Mass Transfer Properties of Osmotic Solutions. I. Water Activity And

Home , Osmotic concentration

This article was downloaded by: [b-on: Biblioteca do conhecimento online UP] On: 24 October 2011, At: 16:02 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

International Journal of Food Properties Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ljfp20 Mass transfer properties of osmotic solutions. I. Water activity and osmotic pressure Vassilis Gekas a , Chelo Gonzalez b , Alberto Sereno c , Amparo Chiralt b & Pedro Fito b a Food Engineering, Lund University, Lund, Sweden b Universidad Politecnica de Valencia, Valencia, Spain c Escola Superior de Biotecnologia, Oporto, Portugal Available online: 02 Sep 2009

To cite this article: Vassilis Gekas, Chelo Gonzalez, Alberto Sereno, Amparo Chiralt & Pedro Fito (1998): Mass transfer properties of osmotic solutions. I. Water activity and osmotic pressure, International Journal of Food Properties, 1:2, 95-112 To link to this article: http://dx.doi.org/10.1080/10942919809524570

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and- conditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. INTERNATIONAL JOURNAL OF FOOD PROPERTIES, 1(2), 95-112 (1998)

MASS TRANSFER PROPERTIES OF OSMOTIC SOLUTIONS. I. WATER ACTIVITY AND OSMOTIC PRESSURE

Vassilis Gekas1'*, Chelo Gonzalez2, Alberto Sereno3, Amparo Chiralt2, and Pedro Fito2

1Food Engineering, Lund University, Lund, Sweden

2Universidad Politecnica de Valencia, Valencia, Spain

3Escola Superior de Biotecnologia, Oporto, Portugal. *Corresponding author

ABSTRACT

In this review paper data on water activity, solute activity and osmotic pressure of" binary and multi-component osmotic solutions are provided. The Characteristics of the osmotic solutions are needed for the optimization of mass transfer during osmotic process, and for the improvement of final product quality. The vant Hoff equation and Gibbs Duhem theorem are commonly used to estimate osmotic pressure and solute activity. Water activities can be easily estimated through experimental determination of the freezing point depression. The possibilities of the group contribution models such as the Analytical Solution of Groups (ASOG) approach are also explored. The future needs especially in the case of multicomponent solutions consisting of electrolyte and non-electrolyte mixtures are pointed out.

INTRODUCTION

A number of food processing unit operations imply immersion of the food in a high osmotic pressure medium containing sugars, such as sucrose, glucose, fructose, syrups Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 and salts, such as sodium chloride or their mixtures. Foods that are treated this way are fruits and vegetables or also meat and fish (Fito et al., 1994; Lazarides, 1994; Lenart and Flink 1984a, 1984b; Lenart, 1994; Lerici et al., 1985). The aims of the osmotic process are: partial dehydration before the final treatment such as drying or freezing, impregnation of solute to improve quality (i.e., cryoprotectant), osmo-freezing or thawing directly in an osmotic medium, and direct formulation of food products.

Research on the above topics so far has shown that the performance of the osmotic unit operations depends on the properties of the osmotic solutions used. From the mass transfer point of view the most important osmotic solution parameter is its water activity lowering capacity in terms of water activity or osmotic pressure- this is an important property for the purpose of dehydration. Due to the simultaneous mass transfer, i.e. water transport from the food to the osmotic medium and solute transport from the osmotic medium to the food, additional information on the solute size and solute activities are also important. A literature review has shown that so far the properties of the osmotic solution considered in order to interprete the unit operations results were solute concentration and only in a few cases there has been reported solution water activity data and to the authors' knowledge there is absent of osmotic pressure data or solute activity data. It is also known that concentrated solutions used in osmosis are real solutions which might deviate strongly from the ideal situations, thus activities in addition to concentrations should provide a more sound theoretical basis for the characterization of the osmotic solutions and better interpretation of the osmotic process. Commonly used osmotic solutions, based on FSTA database 1969-1996 are presented in Table 1. As it is shown, a common osmotic medium used especially for fruits is the sucrose solution or syrup of a concentration range of 40-70 Brix and most frequently used one is 60 Brix. Other sugars such as glucose, fructose, lactose have also been used. Various Dextrose Equivalent (DE) corn syrups have been used for fruits and vegetables whereas for potato, fish and meat, salt solutions (NaCl 15% are being the most common among them) were the preferred media. In a few cases a combination between a sugar(s) and a salt was used. The objective of this paper is to review data of water activity and their prediction models for binary and multi-component osmotic solutions commonly used for osmotic dehydration of fruits.

A. PUBLISED EXPERIMENTAL DATA ON WATER ACTIVITIES.

In Table 2 there are shown values of freezing point depression for various osmotic solutions obtained at our laboratory of Lund University. For comparison litterature values are also presented. In Table 3 the values of water activity of the same solutions as in Table 2 are presented along with the litterature values for comparison. Table 4 shows water activity values of glycerol solutions along with refractive index values of this solutions (Rizvi, 1995). Table 5 contains water activity values of NaCl from Chirife and Resnik (1984). Tables 6 and 7 provide literature data of osmosities and Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 water activity of sugar and electrolyte solutions.

B. METHODS OF MEASUREMENT

Freezing Point Depression

Different methods of water activity measurement are reviewed by Labuza (1984), Rizvi (1995), and Rahman (1995). It is common and simple to measure water activity (or osmotic pressure) of two-component and three-component osmotic solution using OSMOTIC SOLUTIONS. I 97

Table 1. Osmotic solutions commonly used in osmosis

Solution Type Concentration & temperature Types of foods BINARY Sucrose 40-70B, 30-70°C Apple, pineapple, carrot, kiwi, grapes, mushroom, papaya, coconut.

Glucose 40-60 B, 25-40°C Strawberries, plum, pineapple, apple, pear, cherry, apricot, carrot

Glyserole 10/25%, 5°C Strawberries

NaCl 8-25% , 8-40°C Potato, okra, pepper, carrot, aubergine, green beans, meat, fish MULTICOMPONENT

Sucrose + NaCl 45 % - 15% Potato, apple, pineapple or 50%-10% 20-40°C Sucrose +Xylitol 30% + 70% Vegetables Corn syrup solids 34-70%, Papaya, apple, some vegetables DE10-40 35-55°C Corn Syrup / Sucrose/Water 5/3/1, 70B Cherries

Table 2. Freezing point depression of sugars (Gonzalez et al., 1995)

Solution type Measured values Litterature values in the authors's laboratory Mean Fructose 30% -4.84 -4.75 -4.79 -4.70 Sucrose 50% -7.24 Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 -7.64 -7.54 -7.61 -7.64 Sucrose 52% -8.97 -8.87 -8.92 -8.40 Sucrose 60% -12.30 -12.90 -12.70 -12.45 -12.90 Sucrose 60% + NaCl 10% -28.54 -28.14 -28.34 98 GEKAS ET AL.

Table 3. Water activity measured in the authors' laboratory and the litterature values

Solution From measured values From litterature values Fructose 30% 0.955 0.954 (a) 0.961 (b) Sucrose 50% 0.930 0.929 (a) Sucrose 52% 0.918 0.922 (a) Sucrose 60% 0.882 0.874 (a) Sucrose 60% + NaCl 10% 0.757 0.751 (c) (a) From Ferro-Fontan-Chirife Equation (b) Measured by electric hygrometer (c) From Caurie model

Table 4. Water activity of glycerol solutions (Rizvi, 1995)

Concentration Refractive Index Water Activity (kg/L) 1.3463 0.98 1.3560 0.96 0.2315 1.3602 0.95 0.3789 1.3773 0.90 0.4973 1.3905 0.85 0.5923 1.4015 0.80 0.6751 1.4109 0.75 0.7474 1.4191 0.70 0.8139 1.4264 0.65 Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 0.9285 1.4387 0.55 0.9760 1.4440 0.50 1.4529 0.40 OSMOTIC SOLUTIONS. I 99

Table 5. Water activity of NaCl solutions1"2

Concentration Water Activity Concentration Water Activity (%, w/w) (%, w/w) 0.5 0.997 10 0.935 1.0 0.994 11 0.927 1.5 0.991 12 0.919 2.0 0.989 13 0.911 2.5 0.986 14 0.902 3.0 0.983 15 0.892 3.5 0.980 16 0.883 4.0 0.977 17 0.873 4.5 0.973 18 0.862 5.0 0.970 19 0.851 5.5 0.967 20 0.839 6.0 0.964 21 0.827 6.5 0.960 22 0.815 7.0 0.957 23 0.802 7.5 0.954 24 0.788 8.0 0.950 25 0.774 9.0 0.943 26 0.759" 1 In the temperature range 15-5O°C 2 Data source Chirife and Resnik (1984) a Saturation point

freezing point depression method. The solutions were immersed in an ethanol thermostatized bath, kept at a temperature of approximately -46 °C (Lerici et al., 1983). The solutions were vigorously agitated in order to avoid external resistances in heat transfer. Agitation was found very important and also the ethanol temperature to be kept at least 30 degrees below the freezing point (FP) of the solution. To" obtain water activities from freezing point depression values three alternative equations were used i.e. one for ideal solutions, secondly one for real ones, and lastly a numerical

Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 approximation of the equation for real solutions suggested by Ferro-Fontan and Chirife. The differences between the second and third case were minimal. The equation for the real solutions is:

-lnaw= LmAT/(RTTo) (7)

The equation of Ferro-Fontan and Chirife (1981):

2 - In aw = 9.9693 E-3 (To - T) + 4.761 E-6 (To - T) (8) 100 GEKAS ET AL.

Table 6. Osmosities1 and water activities of other sugar solutions (Wolf et al., 1974)

Concentration Glucose Fructose Lactose (%, w/w) 6 0.194 0.993 0.192 0.933 0.103 0.996 8 0.266 0.991 0.263 0.991 0.143 0.995 10 0.361 0.988 0.338 0.989 12 0.422 0.987 0.417 0.987 14 0.506 0.984 0.500 0.984 16 0.594 0.981 0.587 0.982 18 0.687 0.978 0.677 0.981 20 0.785 0.973 0.769 0.979 22 0.892 0.969 0.887 0.970 24 1.007 0.965 0.993 0.966 26 1.124 0.962 1.100 0.962 28 1.244 0.957 1.205 0.960 30 1.369 0.953 1 Osmolality or Osmosity is the molar concentration of the isoosmotic NaCl solution, i.e. the solution of equal water activity or osmotic pressure or freezing point as the ones of the given solution

Table 7. Solute activities of salts (Vanysek, 1994)

1 Molality NaCl KC1 CaCl2 K2SO4 Na2S04 0.001 0.965 0.965 0.888 0.885 0.886 0.005 0.928 0.927 0.787 0.772 0.777 0.01 0.903 0.901 0.727 0.704 0.712 0.05 0.822 0.816 0.577 0.511 0.529

Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 0.1 0.779 0.768 0.517 0.424 0.446 0.2 0.734 0.717 0.469 0.343 0.366 0.5 0.681 0.649 0.444 0.251 0.268 1 0.657 0.604 0.495 0.204 2 0.668 0.573 0.784 0.155 5 0.874 0.593 0.591 10 0.431 OSMOTIC SOLUTIONS. I 101

where aw is water activity, Lm the molar latent heat of freezing of the pure water and AT =To - T the freezing depression value of the solution. A good agreement with water activities (effective concentrations) measured using the freezing point depression method and with textbook values for both sucrose and NaCl solutions was also found recently by Chen et al. (1996). Other measuring methods are: standardized solutions (Dora and Favetto, 1988), vapor pressure measurement, hygrometric instruments (especially the electrical hygrometer), isopiestic transfer measurement, and suction potential. Details of the various methods are discussed by Rizvi (1995), Rahman (1995), Trailer (1983), Uedaira and Uedaira (1969). There is no single method to be a good choice for all applications. The freezing point depression method has chosen for osmotic solutions based on its preference by the researchers working in the field (Marcotte and Le Maguer, 1991; Lerici et al., 1983 ).

C. AVAILABLE MODELS FOR WATER ACTTVITIES

Water Activity

Models for water activity of solutions in general were recently reviewed by Rahman (1995) and Gonzalez et al (1996). A number of the models used for the prediction of water activities are shown in Table 8. Some of the models make use of the concept of the activity coefficient, such as the Norrish, the Margules and the Crapiste ones. Others, such as Chen and Schwarzberg provide a direct correction of Raoult's law which is valid for ideal solutions. The nonideality of the solutions are due to: solute size, intermolecular forces, solvation effects, solute-solute interaction, solute-solvent interaction, dissociation effects of ionic solutes, order of mixing (Rahman, 1995). Flory and Huggins (1941), as cited by Rahman (1995), were the first to express non-ideality due to size differences between solute and solvent. Lilley and Sutton (1991) combined the effects of size, solvation and solute-solute interaction in one equation. In the absence of heterotactic interactions their model reduce to the well known multicomponent Ross equation (Rahman, 1995). The order of mixing was found to play a negligible role as mentioned by Bonne and Shannon (1991).

Osmotic Pressure

For osmotic pressure the non simplified vant Hoff equation can be used. Thus osmotic pressure is another way of expressing water activity of a solution. Plant physiologists Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 are users of the components of the "water potential" in units of pressure. The relationship between water activity and osmotic pressure is shown in Figure 2. The non simplified Van't Hoff equation is as follows:

lnaw (3)

where Vm is the partial molar volume of water. 102 GEKAS ET AL.

Table 8. List of water activity models

Equation Mathematical expression Use = 1. Norrish aw xw exp (-kx^ ) aw binary

2 2. Caurie aw = 1- (w/k) (1+ Aw + Bw ) aw binary

3. Crapiste aw /xw = exp{-A(l-xw )^ } aw binary

4. Margules aw /xwr exp(-Axs) aw binary

5. Favetto-Chirife aw = 1- km aw binary

6. Ross Modified aw = ns(aw>s)ms/m multi component

7. Caurie aw = (aw)i (aw)2 - 2 WjWj/kjkj multicomponent In the above equations, a« is activity, A is constant of non-ideality, A and B constants in the Caurie model, k is a constant defined differently in each model, x is molar feaction, m is molality. w grams per kg of water and y is activity coefficient. Superscript, q ia an exponent equal to 2 for sugars and to 1 for salts Subscripts denote, w water, s solute, i =1,2 etc denote components and T denotes total. in Equation 1 is osmotic coefficient defined as -55.51n aw/ms

Solute Activities

For solute activities in binary systems, Gibbs-Duhem theorem can be used and be solved by numerical integration.

(4)

where Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 m = -55.5 In aw

In the above equation y is the activity coefficient of the solute, m is the molality of the solute and § is known as the osmotic coefficient defined in terms of water activity and solute molality.

Temperature and pressure dependence of activities

The Clausius-Clapeyron equation is used to predict the temperature effect on water activity as: OSMOTIC SOLUTIONS. I 103

Table 9. Comparison of various activity prediction models. Modelled activities

Solution Water activity (%) Solute activity Cone. (%) 1 2 3 4 5 6 7 8 Fructose 30 95.44 95.46 95.46 95.33 2.730 50 89.22 89.42 89.33 89.22 8.534 Glucose 40 92.89 92.96 92.94 92.74 4.556 55 86.74 87.03 86.90 86.71 9.574 60 83.65 84.00 83.87 83.69 12.448 Sucrose 40 95.88 95.93 95.88 95.89 95.18 2.898 50 93.47 93.47 93.49 93.51 92.75 5.180 60 89.46 89.48 89.42 89.54 89.13 10.037 65 86.45 86.54 86.41 86.57 86.55 14.786 70 82.41 82.71 82.38 82.58 83.11 22.976 75 76.65 77.93 76.78 76.89 78.27 38.834 Sucrose 50 + NaCl 10 70.97 75.12 Sucrose 45 +NaCl 15 58.65 65.67 Note. The numbers refer to following models: 1. Norrish, 2. binary Caurie, 3. Crapiste, 4. Margules, 5. Chirife, 6. Modified Ross, 7. Caurie, and 8. Gibbs-Duhem (Equation 4)

ln(a2/a1) = - 1/T2) (5)

The effect of pressure is usually small. The Okos relationship accounts for this effect (Rahman, 1995) as:

(6)

In the above equations, the subscripts 1 and 2 refer water activity values at two different temperatures or pressures, Q is heat of sorption, R gas constant, and A^, and pw heat and density of water respectively. In Table 9 there is a comparison of water activity values obtained through the Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 use of some of the models compiled in Table 8. The concentrations in % are meant by weight (g of solute per 100 g of solution). There may be some possible variations in litterature data from different sources, such as the degree of purity, since impurities may alter the water activityof the solutions, and the degree of hydrated solutes (Reiser et al., 1995). In order to apply these models compiled in Table 8, the concentrations are also required as molalities (number of moles per kg of water) or as molar fractions (moles of solute/ total number of moles). In Table 10, the concentrations of sucrose solutions in terms of concentration (% w/w), molality and mole fraction are given along with water activity values based on the Norish model. Tables 11 and 12 provide 104 GEKAS ET AL.

Table 10. Concentration and water activity of sucrose solutions (Reiser et al., 1995)

(%)w/w Molality Molar fraction Water activity (mol/kg of water) 50.0 2.921 0.050 0.936 52.0 3.165 0.054 0.929 54.0 3.429 0.058 0.923 55.0 3.571 0.060 0.919 57.0 3.873 0.065 0.911 59.0 4.204 0.070 0.902 59.6 4.325 0.073 0.900 60.0 4.382 0.073 0.898 62.0 4.767 0.079 0.887 64.0 5.194 0.085 0.875 65.0 5.425 0.089 0.869 67.0 5.931 0.096 0.854 69.0 6.503 0.105 0.838 70.0 6.817 0.109 0.829 72.0 7.512 0.119 0.810 72.8 7.810 0.124 0.800 97.8 143.150 0.700 0.016

Table 11. Water activity of aqueous electrolyte and non-electrolyte mixtures (Rahman, 1995)

Concentrations Water Activity Norrish Experimental A. 20%w/wNaCl 0.769 0.744 20% w/w sucrose Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 B. molality Caurie Experimental sucrose 2.72 0.827 0.822 NaCl 0.75 KC1 0.40 PEG 3.68 OSMOTIC SOLUTIONS. I 105

Table 12. Water activity of starch mixtures (Rahman, 1995)

Concentration Water Activity Water Starch Sucrose Salts Ross Experimental

22% 70.2% 7.8% 0.979 0.900 44% 45% 5.5% 5.5% 0.933 0.900

additional comparison of water activity of various osmotic solutions using the Norrish, Ross and Caurie models. There are only marginal differences among the results of the various models for the osmotic solutions water activities. Then for both binary and multicomponent sugar solutions the Norrish model could be selected as a good for engineering purposes model. The model used by Crapiste is also a good for engineering purposes model and besides, it covers the case of both electrolytes and non electrolytes. For the non electrolyte, it can be shown that the Norrish, Crapiste and Margules models, with a slight different formulation, they are based on the same idea, i.e. that the logarithm of the water activity cofficient is proportional to the square of the molar fraction of the solute. Crapiste extends the applicability to electrolytes with the difference that the logarithm of the water activity coefficient is proportional to the molar fraction of the salt. The involved k or A constants in these three models are measuring the non-ideality of the solution, the higher the constants the more non-ideal becoming the system. The Norrish equation is possible to be used for other types of solutions (more "practical") as for example corn syrups, in that case the constants k for the solutes were taken as follows glucose 0.7, maltose 2.6, triose and above 2.48 (Lazarides et al., 1997; Palou et al., 1994; Palou et al, 1993). For multicomponent systems including both sugars and salts there was a worse agreement between the two models used, the modified Ross equation and the Caurie model. The latter model has been found by us as well as by others to give controversial results, it works well in some cases and not in others. In our case it gave good results in the case of sucrose solutions (but not in the case of glucose and fructose) and also in the multicomponent case with sucrose and NaCl it gave a good agreement with the

Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 experimental value. Lilley and Sutton (1991) also reported better agreement of their model than the Ross models, for the systems glucose/sucrose and glucose/glycerol up to molalities of 3 mol/kg (25°C). The agreement between modeled values and experimental values using the freezing point depression method was satisfactory. Then the aim is to obtain data for difficult multicomponent systems (such as mixtures electrolytes-non electrolytes) for which no satisfactory modeling up to now exist, through this experimantal method or try to obtain more adequate models using experimental data of this type. 106 GEKAS ET AL.

Water and solute activities from the group contribution models

Le Maguer (1992) pointed out few limitations of semi-empirical correlation models for water activity. A different approach is suggested, based on the application of fluid phase thermodynamics and excess Gibbs energy, G^, models. Attempts to use this approach have been described with encouraging results and have been reviewed by Le Maguer (1992). A further step in the use of G^ models to predict water activities in aqueous solutions consists of the use of group contribution methods. In many cases, equilibrium data involving the desired components are not available for parameter regression. In such cases, it is possible to use group contribution methods. These methods are based on the assumption that molecular interactions can be represented by the combination of interactions among the functional groups constituting them. This concept was developed for non-electrolyte solutions leading to Analytical Solution of Groups (ASOG) based on Wilson equation (Derr and Deal, 1969; Kojima and Tochigi, 1979) and UNIFAC (UNIQUAC Functional Group Activity Coefficients) based on UNIQUAC equation (Fredenslund et al., 1975). Sorrentino et al. (1986) used both ASOG and UNIFAC to predict infinite dilution activity coefficients of aroma compounds in water-carbohydrate and water- polyethylene glycol solutions and Choudhury and Le Maguer (1986) used UNIFAC to predict aw in glucose solutions. UNIFAC method has been used both by Gabas and Laguerie (1992) and Abed et al. (1992) used to predict solid-liquid equilibrium of water-sugar systems. Achard et al. (1992) on the other hand, described the use of UNIFAC-LARSEN model (Larsen et al., 1987) to estimate activity coefficients in aqueous systems containing saccharides, using the standard UNIFAC groups. The authors reported moderate, relative-deviations between experimental and predicted aw values for such systems, particularly with ternary systems. Although the ASOG group contribution method has not been so widely tested for prediction of aw, Correa and Correa (1992) and Correa et al. (1993) used the method to predict water activities in aqueous solutions of sugars and urea with polyols. Kawaguchi et al. (1981) and Correa (1997) used the same method to predict the water activity of binary and ternary aqueous electrolyte solutions. The methodology used by the latter is based on the former, but it includes simplifications with respect to anion contribution. It was realised that the electrical field created around most anions is

Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 significantly weaker than the one corresponding to cations due to their larger ionic radius. Only fluoride ion, the smallest but less frequent in these systems, has an ionic radius similar to potassium. In addition, steric hindrance between water molecules and hydration water may limit new hydration opportunities. These two aspects led to the assumption that anions remain essentially in a non-hydrated state. According to the model mentioned above, an aqueous electrolyte solution was then considered formed by water (W), hydrated cations (CH) and anions (A) leading to the following binary group interactions: water-hydrated water, water-anion, hydrated water-anion. Correa et al. (1994) used the ASOG method in order to estimate water activities of solutions of food engineering interest. Water activities in aqueous solutions of urea OSMOTIC SOLUTIONS. I 107

with sugars (glucose and fructose) and polyols (glycerol, sorbitol and mannitol) at 25°C were measured with an electric hygrometer. Concentration ranges considered in this study reached solubility limits for each solute. Correlation and prediction of water activities using ASOG group contribution method required the use of a set of new specific groups. The interaction parameters for such new groups were calculated from new and previously published experimental data. Average percent deviations of 0.4 % between experimental and predicted aw values were obtained.

A set of new ionic type functional groups for the prediction water activities (aw) in aqueous solutions of electrolyte solutes using ASOG group contribution method is proposed. Previously published experimental data on water activities, osmotic coefficients and freezing temperatures for binary solutions of electrolyte salts and water at different temperatures were used to calculate interaction parameters. With such parameters values of aw for binary (14), ternary (28) and quaternary (3) systems, at different temperatures, were predicted and compared with experimental data. This data included both published and new data, measured with an electric hygrometer, for sodium nitrate (at 20°C) and potassium nitrate (at 20°C and 30° C). Calculated average relative deviations of aw predictions using the ASOG method were 0.21%, 0.28% and 0.20% respectively. On the basis of the results obtained for the prediction of water activities in aqueous solutions of urea with either sugars or polyols and of other sugar/sugar and sugar/polyol solutes, it can be concluded that ASOG group contribution method as described by Kojima and Tochigi (1979) complemented by a set of five new interaction groups proposed here, was able to produce results with an average relative deviation of 0.4 %, which can be considered very acceptable and suggests the possibility of its extension to other similar systems. Concerning electrolyte solutions and to check the applicability of the proposed model, water activities predicted by this method were compared with the predictions obtained by Teng and Seow (1981) using the Ross, and modified Ross methods. Results obtained are clearly better than the ones obtained with Ross's method and a little worse than modified Ross methods. It should be stressed, however, that ASOG predictions were based on general group contributions obtained from data obtained for completely different systems. The other mentioned methods represent essentially interpolating models requiring experimental binary data for all pairs of the system being studied, at the desired temperature; in principle, any extrapolation to other conditions or system is not possible. Interaction parameters presented were calculated from different types of Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 experimental data obtained at several temperatures. This single set parameters was able to make aw predictions for other binary, ternary and quaternary systems at different temperatures with acceptable deviations from experimental data.

D. CURRENT LIMITATIONS AND FUTURE DIRECTION

Water activity data is possible to find in the litterature. Those data are still good for engineering purposes and to develop simplified semitheoretical models. To the author's opinion approaches such as the Lilley Sutton (1991), and use of the Gibbs Duhem 108 GEKAS ET AL.

Sucrose solutions 40%- 75%

0,8 0,9 1.0 water activity

Figure 1. Dependency of solute concentration and water activity of a sucrose solution.

4,00e+6

•fa 3,00e+6 Q.

0,00e+0 40 50 60 70 80 Concentration (%)

Figure 2. Relationship between osmotic pressure and concentration for sucrose solution. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011

theorem for solute activities are most promising. Application of group contribution models are also at their infancy as far as application to the osmotic solutions is concerned.

Future Needs

1. Although water activity data exist in many cases of osmotic solutions, the expression of the osmotic capacity of the media in other equivalent terms such as OSMOTIC SOLUTIONS. I 109

osmotic pressure or osmosity could be helpful since water activity is not sensitive in the region 0.9-1.0. (small differences in a«, give high differences in "Ina«,"). This remark should also to be considered in the case of sorption isotherms in the high range of water activity.

2. More fundamental effords are needed to estimate water and solutes' activities of mixtures and solutes interations.

3. In general, solute activity data are very scarse and there is a future need to obtain such data for the various osmotic solutions. Our results show that solute activity data for concentrated sugar solutions are promising (Figure 1). The application of the water activity in order to estimate osmotic pressure of concentrated solutions give high pressure values in the order of 100 MPa (Figure 2). Solute activity growing smoothly up to more or less 50% sugar concentration shows an exponential trend as the concentration approaches the sugar solubility limit. This fact could explain certain observations from studies of osmotic dehydration reported in the literature, for example the levelling off water loss and solid uptake attained at high concentration values. It could also reveal other kinds of non idealities, probable inflecion points etc. In conclusion, both activity coefficient models and group contribution models could be considered to fill the gap.

REFERENCES

Abed, Y., Gabas, N., Delia, M. L., and Bounahmidi, T., 1992. Measurement of liquid- solid phase equilibrium in ternary systems of water-sucrose-glucose and water- sucrose-fructose, and predictions with UNIFAC. Fluid Phase Equilibria. 73: 175-184. Achard, C., Dussap, C. G., and Gros, J.B., 1992. Prédiction de l'activité de l'eau, des températures d'ébullition et de congélation de solutions aqueuses de sucres par un modèle UNIF AC. Industries Agricoles et Alimentaires. 109 (March): 93- 101. Caurie, M. 1986. A general method for predicting water activity of simple and multicomponent mixtures of solutes and non-solutes. Journal of Food Technology. 21: 221-228. Chen, P., Chen, X. D., and Free, K. W. 1996. Measurement and data interpretation of the freezing point depression of milks. Journal of Food Engineering. 30: 239- Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 253. Chirife, J., and Resnik, S. L. 1984. Saturated solutions of sodium chloride as reference sources of water activity at various temperatures. Journal of Food Science. 49: 1486-1488. Choudhury, G. S., and Le Maguer, M. 1986. Prediction of water activity in glucose solution using the UNEFAC model. In: Food Engineering and Process Applications. Le Maguer, M., and Jelen, P. eds. Elsevier, New York. Correa, A. 1997. Personal communication. Correa, A., and Correa, J. M. 1992. Prediccion de la actividad del agua en disoluciones HO GEKASETAL.

acuosas de electrolitos. Communication to 24th Reunion Bienal de la R.S.E.Q., Malaga, Spain. Correa, A., Comesaña, F., and Sereno, A. M. 1993. Prediction of water activity in non- electrolyte solutions by ASOG group contribution method. Communication to CHEMPOR'93, International Chemical Engineering Conference, Porto, Portugal. Correa, A., Comesaña, F., and Sereno, A. M. 1994. Measurement of water activity in water-urea-sugar and water-urea-polyol systems and its prediction with ASOG group contribution method. Fluid Phase Equilibria. 98: 189-198 Crapiste, G. H. 1985. Fundamentals of Drying of Fruits and Vegetables. PhD Thesis, Bahia Blanca, Argentina. Derr, E. L., and Deal, C. H. 1969. Analytical solution of groups: correlation of activity coefficients through structural group parameters. Institution of Chemical Engineers Symposium Series. 32: 44-51. Dora, K., and Favetto G. 1988. Mixed Saturated Salt Solutions as standards for water activity measurement in the microbiological growth range. Journal of Food Science. 53: 578-579. Favetto, G., Resnik S., and Chirife J. 1983. Statistical Evaluation of water activity measurements obtained with the Vaisala Humidity Meter. Journal of Food Science. 48:534-538. Ferro-Fontan, C. and Chirife, J. 1981. A refinement of Ross's equation for predicting the water activity of Non-electrolyte mixtures. Journal of Food Technology. 16, 219-221. Ferro-Fontan, C., and Chirife J. 1980. The prediction of water activity of aqueous solutions in connection with intermediate moisture foods III. Journal of Food Technology. 5:47-58. Ferro-Fontan, C., and Chirife, J. 1980. The prediction of water activity of aqueos solutions in connection with intermediate moisture foods IV. Journal of Food Technology. 15:59-70. Fito, P., Andres, A., Pastor, P. and Chiralt, A. 1994. Vacuum Osmotic Dehydration of Fruits. In Minimal Processing of Foods and Process Optimization. Singh, R. P., and Oliveira, F. eds. CRC Press, Boca Raton, p. 107 - 121. Fredenslund, A. A., Jones, R. L., and Prausnitz, J. M. 1975. Group contribution estimation of activity coefficients in non ideal liquid mixtures. AIChE Journal. 21: 1086-1099 Gabas, N., and Laguérie, C. 1992. Modelling of liquid-solid phase equilibria with UNIFAC. Application to sugar systems. Communication to European Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 Symposium on Computer Aided Process Engineering - ESCAPE 2. Toulouse, France. Geankoplis, C. 1993. Transport Processes and Unit Operations. Third edition. Prentice-Hall, NJ. Gekas, V. 1992. Transport Phenomena of Foods and Biological Materials. Singh, R. P., and Heldman, D. eds. CRC Press, Boca Raton. Gonzalez, C. 1996. Osmotic Dehydration of Apples - Some Equilibrium Studies, Master Thesis, Lund University and Valencia University. Gonzalez, C., Gekas, V., Fito, P., Lazarides, H., and Sjöholm, I. 1996. Osmotic solution characterization, In Minimal Processing of Foods and Process OSMOTIC SOLUTIONS. I 111

Optimization. Vol. 3. Singh, R. P., and Oliveira, J. eds. CRC Press, Boca Raton. p. 76-82. Haase, R. 1990. Thermodynamics of Irreversible Processes. Second edition. Dover Publications, NY. Kawaguchi, Y., Kanai, H., Kajiwara, H., and Arai, Y. 1981. Correlation for activities of water in aqueous electrolyte solutions using ASOG model. Journal of Chemical Engineering Japan. 14 (3): 243-246 and 15 (1): 62-63, 1982. Kojima, K., and Tochigi, K. 1979. Prediction of Vapour-Liquid Equilibria by the ASOG Method. Elsevier, Tokyo, p. 11-27. Labuza, T. P. 1984. Moisture Sorptions: Practical Aspects of Isotherm Measurement and Use. American Association of Cereal Chemists, St Paul, MN. Larsen, B. L., Rasmussen, P., and Fredenslund, A. 1987. A modified UNIFAC group- contribution model for prediction of phase equilibria and heats of mixing. Industrial and Engineering Chemistry Research. 26 (11): 2274-2286. Lazarides, H. 1994. Osmotic preconcentration- develpoments and prospects. In: Minimal Processing of Foods and Process Optimization, Singh, R. P., and Oliveira, F. eds. CRC Press, Boca Raton, p. 73 - 84. Lazarides, H., Gekas, V., and Mavroudis, N. 1997. Mass diffusivities in fruit and vegetable tissues undergoing osmotic processing. Journal of Food Engineering. 31(3): 315-324. Le Maguer, M. 1992. Thermodynamics and vapour-liquid equilibria. In: Physical Chemistry of Foods. Schwartzberg, H. G., and Hartel, R.W. eds. Marcel Dekker, New York. Lenart, A. 1994. Osmotic dehydration of fruits before drying, developments and prospects. In: Minimal Processing of Foods and Process Optimization. Singh, R. P., and Oliveira, F. eds. CRC Press, Boca Raton, p. 87-106. Lenart, A., and Flink, J. M. 1984a. Osmotic concentration of potato. I.Criteria for the end-point of the osmosis. Journal of Food Technology. 19: 45-63. Lenart, A., and Flink, J. M. 1984b. Osmotic concentration of potato. II. Spatial distribution of osmotic effect. Journal of Food Technology. 19: 45-63. Lerici, C., Pinnavaia, G., and Dalla Rosa, M. 1983. Water activity and freeezing point depression of aqueous solutions and liquid foods. Journal of Food Science. 48: 1667-1669. Lerici, C., Pinnavaia, G., Dalla Rosa, M., Bartilucci, L. 1985. Osmotic dehydration of fruits: influence of osmotic agents on drying behaviour and product quality. Journal of Food Science. 50: 1217-1226. Lilley, T., and Sutton, R. 1991. The prediction of water activities in multicomponent Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011 systems. In: Water Relationships in Food Levine, H, and Slade, L. eds. Plenum Press, NY. p. 291-313. Marcotte, M., and LeMaguer, M. 1991. Repartition of water in plant tissues subjected to osmotic processes. Journal of Food Process Engineering. 13: 297-320. Palou, E., Lopez Malo, A., Argaiz, A., and Welti, J. 1993. Osmotic dehydration of papaya. Effect of syrup concentration. Revista Espanola de Ciencia y Tecnologia de los Alimentos. 33(6): 21-630. Palou, E., Lopez Malo, A., Argaiz, A., and Welti, J. 1994. Osmotic dehydration of papaya with corn syrup solids. Drying Technology. 12(7): 1709-1725. 112 GEKASETAL.

Rahman, M. S. 1995. Food Properties Handbook. CRC Press, Boca Raton, FL. Reiser, P., Birch, G. G., and Mathlouti, M. 1995. Physical Properties. In: Sucrose, Properties and Applications. Mathlouti, M., and Reiser, P. eds. Blackie Academic & Professional, Glasgow, p. 186-222. Rizvi, S. S. H. 1995. Thermodynamic properties of foods in dehydration. In: Engineering Properties of Foods. Second edition. Rao, M. A., and Rizvi, S. S. H. eds. Marcel Dekker, New York. Sorrentino, F., Voilley, A., and Richon, D. 1986. Activity coefficients of aroma compounds in model food systems. AIChE Journal. 32 (12): 1988-1993. Stokes, R. H., and Robinson, R. A. 1966. Interactions in aqueous nonelectrolyte solutions. 1. Solute-solvent equilibria. Journal of Physical Chemistry. 70: 2126- 2130. Teng, T., and Seow, C. A. 1981. Comparative study of methods for prediction of water activity of multicomponent aqueous solution. Journal of Food Technology. 16: 409-419. Tochigi, K., Tiegs, D., Gmehling, J., and Kojima, K. 1990. Determination of new ASOG parameters. Journal of Chemical Engineering Japan. 23 (4): 453-463. Trailer, J. A. 1983. Water activity measurements with a capacitance manometer. Journal of Food Science. 48: 739-741. Uedaira, H., and Uedaira, H. 1969. Activity coefficients of aqueous xylose and maltose solutions. Bulletin of Chemical Society of Japan. 42: 2137-2140. Vanysek., P. 1994. Handbook of Chemistry and Physics. CRC Press, Boca Raton, FL. Warin, F., Gekas, V., Voirin, A., and Dejmek, P. 1997. Migration of sugar in agar gel/milk bilayer systems. Journal of Food Science, (in press). Wolf, A. V., Brown, M. G., and Prentiss, P. B. 1974. Concentrative properties of aqueous solutions: Conversion Tables. In: Handbook of Chemistry and Physics. 55th Edition. p. D-194. Downloaded by [b-on: Biblioteca do conhecimento online UP] at 16:02 24 October 2011