419 Journal of Food Protection, Vol. 51, No. 5, Pages 419-423 (May, 1988) Copyright© International Association of Milk, Food and Environmental Sanitarians Proposed Theoretical Water Activity Values at Various Temperatures for Selected Solutions to be Used as Reference Sources in the Range of Microbial Growth SILVIA L. RESNIK1* and JORGE CHIRIFE Downloaded from http://meridian.allenpress.com/jfp/article-pdf/51/5/419/1658009/0362-028x-51_5_419.pdf by guest on 26 September 2021 Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria, 1428 Buenos Aires, Republic of Argentina (Received for publication March 26, 1987) ABSTRACT based on the thermodynamic properties of strong electro­ lyte aqueous solutions, and was applied either to satu­ This paper compiles recent data on the theoretical prediction rated or unsaturated solutions at different temperatures. of the water activity (a ) of selected saturated salt solutions, w The system of equations developed by Pitzer (7) and his unsaturated NaCl and LiCl solutions, and H2S04 solutions. These results are presented in tabular form in such a way that associates (8,9,14) was successfully used to predict the they can be safely used as reference sources for aw determina­ aw of selected saturated salt solutions, unsaturated NaCl tion in the range of microbial growth (i.e. about 0.57-0.97) and and LiCl solutions (2,6), and H2S04 solutions (13). at different temperatures. It is the purpose of the present work to compile all these results in such a way that they can be safely used as reference sources for aw determinations in the range Adjustment of the water activity (aw) of foods is an of microbial growth (i.e. about 0.57-0.97) and at differ­ important method of controlling microbial spoilage (14). ent temperatures. This has been recognized by regulatory agencies in sev­ eral countries who have incorporated the aw principle in EXPERIMENTAL METHODS the definition of standards for various preserved foods. The U.S. Food and Drug Administration has adopted A brief explanation is given on how the predicted aw values were obtained. the water activity specification in the definition of low- The osmotic coefficient (<)>) is given by the equations de­ acid food (5) as has the FAO/WHO Codex Alimentarius veloped by Pitzer (7) as, Comission (4) and the European Community for the trade of meat products (3). The reliability of most aw measur­ 2 x <H = |ZMZX| f + m (^^)BMX + m (^L f CMX [1] ing devices such as electric or fiber-dimensional hy­ v v grometers, strongly depends on proper calibration using where, vM and vx are the number of M and X ions in the reference sources of aw (usually saturated or unsaturated formula and ZM and Zx give their respective charges in elec­ salt solutions). Unfortunately, there is not an agreement tronic units; also, vXvM + vx. The other quantities have the on the exact water activity of the different reference form: sources, mainly saturated salt solutions (12). It is obvious that for different regulatory agencies to adopt an aw specification for food, and for different researchers to re­ 1 + bl produce each others aw measurements, there must be "universal" agreement on the values to be assigned to re­ ference standard solutions (either saturated or unsatu­ ( ) , 2 rated). This agreement should include not only 25°C BMX = B S1 x + B^ x-exp(-al" ) [3] (usual temperature reference) but also other temperatures where, of interest to microbial growth. In the past few years several works were done at this I: ionic strength = 1/2 £ m{Z] laboratory on use of theoretical models to predict the aw of reference solutions. This theoretical approach was A is the Debye-Huckel coefficient for the osmotic function and has a value of 0.392 at 25°C. Constant b was taken equal to 'Member of Comision de Investigaciones Cientificas de la Provincia de 1.2 for all solutes and also the value of a = 2 was found to Buenos Aires. be satisfactory for all solutes considered in the present work. JOURNAL OF FOOD PROTECTION, VOL. 51, MAY 1988 420 RESNIK AND CHIRIFE In the case of H2S04 solutions the equations also considered RESULTS AND DISCUSSION + = the dissociation HSO"4 -* i H + SO 4 (9). Values of pa­ The predicted aw of selected saturated salt solutions in (0) (,) rameters B , B and the approximate range 0.75-0.98 and between 10 and CMX for the solutes of interest to this 37CC is shown in Table 1; it is believed that these values work are reported in the literature (8,14). The parameters 6<0), are "accurate" to about 0.002-0.003 aw (6,11). Results 8(1) and CMX are described by arbitrary functions of temperature for other saturated salt solutions commonly used in this f(T) (6,14). The osmotic coefficient was cowerted to a accord- w range, such as NaBr and KN03 are not given because ing to of some discrepancies between prediction and measured -55.51 In aw [4] 4> = vm values. Table 2 shows the aw of unsaturated NaCl solutions at any temperature between 15 and 50CC. As shown by where v is the number of particles into which the solute of Chirife and Resnik (2), the a of NaCl solutions in the molality m dissociates. To calculate the a of saturated salt sol­ w w microbiologically important range, 15C-50CC, may be utions, the molality of solute at saturation (ms) was used in equation [4]. considered almost equal; maximum error is about 0.002 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/51/5/419/1658009/0362-028x-51_5_419.pdf by guest on 26 September 2021 aw. Of course, NaCl solutions are useful as standards only for values of aw above = 0.76. Table 3 shows the aw of LiCl solutions between 5 and Accuracy of the predictions 45CC (6). In this instance, the available data extend to The thermodynamic properties of most common inorganic the lower range of microbial growth (about 0.6 aw). Most electrolyte solutions have been exhaustively determined over the of these values may be considered "accurate" to 0.002 past 50 years. Dozens of workers in the physical-chemistry area aw; only at low aw (i.e. below about 0.7 aw) the error determined the activity of water in solutions of NaCl, KC1, may be somewhat higher (i.e. up to about 0.005 aw). LiCl, H2S04, and the like, and reported their data as osmotic It is to be noted that at water activity values above about coefficients for the whole range of concentration and at various C 0.90 the effect of temperature (between 5 and 45 C) on temperatures. Experimental determinations were based on pre­ a of LiCl solutions does not exceed the 0.002 a level. cise and accurate methods, mainly isopiestic ones. From time w w to time well known researchers critically evaluated the literature At lower water activities the effect of temperature be­ data on measurement of the osmotic coefficient of most com­ comes more significant. Finally, Table 4 (13) shows the mon electrolytes and published "best" values of <(> for rounded predicted aw of sulfuric acid solutions between 0 and molalities. Platford (10) critically reviewed the accuracy of 55°C. In this instance, the data cover the entire range isopiestic determinations and concluded that most of this compi­ of microbial growth. These data may be considered "ac­ lations are in general accurate to about 0.001 in <(>, which curate" to the 0.002-0.003 level. As observed with LiCl means that aw values may be considered in most instances to solutions, the effect of temperature on the aw of H2S04 be accurate to about 0.001-0.002 aw. solutions becomes more noticeable at lower water ac­ As discussed later, the predicted values were compared to tivities (i.e. more concentrated solutions). these tabulated values (compilations) and the deviation between The data in Tables 1 to 4 are thus recommended as predicted and tabulated (measured) values were used as an indi­ reference sources of a at various temperatures in the cation of the "accuracy" of the predictions. Admittedly, this is w not accurate unless the measured values were absolutely accu­ range of microbial growth. The accuracy of these pro­ rate which is not true; however, since the accuracy of the tabu­ posed values fulfill the actual needs for aw measuring in lated values largely exceeds the needs in food microbiology re­ food microbiology research and development. It is ex­ search (i.e. 0.005 aw), we arbitrarily named "accuracy" to the pected that this proposal may contribute to reach a "uni­ deviation between preditions and tabulated compilations. versal" agreement on the values to be assigned to solu­ tions used as reference sources of aw. TABLE 1. Predicted aw of selected saturated salt solutions between 10°C and 37°Ca Temp. (°C) NaCl (NH4)2S04 KC1 BaCl2 K2S04 10 0.754 0.809 0.867 0.913 0.980 15 0.753 0.808 0.859 0.910 0.979 17 0.753 0.806 0.856 0.909 0.978 19 0.752 0.805 0.852 0.907 0.977 21 0.752 0.804 0.849 0.906 0.977 23 0.751 0.803 0.846 0.905 0.976 25 0.751 0.803 0.842 0.903 0.975 27 0.750 0.802 0.840 0.902 0.975 29 0.750 0.801 0.836 0.900 0.974 31 0.750 0.800 0.833 0.899 0.973 33 0.749 0.799 0.830 0.898 0.973 35 0.749 0.798 0.827 0.895 0.972 37 0.748 0.797 0.823 0.894 0.971 aData at I0°C are from Pollio et al.