the microbiological growth range. This is the objective of the present review. Equilibrium in water sorption measurements The main criticisms of the utilization of the term water activity in the food area may be summarized as follows. In low moisture foods and intermediate moisture systems, the concept of A, becomes meaningless because the measured vapor pressure of water is no longer the equilibrium vapor pressure: at best, a stationary state is reached under a given set of environmental conditions and mistaken for equilibrium. A recorded low value for water vapor pressure is likely to be due to the inability of water to diffuse rapidly (within the observation period) through the substrate and to equilibrate with the vapor above it (Slade & Levine, 199 1 a; Franks, 199 1 a). In moisture sorption studies. the situation may be further complicated if the amorphous material undergoes a glass-rubber transition during the course of the measurement (Franks, 199 1 a), since the rates of diffusion limited changes (e.g. water equilibration) are much more rapid in the rubbery liquid state than in the glassy solid state (Slade 8r Levine, 1991 a). It is important to analyse literature sorption data on the basis of these statements.
Precision in the determination of sorption isotherms

In order to evaluate literature data, it is first necessary to have an approximate idea of the usual precision in the experimental determination of sorption isotherms. In a comprehensive collaborative study within the framework of the European Cooperation in the Field of Scientific and Technical Research (COST ), precision data (e.g. repeatability/reproducibility) in the determination of sorption isotherms were determined. In a study with 24 laboratories, the following precision data were obtained for the adsorption isotherms of microcrystalline cellulose (MCC) and potato starch. In a water activity range of interest to microbial growth (e.g. 0*60-O-90), the average standard deviation of all data was 2.6% of the equilibrium moisture content for microcrystalline cellulose and 3.8% for potato starch (Wolf et al., 1984). The repeatability (this refers to tests performed at short intervals at one laboratory by one operator using the same equipment) was 2.0% for both MCC and potato starch. The reproducibility (this refers to tests performed at different laboratories which implies different operators and different equipment) was 8.2% and 12.9% for MCC and potato starch, respectively. On line with these results, Macchia and Bettelheim (1964) reported that the repeatability of their isotherm data for amylose was 2%; Andrieu et al. (1985) also reported a repeatability of 2% for their isotherms of durum wheat semolina and spaghetti. Jayas and Mazza ( 199 1) studied the variability among replicates in the determination of sorption isotherms of safflower seeds (in the range of A, O-1l-0.96 at 10, 25, 40 and 55C) and reported that the coefficient of variation [(standard deviation/mean equilibrium moisture content) X 1001 varied between 2% and 5% for 35 data points.

534 Equilibrium

J. Chirife, M. P. Buera time and criterion for equilibrium weight

From the point of view of the present review, the definition of the equilibrium between food samples and the water vapor source is a critical aspect in performing sorption experiments. Although theoretically this state can be in some cases expected after a very long period of time, practically the sorption process could be stopped when the weight difference left in the final equilibrium becomes less than the sensitivity of the balance used. But even this stage may require a very long waiting time and therefore an apparent equilibrium state is commonly defined in terms of a maximum tolerable weight change during an arbitrary time interval (Gal, 1981). This is the most usual procedure followed in the literature. It is important to note that the equilibration time in the determination of the isotherms will depend not only on water diffusion characteristics in the substrate, but also on mass transfer characteristics in the head space over the sample. In other ways the heat and mass transfer within the equipment used to determine the isotherms (e.g. desiccator) may or may not be lower than that in the product itself (King, 1968). For example, the COST project determined (Spiess & Wolf, 1985) that at low pressures the equilibration times were shorter than at higher pressures because of reduced transport resistance. At pressures below 130 mbar, the equilibration times can be reduced by a factor of two to three as compared to test under normal pressure, irrespective of the relative humidity prevailing in the head space. Of course, equilibration times will also depend on the thickness of the sample (Spiess & Wolf, 1985). Lomauro et al. ( 1985) reported on a study to test an objective criterion for determining the equilibrium moisture content, when sorption isotherms are prepared. Foods of different physical characteristics were exposed to an atmosphere of 752% relative humidity (RH) at 25C in non-evacuated desiccators, and the moisture content recorded as a function of time. Foods included flour, freeze-dried apple, freeze-dried turnip, freeze-dried ground beef, oatmeal cookie, shredded wheat and raisins. These authors considered that a pseudo equilibrium was reached when the moisture content (dry basis) did not change more than 0.5% during three consecutive sampling periods at not more than 7 day intervals. This criterion for equilibrium moisture content was compared to the values obtained after 6 months storage in closed Manson jars (over constant RH of 75.2%), which were considered to be very close to the equilibrium moisture content. They concluded that the criterion of no more than 0.5% dry basis moisture difference over three successive readings taken at 1 week intervals seemed satisfactory, and noted that the foods tested reached equilibrium (or very close to) within 1 month, based on the above criterion. Also the values compared well (differences between 0.5-3.4%) with the 6 month storage study. It is important to note that the differences between the 1 month and 6 month studies are in the range of the repeatability/reproducibility values reported for sorption isotherms, as noted previously. Various authors reported their equilibrium times for isotherm determinations (using the gravimetric, static method over saturated salt solutions). The data are summarized in Table 1. As noted, most authors utilized a small amount of sample (l-3 g) to diminish water transport resistances in the substrate, and their equilibrium times ranged mostly between 1 and 3 weeks. Some authors also reported the criterion used for equilibrium time. For example, Jayas and Mazza (1991) determined the sorption isotherms of

TABLE1
Static

Some Equilibration
Temperature (XI Reference Equilibration time (days)

Times Reported

in the Literature for the Determination of Sorption Values in Foods Using the Gravimetric, Method over Saturated Salt Solutions*

Product

MUSS

05 1 3 1 1 2 ;:

3 5 R

$.
3 $ 8 4 3 &

Dried spaghetti Amaranth flour Peanut flakes Parboiled rice Soy protein concentrate Wafer biscuits Banana powder Corn flour Fish protein concentrate Cowpea flour Starch White beans Pea seeds Orgeat (freeze-dried) Fruits (various) Prunes (dried) and raisins Sultana raisins Apricots, figs and raisins Coffee (freeze-dried) 5-35 20-50 21-37 12.5 38 25-50 25-42 21 25 16-49 10-5.5 25 15-60 25 20-35 20-36 20-30 21 10-14 10-14 14 2-7 4 14 14-21 7 10-12 7 31 7-35 15-45 15 5-10 15 21-28 14

2 4 1 13 10 5

8 0.4 0.5 0.5

Lagoudaki et al. (1993) Valdez-Niebla et al. (1993) Hill & Rizvi (1982) Engels et al. (1987) Hansen (1976)+ Wedzicha & Qume (1983) Lima & Cal-Vidal ( 1983 j Kumar (1974) Rasekh et al. (197 1) Chhinnan & Beuchat (1985) Bizot (1985) Hutchinson & Otten ( 1984) Mazza & Jayas ( 199 1) Pidaga & Lafuente ( 1965) Tsami et al. (1990) Bolin (1980) Saravacos et al. (1986) Ayranci et al. ( 1990) Hayakawa et al. ( 1978)

*Exact values depended on relative humidity and temperature. +In evacuated desiccators. Agitation of air space in the sorption chamber. Sin evacuated sorption chamber (flask) agitated continuously.

536

J. Chirife, M. P. Buera

safflower seeds and weighed their samples every second day until the change of sample mass between two successive readings was less than 0.01 g. From their data it can be calculated that this mass change represents 1% of the moisture content at water activity 0.75 (25C). Tsami et al. (1990) determined the sorption isotherm of various fruits (between 15 and 60C) and the equilibrium was based on weights in two successive measurements taken at 1 day intervals not exceeding the accuracy of the balance ( -t O*OOO 1 g). Hutchinson and Otten ( 1984) determined the sorption isotherms of white beans and considered that the samples have reached equilibrium when the change in sample mass between several successive measurements (every 3 days) was 0.01 g or less. A change of 0.01 g in their samples was calculated to represent a change of 05% in the total water content at 75% relative humidity (at 16C). Rapusas et al. (1993) determined the moisture desorption of raw onion slices and reported that the equilibrium moisture content was reached when two successive daily weighings gave a difference of not more than 0.01 g. It was calculated from their data that this represents a change in the total moisture content of the sample at 75% RH (at 30C) of about 1.5%. S6 and Sereno (1993) determined the sorption isotherms of quince jam and considered that equilibrium has been reached when all samples showed a change of 0.4% or less in two successive weekly weighings. Biquet and Labuza (1988) determined the moisture sorption isotherm of chocolate. Equilibrium was assumed to be reached when the change in weight expressed on a dry basis did not exceed 0.1% for three consecutive weighings at no less than 5 day intervals. The data reported by Bizot et al. (1985) for starch are noteworthy to the purposes of the present analysis. They utilized a practical equilibration time of about 7 days ( IL 0.02% water per 24 h) for a 1 g sample but also stored their starch samples over saturated salt solutions for two years. They noted a slow drift of desorption pseudo equilibria but it was only 1% water on a dry basis after such a long time. On the other hand, adsorbing samples remained stable. van den Berg ( 199 1) reported that a O-15 g sample of starch in a sorption balance under high vacuum exhibits equilibration times of O-5-3 days, producing states which are steady up to several months. Lewicki (1975) stored previously equilibrated freeze-dried parsley and leek in hermetically sealed containers for a period of 1 year and did not find relevant changes in the isotherms, which suggested that difficulties for attainment of the equilibrium were not important. From the above considerations it may be assumed that sorption determinations performed with the usual precautions regarding a practical weight constancy are not likely to be too far from equilibrium, and the differences are probably within the uncertainties associated with the experimental determination of isotherms, as previously discussed.
Sorption isotherms offoods containing crystallizable sugars

Although certain species of yeasts as well as molds will grow, albeit slowly, at A, 0.62-0.70, most fungi do not develop below a water activity of 0.70 (Pitt, 1975; Beuchat, 1983) and this value may be regarded as a practical upper limit for safe storage. For this reason, the practical microbiological growth range may be considered here to be 0.70 and higher. The so-called intermediate moisture foods have water activity values within this range (e.g. 0.70-0.90) (Leistner, 1987; Ferro Fontan et al., 1980; Chirife & Favetto, 1992). Dried fruits such as

Water activity values offoods

537

raisins, figs, prunes, etc. are good examples of non-equilibrium, intermediate moisture systems and may be used to challenge the applicability of the water activity concept. Sugars (glucose, fructose, sucrose) constitute between about 50-8351 of the dry matter in fruits (Tsami et al., 1990), and in the intermediate moisture range, they are likely to be in an amorphous unstable rubbery state when stored at microbial incubation temperatures (25/35(Z) (Roos, 1987, 1993; Chirife & Buera, 1994). Figure 1 compares the sorption isotherm data for raisins (glucose + fructose = 83% dry basis) in the intermediate moisture range, as reported by different workers. If we take into consideration the usual precision in the determination of sorption measurements, the agreement among all data may be considered reasonable, despite the fact that raisins constitute a non-equilibrium system with a high sugar concentration. Slade and Levine ( 1987) suggested, and Roos and Karel ( 1991~) confirmed that sugar (sucrose) crystallization is possible only above the glass transition temperature Tg and that the rate of crystallization depends on the ( T- I!) value. Since raisins at moisture contents (or A,) corresponding to the microbiological growth range and stored at 30C would be expected to have large ( T- YTg) values (Chirife & Buera, 1994; Roos, 1993), sugar crystallization is theoretically possible. Crystallization of amorphous sugars results in the release of moisture and this increases product water activity (Karel & Roos, 1993). However, the absence of discontinuities (Karmas et al., 1992) in the isotherm of raisins is noticeable, suggesting that sugars remained amorphous. This may be attributed to the presence of high molecular weight compounds in the raisins matrix, which delay (or inhibit) crystallization (Iglesias & Chirife, 1978; Karmas et al., 1992). Similarly, the sorption isotherms of other dried fruits reported in the literature, such as figs, currants, apricots and prunes, do not show discontinuities (Tsami et al., 1990) suggesting the absence of sugar crystallization, a non-equilibrium situation

130.87 120.8.$? t P 2 0 8 s 5 2 0 ?! 2 x B 110.8loo.& SO& 80% 70.8. 60.850.84o.c 30.820% IOB,, 0.81 0.3 I 0.4 =x 0.5 n x . Ii f v I1 x F I-1 * RAISINS

* I

0.6

0.7

0.8

0.9

Water activity

Fig. 1.

Water adsorption isotherms in raisins: comparison of several literature data. ( 1) Bolin (1980); (2) Saravacos et al. (1986); (3) Ayranci et al. (1990); (4) Tsami et al. (1990).

538

J. Chirfe, M. P. Buera

which, when it occurs, may alter moisture/A, values in the system. Bohn (1980) reported water activity values and equilibrium moisture content values for raisins sealed in glass jars and held at 21 or 32C for up to 12 months. His data are plotted in Figs 2 and 3; it can be seen that storing raisins in sealed jars for 9-12 months had little effect on the moisture isotherm, in comparison with the uncertainties associated with the experimental determination of sorption

60a a g i? ; s 5 c 8 g iii ._ I lo20304050-

RAISINS,

30

0-I 0.2

0.3

0.4

0.5

0.6

0.7

0.6

0.9

Water activity

Fig. 2.

Effect of storage time at 21C on the adsorption isotherm of raisins [from data reported by Bolin ( 1980)].

60RAISINS, 21 C
Initial time -V3 months t 6 months 12 months

ij : c: E E 8 El ; P

50.

40. _ 30-

20-

IO-

0, 0.2

0.3

0.4

0.5

0.6

0.7

0.6

0.9

Water activity

Fig. 3.

Effect of storage time at 30C on the adsorption isotherm of raisins [from data reported by Bolin (1980)].

Water activity values offoods

539

measurements in foods. This suggests that non-equilibrium effects (e.g. inability of water to diffuse rapidly) are very slow (in the time frame of these experiments) and/or small. Caiiellas et al. (1993) did not detect substantial changes in water activity (initial A, 0.61) of raisins stored at 4, 11 and 23C for up to 11 months. Samples stored at 11 and 23C showed an increase from O-61 to 0.65 A, after 11 months; the sample stored at 4C had least variation, and after 8 months storage, remained constant. At this point, it is important to note that any observed time dependency of measured RVP in semi-moist foods may not necessarily indicate the inability of water to diffuse rapidly within the food. It may be due also to chemical changes that may occur during the time of observation. Cafiellas et al. (1993) stated that in stored raisins, the sugars (glucose +fructose) are involved in Maillard reactions, and consequently a decrease in sugar content was observed during storage. This decrease in sugar content would influence water activity; moreover, during Maillards browning reactions, water is formed and this could also increase the A,,,. Since Maillard reactions have a strong temperature dependence (Hendel et al., 195 5; Resnik & Chirife, 1979; Buera et al., 1987), it is expected that at 4C reactions are more inhibited and water activity would remain almost constant, as was experimentally observed. Glass transitions in the microbiological growth range Moisture content profoundly affects the glass transition temperature in foods (Slade & Levine, 1987). The inability of water to diffuse rapidly in the food precludes attainment of equilibrium (Slade & Levine, 1991 a) and to some extent controls the apparent value of water activity. Thus, it is important to check whether or not amorphous foods undergo a glass-rubber transition during the course of water activity equilibration. This is because the rates of water diffusion would be much more rapid in the rubbery liquid state (governed by WLF kinetics) than in the glassy solid state (governed by Arrhenius kinetics) (Slade & Levine, 199 1 a). Recently, Chirife and Buera ( 1994) examined available Tg versus moisture content data for various foods (strawberry, potatoes, horseradish, carrots, cabbage) and food components (biopolymers). They found that at microbial incubation temperatures (e.g. 30C) they are in many cases in the rubbery state for most of the microbiological growth range. The behaviour of Tg versus water activity for some plant materials (carrots, strawberries and potatoes) and biopolymers (wheat starch and gluten, gelatin) was calculated from literature data (Karmas et al., 1992; ROOS, 1987; Zeleznak & Hoseney, 1987; Hoseney et al., 1986; Slade & Finley, 1989; Iglesias & Chirife, 1982) as shown in Fig. 4. Plant materials are rubbery throughout the microbiological growth range. The T, values for biopolymers are much higher than in plants, although some of them also appear to be rubbery at incubation temperature in the water activity range of most practical interest regarding microbial growth. For starch, however, the glassy state extends to most of the important microbiological growth range. Although some biopolymers, such as starch, have high Tg values even at high values of water activity, it is likely that in real foods the presence of relatively small amounts of low molecular weight solutes (e.g. sugars) depresses the glass transition temperature, in such a way that the rubbery state (e.g. ( T- Tg> 0)) may prevail in most of the intermediate moisture range. If we accept that RVP is a time dependent property, the rate of

540
90 80

J. Chirife, A4. P. Buera

Wheat Starch

0: o
. g at Is E E E I2 m 0

.. ....
...

... -_.
\\

60:
5040. 30. 20. IO0-m -IO-20; -3o-40. -5o-60. -701 0.2 0.3

.. . ..y

carrots

0.4

0.5

0.6

0.7

0.8

0.9

Water activity

Fig. 4. Relationship between water activity and glass transition temperature of various plant materials and biopolymers [adapted from data reported by Karmas et af. (1992), Roos (1987), Hoseney et a/. (1986), Zeleznak & Hoseney (1987) and Slade & Finlay ( 1989)]. Isotherm data of starch, gluten and gelatin from lglesias and Chirife ( 1982).

approach to the equilibrium will be a function of c/t, where 5 is the relaxation time. However, the relaxation time is a function of T- Tg and since in the microbiological growth range, T- T is usually large for foods (Chirife & Buera,

1994), the approach

T-water

to the equilibrium

value will be relatively

fast. The

activity relationship for potatoes is much different from that of starch (about 65% of the dry matter). The pres(&g. 4), w hICh is its main component

ence of low molecular weight sugars in potatoes may account partially for this difference (Slade & Levine, 199 1a). Figure 5 shows the effect of low molecular weight solutes on the ?g values of some biopolymers. The glass transition temperature of amylopectm [from data reported by Kalichevsky and Blanshard (1993)] and maltodextrin DE 10 [from data reported by Roos and Himberg (1993)], at various moisture contents, are reduced by the addition of 9% fructose or 7.5% (total solid basis) of o-xylose + lysine, respectively. For example, the addition of 7.5% of lysine plus D-xylose reduced the Tg of maltodextrin by about 15-23C.
Prediction of water activity in semi-moist foods

In concern with the experimental determination of water activity in semi-moist foods, Franks (1991~) and Slade and Levine (1991a) have noted that the high viscosity of semi-moist systems is related to a much slower rate of attainment of equilibrium and what is measured in such cases is a non-equilibrium RVP and not A,. Franks ( 199 1a) noted, however, that in liquid systems (simple solutions), the assumption of equilibrium (and hence the use of water activity) can be valid, because the diffusion rate of water molecules is high compared to the time scale of the measurement of vapor pressure. Thus, it is interesting to compare the

Water activity values offoods

541

-101
0

IO

15

20

25

30

35

40

Moisture content, % d.b.

Fig. 5. Effect of the addition of low molecular weight compounds on the glass curves of amylopectin-water [from data reported by Kalichevsky and Blanshard (1993)] and maltodextrin DE lo-water [from data reported by Roos and Himberg (1993)l.

experimentally determined A, of some semi-moist foods (non-equilibrium values) with predictions made from knowledge of the water activity of simple solutions (equilibrium values) representative of their main soluble substances. In semi-moist foods containing a relatively high proportion of soluble substances (usually NaCl and/or low molecular weight sugars), A,, is mainly determined by the molality of solute(s) in the aqueous phase (Lupin et al., 198 1). Figures 6 and 7 compare predicted and measured water activity values in figs and raisins in the intermediate moisture range. The sugar content of figs and raisins was (dry basis) 75.2% and 853%, respectively (Tsami et al., 1990), and the main sugars were glucose and fructose. Predicted values were calculated by applying Roos equation (Chen, 1990) to the data reported by Chirife et al. ( 198 1) for the water activities of glucose and fructose solutions. In both cases (figs and raisins), a good agreement between predictions and measurements was observed. The goodness of the agreement should be judged on the basis of the usual uncertainties associated with the determination of sorption values in foods. The comparisons were extended to a wide variety of semi-moist foods comprising many physical states, such as solids (fish), pastes (ketchup, milk, jam), jellies (fruit preserves, marmalades), emulsions (mayonnaises) and very concentrated liquid food solutions (sugar molasses and honey), as shown in Fig. 8. Again, a relatively good agreement was observed between water activity determinations made for foods (some of them presumably non-equilibrium systems) and predictions made from the knowledge of literature values for the water activity of aqueous solutions (equilibrium systems) of their main soluble solids (usually NaCl, sucrose, glucose, fructose). It is to be noted that the slight discrepancies between predictions and measurements may also be attributed among other factors to some inaccuracies in the electronic hygrometers (calibrations,

542
llOFIGS, ioo30 C

J. Chirife, AI. P. Buera

20

0.6

0.6

Water activity

Fig. 6. Comparison of predicted and measured water sorption data in figs at 30C [measured values calculated from isotherm data reported by Tsami et af. (1990)].

Experimental - Ads.

Predicted (Glucose + Fructose)

0.3

0.4

0.5

0.6

0.7

0.0

0.9

Water activity

Fig. 7. Comparison of predicted and measured water sorption in raisins at 30C [measured values calculated from isotherm data reported by Tsami et al. ( 1990)].

standards for calibration, etc.) mostly utilized for the experimental determination of A, in the foods studied (van den Berg, 1991; Chirife, 1993). On the contrary, the literature data for simple aqueous solutions were obtained by the accurate isopiestic vapor pressure technique widely used in the physico-chemical area (Robinson & Stokes, 1965). The observed agreement between water activity measurements in foods (non-equilibrium systems) and predictions made

Water activity values offoods

543

0.6

0.64

0.68

0.72

0.76

0.8

0.84

81

0.88

I,

0.92

0.96

Measured water activity Fig. 8. Predicted and measured water activity values in various semi-moist and concentrated foods. Moist salted fish: this includes, trout (smoked), eel (smoked), tuna fish (in oil), cod, salmon (smoked), hake and sardines (minced) (Aguilera et al., 1990, 1992). Moisture content of the different salted fish products ranged between about 44-70% (w/w). Milk jams: household consumption and confectionery use (Ferramondo et al., 1984). Moisture content of the different jams ranged between about 23-28% (w/w). Ketchup and mayonnaises (Aguilera et al., 1990): moisture content of ketchup and mayonnaises samples ranged between ~59-69~~ (w/w) and 16-18S (w/w), respectively. Fruit preserves: includes jellies, marmalades or jams (Aguilera et al., 1990), of one of the following fruits: peach, figs, lemon, mango, orange, papaya, pineapple, custard apple, tamarind, guava, brambleberry, apricots, sloelike fruit; also sugar cane molasses and sweet cucumbers. Moisture content of the different products varied between about 20-60% (w/w). Honey (synthetic): the synthetic honeys were obtained by adding water to a mixture of sugars containing 48% (w/w) fructose, 40% (w/w) glucose, 10% (w/w) maltose, and 2% (w/w) sucrose (Ruegg & Blanc, 1981). Moisture content varied between 18.1 and 28% (w/w).

from equilibrium situations (liquid solutions), may suggest that non-equilibrium effects (due to a slow rate of moisture diffusion) in some semi-moist foods are relatively small, at least if the variations are compared with the usual error in the determinations of A,.
Use of water activity as a predictor of microbial stability in semi-moist foods

The definition of a moisture condition at which pathogenic (or spoilage) microorganisms cannot grow is of paramount importance in food preservation. It has been repeatedly shown in the literature that each microorganism has a critical water activity, below which growth cannot occur (Brown, 1974; Troller & Christian, 1978; Gould & Measures, 1977; Troller, 1980; Leistner et al., 1981; Beuchat, 1983). For example, pathogenic bacteria cannot grow below water activity O-86; yeasts and molds are more tolerant to reduced water activity, but usually no growth exists below a water activity of about 0.62 (Scott, 1953;

544

J. Chirife, M. P. Buera

Hocking & Pitt, 1979; Beuchat, 1983; Silverman et al., 1983; Leistner, 1987; Jermini & Schmidt-Lorenz, 1987; Chirife & Favetto, 1992; Ballesteros et al., 1993). Many workers concluded that the water activity is the determinant of growth because it determines the osmotic stress and because the ability to grow is determined by the degree of that stress and the osmoregulatory capacity of a particular microbial cell. A fundamental requirement for growth of microorganisms on substrates of high osmolality is the intracellular accumulation of solutes, either by transport or synthesis, to concentrations that counterbalance the osmolality of the external medium (Prior et al., 1987; Hocking, 1988); this process is often referred to as osmoregulation. The biochemical basis of osmoregulation has been the subject of numerous studies in the past decade (Anderson & Witter, 1982; Csonka, 1989; Landfald & Strom, 1986; Larsen et al., 1987; Gervais et al., 1992). Slade and Levine ( 1991 a) have questioned the usefulness of specifications of values of A, required for the microbial safety and stability of semi-moist foods on the basis of the so-called specific solute effect. Since the early reports by Scott (1953) it has been known that the microbial response may differ at a particular A, when the latter is obtained with different solutes. It is more or less established that the A, of the medium is not the only determining factor regulating the biological response but that the nature of the A, controlling solute also plays a role (Christian, 198 1; Gould, 1988; Ballesteros et al., 1993). Ballesteros et al. (1993) recently examined the growth behaviour of Staphylococcus aureas in solutions of water activity controlled using 16 different solutes comprising inorganic and organic salts, sugars, polyols, etc. and including solutes which are unlikely to ever be used in foods. They reported that although the minimal A, for growth depended in various cases on the solute used to adjust it, S. uureus could not grow below the current widely accepted minimal A, = O-86. This is a result of the outmost importance regarding food safety (Silverman et al., 1983). Also, when the solutes most often present in reduced A, preserved foods (e.g. NaCl, sucrose, glucose, KCI) are used to control A,, specific solute effects are much less evident. Chirife (1993) reviewed the minimal water activity for growth of several pathogenic bacteria and reported that the minimal A,,, allowing growth was about the same when the solutes used to control it were NaCl, sucrose, glucose or KCI. When glycerol was used, a specific effect on the minimal A, for growth was much more evident. Minimal water activity values for growth of different microorganisms have been mostly determined in liquid laboratory media, using sodium chloride to depress the water activity (Baird-Parker & Freame, 1967; Briozzo et al., 1986; Emodi & Lechowich, 1969). In these conditions there is no doubt that measured RVP are equilibrium ones and thus, A,,, =p/p,, since the viscosity of NaCl solutions having A, in the range O-85-0.95 would be only slightly above that for pure water (e.g. > 1 < 2 cp) (Ballesteros et al., 1993). These values are far away from the value of 10 Pa/s corresponding to the onset of rubbery, non-equilibrium behaviour (Soesanto & Williams, 198 1). However, if we accept the idea (Slade & Levine, 1991 a) that in a solid semi-moist food, equilibrium is not established during the time of observation, the recorded vapor pressure will be mistaken for equilibrium, and this will create some concerns about the safety of measured A, values in foods as regarding microbial growth inhibition. This situation may be investigated using literature data. For

Wateractivity values offoods

545

example, it has been well established that the minimum A, for the growth of Clostridium botulinum types A and B in liquid broth media adjusted with NaCl is 0*94/0.95 (Baird-Parker & Freame, 1967; Ohye & Christian, 1966). Glass and Doyle (1991) challenged these values in a study of the relationship between the water activity of fresh pasta and toxin production by Clostridium botulinum. Four types of fresh pasta (meat- or cheese-filled tortellini and flat noodle linguine or fettucini) were prepared with different water activities, inoculated with C. botulinum, packaged under a modified atmosphere, and stored at 30C for S-10 weeks. The pH of all samples was favourable to C. botufinum growth. No toxin was detected in tortellini with an A, of 0.94 or below held at 30C for 10 weeks. Toxin was produced at 2 weeks in linguine at A, = 0.96 held at 3OC, whereas no linguine or fettucini at A, = 0.93 or 0.95 and held at 30C was toxic during 8 or 10 weeks of incubation. Glass and Doyle ( 1991) concluded that the A, of fresh pasta is a principal factor in preventing botulinial toxin production in temperature-abused products. These results are in very good agreement with predictions made from the behaviour of C. botufinum in liquid broth media of known water activity, and suggest that, at least within the time frame of the experiments, non-equilibrium effects in the determination of the water activity of pasta are not significant. Dodds (1989) reported a study on the effects of water activity on toxin production by C. botulinum in cooked, vacuumpackaged potatoes of A, controlled by the addition of NaCl, and incubated for up to 60 days at 25C. Toxin was produced at an A,, of 0.96, but no toxin was detected when the A, was 0.955, in good agreement with predictions made from the behaviour of the bacterium in liquid broth media of measured equilibrium RVP (or A,,,). These results again suggest that, in the time frame of the experiments (which in many cases may be coincident with the expected shelf-life of the product), non-equilibrium effects (e.g. slow rate of moisture diffusion) do not significantly influence the determination of A, in potatoes. Consequently, the bacterial response may be predicted adequately. Valik and Gorner (1993) studied the growth of S. aureus in pasta dough in relation to its water activity and found that the bacterium appeared to multiply until the A,. was below 0.86 at which stage it decreased. This is in good agreement with the minimal water activity (A,, = 0.86) for growth of S. aureus determined in liquid broth media adjusted with NaCl (Vaamonde et al., 1982). Giannuzzi and Parada (1986) studied the behaviour of S. aureus (three strains) in dehydrated milk, beef and pork equilibrated at water activities of 0.84 and 0.90 and incubated at 30C for up to 30 days. No growth was observed in any of the foods studied at A, = O-84; however, evident growth was noticed in all systems at A, = 0.90, and these results agreed with the known behaviour of S. aureus in liquid broth media of water activity adjusted with NaCl (Vaamonde et al., 1982). The glass transition temperature of dried milk as a function of moisture content (and A,,,) has been recently reported by Jouppila and Roos ( 1994). This information is useful for a better analysis of the results reported by Giannuzzi and Parada ( 1986) on S. aureuS growth in dried milk. Figure 9 (from data reported by these authors) shows the behaviour of three S. aureus strains inoculated in dried milk (A, = O-84) and incubated at 30C. No growth was observed for any of the three strains in spite that at this water activity, dried milk should be in the rubbery state (estimated value of T- Tg above 82C) and lactose probably had recrystallized due to the high mobility of the system (Jouppila & Roos, 1994). In

546
l.OE+OO

J. Chirife, M. P. Buera
Sfnphyfococcur

nureun, 30 C

FDA-C243

II:.7, ,y:; 1
0 5 10 15 20 25 30 35 40

Incubation time, day Fig. 9.

Behaviour of Staphylococcus aureus (strains FMl, ATCC 6538 P and FDAC243) in dried milk of A, = 0.84 incubated at 30C [from data reported by Gianuzzi and Parada (1986)]; (T- T,) value estimated from data reported by Jouppila and Roos (1994).

any case, mobility effects do not seem to play a role in the observed growth inhibition which is likely to be controlled by water activity. It is noteworthy that in the S. aureus growth experiments performed by Giannuzzi and Parada ( 1986) water activity should have remained constant in spite of some water loss due to lactose crystallization because dried milk was kept in a closed desiccator over a saturated salt solution (KCl) of water activity 084. Silverman et al. (1983) reported that A, was the main parameter in controlling S. aureus growth in precooked bacon. They reported that the limiting A, for the growth of S. aureus A-100 sealed in cans and stored for up to 28 days at 37C was 0.87 in good agreement with known behaviour in liquid broth medium, as noted above.

CONCLUSIONS The analysis of several literature data suggests that although some amorphous intermediate moisture foods are non-equilibrium systems (e.g. very slow rate of attainment of equilibrium), this may not seriously affect either the stability (time dependency) of measured water sorption isotherms, or the prediction of water activity from equilibrium data corresponding to aqueous solutions of main food solutes (NaCl, sugars). It appears that these non-equilibrium effects, if present, produce changes which would be within the range of the experimental error in sorption isotherm determination and/or A,,, measurement in foods. These conclusions agree with earlier statements of van den Berg (1981). He suggested that although in some foods there exist no true thermodynamic equilibrium, some pseudo equilibrium exists (which is not clearly definable) leading to the use of an empirical pseudo water activity; he added that the

Water activity values offoods

547

numerical value of this practical pseudo water activity will not, in many cases, deviate much from the strict thermodynamic equilibrium water activity. The cases where the numerical value of A, may deviate from the equilibrium value are those where a much delayed crystallization phenomenon occurs (also a nonequilibrium situation). Crystallization of amorphous sugars in foods may lead to significant modifications of the A, values. At water activities in the microbiological growth range, real foods are in most cases in the rubbery liquid state, in which the rates of diffusion limited phenomena (as is the case with moisture equilibrium) are much more rapid than in the glassy solid state. As far as the prediction of microbial growth in foods is concerned, the non-equilibrium effects discussed here (inability of water to diffise in the semi-moist food) may be in many situations slow within the time frame (foods shelf-life) of the experiments, and/or so small that they do not seriously affect the application of the water activity concept as a predictor of microbial growth. It is noteworthy that we are not saying here that since changes in non-equilibrium food systems might be slow and/or small, and since glass transitions might not occur during shelf storage and if such a system were to remain in a rubbery liquid state, then considerations of non-equilibrium effects, glassy and rubbery states, etc. are unnecessary and that the water activity concept is all that one needs to ensure the safety of semi-moist foods. This is not so. Water activity has several limitations and should be always used carefully and this must include considerations (and precautions) regarding the possible influence of some non-equilibrium situations, as reported by Slade and Levine in their pioneer work ( 1987). It is noteworthy, however, that these warnings are not new. Almost twenty years ago they were clearly stated by Reid (1976) in an international symposium entirely devoted to Intermediate Moisture Foods, held in Reading, UK. He pointed out that since A, is a thermodynamic concept, it refers only to equilibrium; thus if equilibrium has not been attained, the definition does not apply. One might measure a relative humidity but, if it is not equilibrium relative humidity, it is not a measure of water activity, though it may still be an important quantity. Reid (1976) added the following examples. If a system is not at internal equilibrium, one might measure a steady vapor pressure (over a period of measurement) which is not the true vapor pressure of the system. Another situation which could readily occur in a food system is a pseudo equilibrium, e.g. a solution capable of supersaturation. The vapor pressure of the supersaturated solution is lower than that of the saturated solution, thus the A, is also lower. Whether this observation is meaningful or not depends on the stability of the supersaturated solution. Reids concepts may be summarized as follows: since activity is a thermodynamic concept anyone who is going to employ the term must be aware of the requirements of the definitions. Since Reids (1976) paper was specifically addressed to people working on A,-related food preservation situations, one might hope (or at least desire) that they were taken into account. ACKNOWLEDGEMENT The authors acknowledge financial support from CONICET de Buenos Aires, Republica Argentina. and Universidad

548

J. Chirife, M. P. Buera

REFERENCES Aguilera, J. M., Francke, A., Figueroa, G., Bornhardt, C. & Cifuentes, A. (1992). Preservation of minced pelagic fish by combined methods. Int. J. Food Sci. Technol., 27,
171-7.

Aguilera, J. M., Chirife, J., Tapia de Daza, M. S. & Welti Chanes, J. ( 1990). Inventario de Alimentos de Humedad Intermedia Tradicionales de Iberoamerica. ed. E. Parada Arias, Instituto Politecnico National, Mexico. Anderson, C. B. & Witter, L. D. (1982). Glutamine and proline accumulation by Staphylococcus aureus with reduction in water activity. Appl. Env. Microbial., 43, 1501-3. Andrieu, J., Stamatopoulos, A. & Zafiropoulos, M. (1985). Equation for fitting desorption isotherms of durum wheat pasta. J. Food Technol., 20,65 l-7. Ayranci, E., Ayranci, G. & Dogantan, Z. (1990). Moisture sorption isotherms of dried apricot, fig and raisin at 20 and 36C. J. FoodSci., 55, 1591-3, 1625. Baird-Parker, A. C. & Freame, B. (1967). Combined effects of water activity, pH and temperature on the growth of Clostridium botulinum from spore and vegetative cell inocula. J. Appl. Bacterial., 30,420-9. Ballesteros, S. A., Chirife, J. & Bozzini, J. P. (1993). Specific solute effects on Staphylococcus aureus cells subjected to reduced water activity. Int. J. Food Microbial., 20,
5 l-66.

Beuchat, L. R. (1983). Influence of water activity on growth, metabolic activities and survival of yeasts and molds. J. Food Protect., 46, 135-4 1. Biquet, B. & Labuza, T. P. (1988). Evaluation of the moisture permeability characteristics of chocolate films as an edible moisture barrier. J. Food Sci., 53,989-98. Bizot, H., Buleon, A., Mouhous-Riou, N. & Multon, J. L. (1985). Some facts concerning water vapour sorption hysterisis on potato starch. In Properties of Water in Foods in Relation to Quality and Stability, eds D. Simatos & J. L. Multon, pp. 83-93. Martinus Nijhoff, Dordrecht, The Netherlands. Bolin, H. R. (1980). Relation of moisture to water activity in prunes and raisins. J. Food Sci., 45, 1190-l. Briozzo, J., de Lagarde, E. A., Chirife, J. & Parada, J. L. (1986). Effect of water activity and pH on growth and toxin production by Clostridium botulinum type G. Appl. Environ. Microbial., 5 1,844-g. Brown, A. D. (1974). Microbial water relations: features of the intracellular composition of sugar-tolerant yeasts. J. Bacterial., 118, 769-77. Buera, M. P., Chirife, J., Resnik, S. L. & Wetzler, G. (1987). Nonenzymatic browning in liquid model systems of high water activity. 2. Kinetics of color changes due to Maillard reactions between different single sugars and glycine and comparison with caramelization browning. J. Food Sci., 52, 1063-70. Catiellas, J., Rosello, C., Simal, S., Soler, L. & Mulet, A. (1993). Storage conditions affect quality of raisins. J. Food Sci., 58, 805-9. Chen, C. S. (1990). Predicting water activity in solutions of mixed solutes. J. Food Sci.,
55,494-7,515.

Chhinnan, M. S. & Beuchat, L. R. (1985). Sorption isotherms of whole cowpeas and flours. Lebensm. Wiss. Technol., 18,83. Chirife, J. (1993). Physicochemical aspects of food preservation by combined factors.
Food Control, 4,210-15.

Chirife, J., Ferro Font&r, C. & Favetto, G. J. (1981). The water activity of fructose solutions in the intermediate moisture range. Lebensm. Wiss. Technol., 15, 159-60. Chirife, J. & Favetto, G. J. (1992). Some physicochemical basis of food preservation by combined methods. Food Res. Int., 25,389-96. Chirife, J. & Buera, M. P. (1994). Water activity, glass transition and microbial stability in concentrated/semi-moist food systems. J. Food Sci., 59 (5).

Water activity values offoods

549

Christian, J. H. B. (1981). Specific solute effects on microbial water relations. In Water Activity: Influences on Food Quality, eds L. B. Rockland & G. F. Stewart, pp. 825-54. Academic Press, New York. Csonka, L. N. (1989). Physiological and genetic responses of bacteria to osmotic stress.
Microbial. Rev., 53, 121-47.

Dodds, K. L. (1989). Combined effect of water activity and pH on inhibition of toxin production by Clostridium botulinurn in cooked, vacuum-packaged potatoes. Appl.
Environ. Microbial., S&656-60.

Emodi, A. S. & Lechowich, R. V. (1969). Low temperature growth of type E Clostridium botulinurn spores. 2. Effect of solutes and temperature. J. Food Sci., 34,82-7. Engels, C., Hendrickx, M. & Tobback, P. (1987). Limited multilayer desorption of brown, parboiled rice. ht. J. Food Sci. Technol., 22,219-23. Ferramondo, A., Chirife, J., Parada, J. L. & Vigo, S. (1984). Chemical and microbiological studies on dulce de leche a typical Argentine confectionery product. J.
Food Sci., 49,821-3.

Ferro Font&r, C., Benmergui, E. A. & Chirife, J. (1980). The prediction of water activity in aqueous solutions in connection with intermediate moisture foods. 3. Aw prediction in multicomponent strong electrolyte solutions. J. Food Technol., 15,47-58. Franks, F. ( 199 1 a). Water activity: a credible measure of food safety and quality? Trends
Food Sci. Technol., 68-72.

Franks, F. (199 1 b). Hydration phenomena: an update and implications for the food processing industry. In Water Relationships in Foods, eds H. Levine & L. Slade. Plenum Press, New York, pp. 1-19. Gal, S. (1981). Recent developments in techniques for obtaining complete sorption isotherms. In Water Activity: Znfruences on Food Quality, eds L. B. Rockland & G. F. Stewart. Academic Press, New York, pp. 89-102. Gervais, P., Perrier-Cornet, J. M. & Berner, J. L. (1992). Osmoregulation mechanisms of the yeast Sporidiobolus salmonicolor. J. Agric. Food Chem., 40,17 17-2 1. Giannuzzi, L. & Parada, J. L. (1986). Sobre el crecimiento de Staphylococcus aureus en medios solidos de actividad acuosa inferior a 0.86. Rev. Arg. Microbial. Glass, K. & Doyle, M. P. (1991). Relationship between water activity of fresh pasta and toxin production by proteolytic Clostridium botulinum. J. Food Protect., 54, 162-S. Gould, G. W. (1985). Osmoregulation: is the cell just a simple osmometer? The microbiological experience. In A Discussion Conference: Water Activity: A Credible Measure of Technological Performance and Physiological Viability? Faraday Div., Royal Sot. Chem., Cambridge, UK, l-3 July. Gould, G. W. (1988). Interference with homeostasis - food. In Homeostatic Mechanisms in Microorganisms, eds J. G. Banks, R. T. G. Board, G. W. Gould & R. W. Mittenbury. Bath University Press, UK. Gould, G. W. & Measures, J. C. (1977). Water relations in single cells. Phil. Trans. R. Sot.
London, 278,151.

Gould, G. W. & Christian, J. H. B. (1988). Characterization of the state of water in foods - Biological aspects. In Food Preservation by Moisture Control, eds C. C. Seow, T. T. Teng & C. H. Quah Elsevier Applied Science, Oxford, pp. 43-56. Hansen, J. R. (1976). Hydration of soybean protein. J. Agric. Food Chem., 24,1136-4 1. Hayakawa, K. I., Matas, J. & Hwang, M. P. ( 1978). Moisture sorption isotherms of coffee products. J. FoodSci., 43, 1026-7. Hendel, C. E., Silveira, V. G. & Harrington, W. 0. (1955). Rates of nonenzymatic browning of white potato during dehydration. Food Technol., 9,433-g. Hill, P. E. & Rizvi, S. S. H. (1982). Thermodyanamic parameters and storage stability of drum dried peanut flakes. Lebensm. Wiss. Technol., 15,185-90. Hocking, A. D. & Pitt, J. I. (1979). Water relations of some Penicillium species at 25C.
Trans. Brit. Mycol. Sot., 73, 141-5.

550

J. Chirife, M. P. Buera

Hoseney, R. C., Zeleznak, K. & Lai, C. S. (1986). Wheat gluten: a glassy polymer. Cereal Chem., 63,285-6. Hutchinson, D. H. & Otten, L. (1984). Equilibrium moisture content of white beans.
Cereal Chem., 61, 155-8.

Iglesias, H. A. & Chirife, J. (1978). Delayed crystallization of amorphous sucrose in humidified freeze dried model systems. J. Food Technof., 13, 137-44. Iglesias, H. A. & Chirife, J. (1982). Handbook of Food Isotherms, Academic Press, New York. Jayas, D. S. & Mazza, G. ( 199 1). Equilibrium moisture characteristics of safflower seeds.
Trans. ASAE, 342099-103.

Jermini, M. F. G. & Schmidt-Lorenz, W. ( 1987). Growth of osmotolerant ent water activity values. J. Food Protect., 50,404-10. Jouppila, K. & Roos, Y. H. (1994). Glass transitions and crystallization Submitted to J. Dairy Sci. Karmas, R., Buera, M. P. & Karel, M. (1992). Effect of glass transitions enzymatic browning in food systems. J. Agric. Food Chem., 40,873-9. Kalichevsky, M. T. & Blanshard, J. M. V. (1993). The effect of fructose glass transition of amylopectin. Carbohy. Polym., 20,107-13. King, C. J. (1968). Rates of moisture sorption and desorption in porous,
Food Technol., 22,509-15.

yeasts at differin milk powders. on rates of nonand water on the dried foodstuffs.

Kumar, M. (1974). Water vapour adsorption on whole corn flour, degermed corn flour and germ flour. J. Food Technol., 9,433-44. Lagoudaki, M., Demertzis, P. G. & Kontominas, M. G. (1993). Moisture adsorption behaviour of pasta products. Lebensm. Wiss. Technol., 26,512-16. Landfald, B. & Strom, A. R. (1986). Choline-glycine betaine pathway confers a high level of osmotic tolerance in Escherichia coli. J. Bacterial., 165,849-55. Larsen, P. I., Sydnes, L. K., Landfald, B. & Strom, A. R. (1987). Osmoregulation in Escherichia coli by accumulation of organic osmolytes: betaines, glutamic acid and trehalose. Arch. Microbial., 147, 1-7. Leistner, L. & Rodel, W. (1975). The significance of water activity for micro-organisms in meats. In Water Relations of Foods, ed. R. D. Duckworth. Academic Press, New York, pp. 309-23. Leistner, L., Rodel, W. & Krispien, K. (1981). Microbiology of meat and meat products in high and intermediate moisture ranges. In Water Activity: Infruences on Food Quality, eds L. B. Rockland & G. F. Stewart. Academic Press, New York, pp. 855-916. Leistner, L. (1987). Shelf stable products and intermediate moisture foods based on meat. In Water Activity: Theory and Applications to Food, ed. L. B. Rockland & L. R. Beuchat. Marcel Dekker, New York, pp. 295-327. Lewicki, P. ( 1975). Powtarzhtosc statycznoexykatorialnej metody yznacznia izoterm sorpocji wody produktow spozywczych. Przemys Spozywczy, XXX, 3-8. Lima, A. W. 0. & Cal-Vidal, J. (1983). Hygroscopic behaviour of freeze dried bananas. J.
Food Technol., 18,687-96.

Lomauro, C. J., Bakshi, A. S. & Labuza, T. P. (1985). Moisture transfer properties of dry and semimoist foods. J. FoodSci., 50,397-400. Lupin, H. N., Boeri, R. L. & Moschiar, S. M. (1981). Water activity and salt content relationship in moist salted fish products. J. Food Technol., 16, 31-8. Mac&a, D. J. & Bettelheim, F. A. (1964). Sorption of water vapor by amylose B. Polymer Lett., 2, 110 l-4. Mazza, G. & Jayas, D. S. (1991). Evaluation of four three-parameter equations for the description of the moisture sorption data of Lathyrus pea seeds. Lebensm. Wiss.
Technol., 24562-5.

Ohye, D. F. & Christian, J. H. B. (1966). Combined effects of temperature,

pH and water

Water activity values offoods

551

activity on growth and toxin production by Clostridium botulinum types A, B and E. In Proc. 5th Int. Symp. Food Microbiology, Moscow, p. 217. Pitt, J. I. (1975). Xerophilic fungi and the spoilage of foods of plant origin. In Water Relations of Foods, ed. R. B. Duckworth, Academic Press, London, pp. 273-307. Piiiaga, F. & Lafuente, B. (1965). Horchata en polvo. I. Humedades de equilibrio de la horchata liofilizada. Rev. Agroq. Tecnol. Alimentos, 5,99-105. Prior, B. A., Kenyon, C. P., van den Veen, M. & Mildenhall, J. P. (1987). Water relations of solute accumulation in Pseudomonasfluorescens. J. Appl. Bacterial., 62, 119-28. Rapusas, R. S., Driscoll, R. H. & Buckle, K. A. (1993). Moisture desorption characteristics of raw onion slices. FoodAustralia, 45,278-84. Rasekh, J. G., Stillings, B. R. & Dubrow, D. L. (197 1). Moisture adsorption of fish protein concentrate at various relative humidities and temperatures. J. Food Sci., 36,
705-7.

Reid, D. S. (1976). Water activity concepts in intermediate moisture foods. In Intermediate Moisture Foods, eds R. Davies, G. G. Birch & K. J. Parker. Applied Science. London, pp. 54-65. Resnik, S. L. & Chirife, J. (1979). Effect of moisture content and temperature on some aspects of nonenzymatic browning in dehydrated apple. J. Food Sci., 44,601-5. Roos, Y. (1987). Effect of moisture on the thermal behavior of strawberries studied using differential scanning calorimetry. J. Food Sci., 52, 146-9. Roos, Y. (1993). Water activity and physical state effects on amorphous food stability. J.
Food Process. Present., 16,433-47.

Roos, Y. & Karel, M. (1991 a). Plasticizing effect of water on thermal behavior and crystallization of amorphous food models. J. Food Sci., 56, 38-43. Roos, Y. & Karel, M. (1991 b). Phase transition of mixtures of amorphous polysaccharides and sugars. Biotechnol. Prog., 7,49-53. Roos, Y. & Himberg, M. J. (1993). Low temperature nonenzymatic browning at the glass transition range of an intermediate moisture food model. Submitted. Ruegg, M. & Blanc, B. (1981). The water activity of honey and related sugar solutions. Lebensm. Wiss. Technol., 14, l-6. Sa, M. M. & Sereno, A. M. (1993). Effect of temperature on sorption isotherms and heats of sorption of quince jam. Znt. J. Food Sci. Technol., 28,24 l-8. Saravacos, G. D., Tsiourvas, D. A. & Tsami, E. (1986). Effect of temperature on the water adsorption isotherms of sultana raisins. J. Food Sci., 5 I,38 l-3,387. Scott, W. J. (1953). Water relations of Staphylococcus aureus at 30C. Austr. J. Biol. Sci..
6,549-56.

Silverman, G. J., Munsey, D. T., Lee, C. & Ebert, E. (1983). Interrelationship between water activity, temperature and 5.5% oxygen on growth and enterotoxin A secretion by Staphylococcus aureus in precooked bacon. J. Food Sci., 48, 1783-6,1795. Slade, L. & Levine, H. (1987). Structural stability of intermediate moisture foods - A new understanding. In Food Structure - Its Creation and Evaluation, eds J. R. Mitchell & J. M. V. Blanshard. Butterworths, London, pp. 115-47. Slade, L. & Finley, J. W. (1989). Protein-water interactions: water as a plasticizer of gluten and other protein polymers. In Protein Quality and the Effects of Processing, eds R. Dixon Phillips & J. W. Finley. Marcel Dekker, New York, pp. 9- 124. Slade, L. & Levine, H. (1991a). Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit. Rev. Food Sci. Slade, L. & Levine, H. ( 1991 b). A food polymer science approach to structure-property relationships in aqueous food systems. In Water Relationships in Foods, eds H. Levine & L. Slade. Plenum Press, New York, pp. 29- 10 1. Soesanto, T. & Williams, M. C. (1981). Volumetric interpretation of viscosity for concentrated and dilute sugar solutions. J. Phys. Chem., 85,3338-41.
Nutrit., 30, 115-360.

552

J. Chirife, M. P. Buera

Spiess, W. E. L. & Wolf, W. R. (1985). The results of the COST 90 project on water activity. In Physical Properties of Foods, eds F. Escher, B. Hallstrom, H. S. Meffert, W. E. L. Spiess & G. Voss. Applied Science, New York, pp. 65-87. Troller, J. A. (1980). Influence of water activity on microorganisms in foods. Food Troller, J. A. & Christian, J. H. B. (1978). WaterActivity and Food. Academic Press, New York. Tsami, W., Marinos-Kouris, D. & Maroulis, Z. B. (1990). Water sorption isotherms of raisins, currants, figs, prunes and apricots. J. Food Sci., 55, 1594-7, 1625. Vaamonde, G., Chirife, J. & Scorza, 0. C. (1982). An examination of the minimal water activity for Staphylococcus aureus ATCC 6538 P growth in laboratory media adjusted with less conventional solutes. J. FoodSci., 47, 1259-62. Valdez-Niebla, J. A., Paredes-Lopez, O., Vargas-Lopez, J. M. & Hernindez-Lopez, D. (1993). Moisture sorption isotherms and other physicochemical properties of nixtamalized amaranth flour. Food Chem., 46,19-23. Valik, L. & Garner, F. (1993). Growth of Staphylococcus aureus in pasta in relation to its water activity. Int. J. Food Microbial., 20,45-8. van den Berg, C. ( 198 1). Water activity and its estimation in food systems: theoretical aspects. In Water Activity: Injluences on Food Quality, eds L. B. Rockland & G. F. Stewart. Academic Press, New York, pp. l-6 1, van den Berg, C. (1991). Food-water relations: progress and integration, comments and thoughts. In Water Relationships in Foods, eds H. Levine & L. Slade. Plenum Press, New York. Wolf, W., Spiess, W. E. L., Jung, G., Weisser, H., Bizot, H. & Duckworth, R. B. (1984). The water-vapour sorption isotherms of microcrystalline cellulose (MCC) and of purified potato starch. Results of a collaborative study. J. Food Engng, 3,51-73. Wedzicha, B. L. & Quine, E. C. (1983). Sorption of water vapour by wafer biscuits. The effects of components, ingredients and the baking process. Lebensm. Wiss. Technol.,
16,115-18. Technol., 34,76-80,82.

Zeleznak, K. J. & Hoseney, R. C. (1987). The glass transition 64,121-4.

in starch. Cereal Chem.,