Theoretical Prediction of the Water Activity of Standard
Saturated Salt Solutions at Various Temperatures
DORA KITIC, DENISE C. PEREIRA JARDIM, GUILLERMO J. FAVETTO, SILVIA L. RESNIK, and JORGE CHIRIFE
ABSTRACT
A theoretical approach was used to predict the water activity (a,) at
15°C and 35°C of selected saturated salt solutions used as standards
in the range of microbial growth (20.57-0.97). For this purpose a
rigorous thermodynamic model to predict osmotic coefficient was used.
The results agreed very well with experimental measurements at 15°C
and 35°C for most salts studied, which included NaBr, NaCl,
(NH&S04, KCl, BaC12, and KzS04. For KNOs the agreement at the
higher temperature was somewhat less satisfactory and the reasons for
this behavior are discussed.
INTRODUCTION
IT IS WELL KNOWN that adjustment of the water activity
(a,) of foods is an important method of controlling spoilage
by microorganisms (Troller and Christian, 1978). This has been
recognized by regulatory agencies in several countries who
have incorporated the a, principle in the definition of standards
for various preserved foods.
The U.S. Food & Drug Administration has adopted the water
activity specification in the definition of low-acid foods (FDA,
1973) as has the FAO/WHO Codex Alimentarius Commission
(FAO/WHO, 1979) and the European Community for the trade
of meats products (EEC, 1976). The reliability of most a,
measuring devices such as electric or fiber-dimensional hygro-
meters, strongly depends on proper calibration using reference
sources of a,. Saturated salt solutions are used by a large
number of people as a convenient way to provide solutions of
known a,. They are reproducible reference standards because
no measurement of concentration is needed and if the salts are
properly chosen no interfering vapors are present (Stoloff, 1978;
Gal, 1981).
Considerable data and several reviews are available on the
a, of standard salt slurries (Stokes and Robinson, 1949; Wex-
ler and Hasegawa, 1954; Young, 1967; Rockland, 1960; In-
ternational Critical Tables, 1926; Labuza et al., 1976; Greenspan,
1977). Unfortunately, most reports in the literature do not agree
on the exact a, of each saturated salt solution. Considerable
doubt has arisen about the accuracy of water activity of satu-
rated salt solutions in view of the recent review by Greenspan
(1977). His work reflects very well the uncertainties of liter-
ature values from various workers. Recently, Resnik et al.
(1984) reported the results of a world survey of a, values at
25°C of selected saturated salt solutions used as standards in
the range of a, 0.57-0.97, which is of most interest to micro-
bial growth. Their results confirmed the discrepanciesbetween
laboratories regarding the correct a, to assign to each salt. The
Authors Kitic, Favetto, and Chirife are with PRO/PA (CONICET-
FCEyN), Depto. de Industrias, Facultad de Ciencias Exactas y
Naturales, Univ. de Buenos Aires, 1428 Buenos Aires, Republica
Argentina. Author Pereira Jardin is with lnstituto de Tecnologia
de Alimentos (ITAL), Campinas, S.P., Brasii. Author Favetto is a
member of CONICET, Argentina. Author Resnik is a member of
Comisi& de lnvestigaciones Cientificas de la Provincia de Bue-
nos Aires, Argentina.
Volume 51, No. 4, 1986JOURNAL
survey was limited to 25°C but it was expected that the dis-
crepancies would be even larger if other temperatures were
considered. It is obvious that for different regulatory agencies
to adopt an a, specification for foods, and for different re-
searchers to reproduce each others a, measurements, there
must be an “universal” agreement on the values to be assigned
to reference standards of saturated salt solutions. This agree-
ment should include not only 25°C but also other temperatures
of interest of microbial growth.
The purpose of this study was to use a theoretical approach
based on the thermodynamic properties of strong electrolyte
aqueous solutions to calculate the a, of selected saturated salt
solutions in the temperature range 15-35°C. The study covered
the range of a, of most interest to microbial growth (=0.57-
0.97). The predicted results were compared with experimental
data obtained with an electric hygrometer previously calibrated
with unsaturated salt solutions for which agreement on their
a, at 15°C and 35°C existed.
MATERIALS & METHODS
Materials
Saturated salt solutions were prepared by mixing a large excess of
each salt and distilled water and stored at 15°C (?0.2”C) or 35°C
(kO.2”C) in a forced convection low-temperature incubator for var-
ious days before using. Seven reagent grade salts were used: sodium
bromide, sodium chloride, ammonium sulfate, potassium chloride,
barium chloride, potassium nitrate and potassium sulfate. These salts
have been used extensively as reference standards of water activity in
the range of interest to the present study. Unsaturated solutions of
sodium chloride and lithium chloride were prepared using anhydrous
reagent grade salts for calibration of the hygrometer at 15°C and 35°C.
Measurement of a,
The a, of the six saturated salt solutions at 15°C and 35°C was
measured using a Novasina Humidat-TH2 Thermoconstanter hygro-
meter (Novasina AC, Ch-8050 Zurich, Switzerland). This instrument
has a built-in accurate temperature control device which maintains the
sensor and the sample at constant temperature in the range 0-50°C.
The hygrometer was operated following the procedure described in
detail by Kitic et al. (1986). Each a, measurement represents the
average of three determinations. As shown previously (Kitic et al.,
1986), three replicates are enough to obtain a precision of f 0.005
a, when using the Novasina Thermoconstanter for measuring the a,
of saturated salt solutions in the range of interest of this study. To
measure the a, of saturated salt solutions at 15°C and 35°C the in-
strument was first calibrated at these temperatures using other stan-
dards for which agreement exists on their a, at 15°C and 35’C. Chirife
and Resnik (1984) and Kitic et al. (1986) proposed the use of unsat-
urated solutions of NaCl and LiCl as isopiestic standards for calibra-
tion of a, measuring devices at various temperatures and showed that
there was excellent agreement (theoretical and experimental values)
on the exact value of the a, of these solutions; thus these values were
used for the calibration of the Novasina Thermoconstanter at 15°C
and 35°C. Table 1 shows the data used for the calibration; at least six
points were used at each temperature. NaCL solutions are useful as
standards only for values of a, of about 0.75 and above (saturation
value); for this reason LiCl solutions were used as standards for lower
a, values. Equilibration times for measuring a, of each saturated salt
were previously determined and ranged between 30-90 min.
OF FOOD SCIENCE-1037
a, OF SALT SOLNS AT VARIOUS TEMPS . . .
Table l--Unsaturated salt solutions used as isopiestic standards for cal- Table 2-Values of the parameters in Eq. (2) for some selected salt so-
ibration of the Novasina Thermoconstanter hygrometer at 15°C and 35°C lutions at 25”Cs
LiCP Salt p pll, CMX mmaxb
Concentratio+ a, a, NaBr 0.0973 0.2791 0.00116 4
% (w/w) (15°C) (35°C) NaCl 0.0785 0.2664 0.00127 8
25.97 0.505 0.521 NW,)304 0.0409 0.8585 -0.00116 5.5
25.00 0.536 0.550 KCI 0.04835 0.2122 - 0.00084 4.8
23.98 0.568 0.581 BatI& 0.2628 1.49825 -0.01938 1.8
23.00 0.597 0.609 KNOB -0.0816 0.0494 0.00660 3.8
22.98 0.598 0.609 K.S.0, 0.04995 0.77925 __- 0.7
22.51 0.612 0.623 a Pitzer and Mayorga (1973)
18.01 0.732 0.739 b Maximum molality for which agreement with experimental data was attained to
17.49 0.744 0.751 0.01 in $ (or for which data were available).
17.48 0.745 0.751
17.36 0.748 0.754
16.98 0.756 0.762
16.39 0.770 0.775 tronic units; also v = u, + x; the other quantities have the
NaCIc form:
Concentratior@ a, aw
% (w/w) (15°C) (35°C)
23.50 0.795 0.795
22.50 0.808 0.808
21.50 0.821 0.821 BMX =. p& + p& exp (- ~1I)“2
20.50 0.833 0.833 (0) (1)
19.02 0.852 0.850 pMx, pMx and CMx are parameters which are fixed for each
18.38 0.859 0.857
17.70 0.867 0.864 electrolyte, A, is the Debye-Hiickel coefficient for the osmotic
17.01 0.874 0.872 function, I is the ionic strength % Z miZi*; b is equal to 1.2
15.25 0.891 0.889 for all solutes and (Y = 2.0 was found to be satisfactory for
13.42 0.908 0.906 all solutes considered by Pitzer and Mayorga (1973).
11.98 0.921 0.920
10.71 0.930 0.929 The ion-interaction model of Pitzer (1973) has been very
8.25 0.949 0.948 successful in describing osmotic results, both at 25°C (Pitzer
6.18 0.963 0.962 and Mayorga, 1973) and at elevated temperatures (Silvester
2.44 0.986 0.986 and Pitzer, 1977; Holmes and Mesmer, 1983) for different
0 Kitic et al. 0966) electrolyte solutions.
bg of anhydrous salt per 1OOgof solution
CChirife and Resnik (1964)
Table 2 shows .the values of the parameters which describe
the a, at 25°C of the salt solutions considered in this work,
namely NaBr, NaCl, (NH4)*S04, KCl, BaC12, KN03 and
K2S04, as reported by Pitzer and Mayorga (1973). These val-
RESULTS & DISCUSSION ues were mostly obtained from careful isopiestic deterrnina-
tions of vapor pressures in such a way that they would allow
Theoretical prediction of the a, of aqueous accuracy to the third (or even fourth) decimal figure in pre-
strong electrolyte solutions dicted a,. In order to predict the a, of the corresponding
saturated salt solution at 25”C, one only needs to have a precise
The thermodynamic properties of aqueous strong electrolyte knowledge of the molality at saturation (m,), and then the a,
solutions have been extensively studied both experimentally at saturation may be calculated through the relationship
and theoretically. Excellent examples may be found in the
standard monographs by Hamed and Owen (1958) and Robin- (aJsat = exp( - 4 0.018 m,v) (3)
son and Stokes (1965) or in the more recent works of Scatchard
(1968), Scatchard et al. (1970), Pitzer (1973) and Bromley Tran and Lenzi (1974) demonstrated that Pitzer’s equation may
(1973). One of the most useful and accurate theoretical models be safely extrapolated even to predict a, of various supersa-
to predict the a, of salt solutions is that of Pitzer (1973), (Pitzer turated salt solutions. Chirife et al. (1983) successfully applied
and Mayorga, 1973). Pitzer (1973) developed a system of Pitzer’s model to predict the a, of saturated NaCl, (NH&S04,
equations (ion-interaction model) for the thermodynamic prop- KCl, BaC12, KN03, and K2S04 at 25°C.
erties of strong electrolytes which yield agreement within ex-
perimental error up to concentrations of several mol/kg (in a, of saturated salt solutions at temperatures
numerous cases very close to saturation). It is interesting to other than 25°C
note that the fit was particularly good for l-l and 2-l electro-
lytes, which are those of concern in the present study. Chirife et al. (1983) have shown that the a, of certain sat-
urated salt solutions at 25°C may be predicted accurately. In
The osmotic coefficient, 4, is defined as order to predict the a, at different temperatures two aspects
-55,51 In a, have to be taken into account: (1) the effect of temperature on
4= (1) the salt solubility, usually, an increase of temperature will
Vi ml increase the amount of soluble salt thereby decreasing a, if
where vi is the number of particles into which each solute of other temperature effects are absent; (2) the thermodynamic
effect of temperature on a, at constant concentration of the
molality mi dissociates. The osmotic coefficient is given by solution; this im lies that the temperature dependence of pa-
the equations developed by Pitzer (1973) as:
rameters p(O), p( P) and C MX in the predictive equation (Eq. 2)
must be known. Thus, the effect of a temperature change on
the a, of a saturated salt solution will be the result of these
BMX
\ - I two factors.
Silvester and Pitzer (1977) extended the ion-interaction model
+ m*2 03/2
V
cMx (2) (Eq. 2) for the calculation of activity and osmotic coefficients
of NaCl in a wide range and applied the equations to an ex-
where, w and ux are the number of M and X ions in the tensive array of the thermodynamic data of aqueous NaCl from
formula and ZM and Zx give their respective charges in elec- 0” to 300°C. The general forms to describe the temperature
1038-JOURNAL OF FOOD SCIENCE-Volume 51, No. 4, 1986
Table 3-Prediction of a,,, of NaCI solutions at different temperatures* - Table 5-Parametersa of the temperature derivatives in Eq. (7) for some
Coefficients for Eq. (4), (5) and (6) selected salt solutions
q1 = 0.0765 q9 = 6.1608 x IO-5 ap@) apfl) acMx
- 777.03 1.0715 x 10-e x 104 x 104 x 105
q2 = %o = Salt 7 7 7 mbmax
q3 = - 4.4706 911 = 0.00127
0.008946 33.317 NaBr 7.6g2 10.7s -9.30 9
q4 = cl12 =
-3.3158 x 10-e 0.09421 NaCl 7.159 7.005 -10.54 6
% = q13 = ___ ___ --_ ___
0, = 0.2664 0.” = -4.655 x 10-s W ’W,SOa
KCI 5.794 10.71 - 5.095 4.5
* Silvester and Pitzer (1977) BaCI, 0.854 4.3, -2.3 1.8
KNO, 2.06 64.5 39.7 2.4
KzSQ, 1.92 8.g3 -__ 0.1
Table 4-Prediction of a, of KCI solutions at different temperaturese - a Silvester and Pitzer (1978).
Coefficients for Eq. (7) b Maximum molalities to which the enthalpy equations were fined.
pw p(l) CMX
Pl 0.04808 0.0476 -7.88 x 1O-4 Table S-Solubility of selected standard salt solutions at various tem-
P2 -758.48 303.9 91.270 oeratures
P3 -4.7062 1.066 0.58643
P4 0.010072 --_ -0.0012980 Solubilitya
P5 -3.7599 x 10-s ___ 4.9567 x lO-7 (g of anhydrous salt par IOOg of solution)
on ___ 0.0470 _-- Salt 15°C 25°C 35°C
B Holmes and Mesmar (1983) NaBr 46.58 it 0.10 48.61 f 0.10 50.48 f 0.48
NaCl 26.35 f 0.03 26.50 f 0.10 26.61 f 0.05
W-L&SO., 42.53 k 0.17 43.42 e 0.10 44.22 + 0.20
KCI 24.64 k 0.20 26.44 k 0.10 27.93 k 0.10
dependenceof B(O),B(l) and C&x for NaCl solutions are given BaCI, 25.68 f 0.10 27.04 f 0.15 27.90 f 0.15
as: KNO, 20.50 f 0.15 27.54 + 0.24 34.84 -c 0.24
KzSO4 9.21 + 0.04 10.73 + 0.06 12.17 + 0.05
s Literature mean value (calculated from several different sources) and estimated
p(O) = qt + 92 (: - f) + q3 ln(+) standard deviation.

; q4(T - '6 + qs(T* - T,*, (4)

Table 7-Comparison of predicted and measured values of a, of selected
saturated salt solutions at various temoeratures
p(l) = q6 + q9(T - T,) + qlo(T’ - ‘L2, (5) Temp Predicteda Predicted* Measured
Salt I%) Ib1 lc) Id1
NaBr 15 ___ 0.601 f 0.002 0.603
CMX = $1 + %2($ - f) 25 0.562 f 0.002 ___ ___
r ---
35 0.525 T 0.008 0.537

+ q13 ln f + qdT - Tr) (6) NaCI 15 0.752 k 0.001 0.753 k 0.001 0.753
25 0.751 +- 0001 ___ ___
r
35 .._. f 0.001
0.750 __--. 0.749 I 0.001 0.751
where Tn is 298.15”C and the values of the parameters PJH&S04 15 --_ __- 0.804
qr . . . . q14 are given in Table 3. These values, along with 25 0.803 k 0.001 ___ ___
35 ___ ___ 0 797
Bq. (3), (4), (5), (2) and (1) allow a rigorous and accurate
prediction of the a, of NaCl solutions for the whole concen- KCI 15 0.858 + 0.002 0.859 f 0.002 0.860
25 0.842 + 0.001 0.842 -+ 0.001 ---
tration range (up to saturation) and any desired temperature. 35 0.827 + 0.001 0.827 f 0.001 0.828
The values of the Debye-Htickel parameters(Eq. 2) at different BaCls 15 ___ 0.910 f 0.001 0.908
temperaturesare given by Ananthaswamy and Atkinson (1984) 25 0.903 f 0.001 ___ ___
and those of b and 01are taken as temperature independent. 35 --_ 0.894 k 0.002 0.893
Recently, Holmes and Mesmer (1983) applied the ion-in- KNOs 15 --_ 0.948 f 0.001 0.943
25 0.926 IT 0.001 -__ ___
teraction model of Pitzer (1973) to aqueous solutions of the
35 ___ 0.889 f 0.002 0.898
alkali metal chlorides to 250°C. Their resulting set of equations
provide a thermodynamic description within the accuracy of KzSQI 15 ___ 0.979 -+ 0.000 0.977
25 0.975 -c 0.000 -__ ___
the reported experimental data. The parameters B(O),B(l) and 35 ___ 0.972 f 0.000 0.971
CMx are described by arbitrary functions of temperature, f(T) OThe fluctuations in a, values arise from the standard deviation of the solubility
of the form data.
b Predicted using Pitzer’s model (Eq. 2) with parameters for 25°C or with the rigorous
temperature dependence for 15’C and 35°C [Eq. (4). (5). 0%. and (7)l.
f(T) = pl + PZ ($ - f) + p3 ln f + ~4 (T - ‘G) (7) c Predicted using Pitzefs model with approximation for the temperature dependence
IEq. WI.
+ ps (T2 - Gr2) + p6 ln;T - 260) d Measured with the Novasina Thermoconstanter (average of triplicate determina-
tions).
where p,,. . . ., p6 are parameters.
with the exception of the term envolving p6 this equation (i.e., 15°C or 35°C as is the case in the present paper) can be
is identical with the form used by Silvester and Pitzer (1977). approximated by
The values of the parameters pt, . . . ., p6 for aqueous so-
lutions of potassium chloride are listed in Table 4; they allow
a very accurate calculation of the a, of KC1 solutions for the fcr, = 4x-c) + -&CAT (8)
whole concentration range (up to saturation) at any tempera-
ture . where f(r) is the value of the function at temperature T. The
Silvester and Pitzer (1978) gave the change with temperature values of the corresponding derivatives for the salts of interest
of the activity and osmotic coefficients. They reported results to the present study are given in Table 5. The most important
for 84 electrolytes of 1- 1, 2- 1, 3- 1 and 2-2 valence types. The conclusion is that for almost all salts considered here the tem-
parameters allow the convenient calculation of properties at perature derivatives are relatively small. Thus, for NaBr, NaCl,
other temperatures not too different from 25°C. The values of KCl, BaC12, and K2S04 where B(O)and B(l) at 25°C had a
B(O),B(l) and CMX at temperaturesnot too different from 25°C magnitude of about 0.05 or a few tenths (Table 2), the deriv-

Volume 51, No. 4, 1986--JOURNAL OF FOOD SCIENCE-1039

a, OF SALT SOLNS AT VARIOUS TEMPS . . .
Table 8-Prediction of the a,,, of selected saturated salt solutions at 15°C experimentally found in the literature for several salts slurries;
and 35°C neglecting the effect of temperature on the thermodynamic
parameters of the ion-interaction model8
however, as reported by Scott and Barnard (1983), the “phys-
b icochemical reasons for the decrease in a,,, were not fully
a, a,
Salt (35°C) understood.” The a, of saturated NaCl remained almost con-
NaBr 0.601 ‘- 0.002 0.522 f 0.008
stant between 15°C and 35°C (ha, = 0.002) and this was
NaCl 0.752 k 0.001 0.750 f 0.001 expected on the basis of the solubility changes and very small
N-L&S04 0.808 f 0.002 0.798 e 0.002 effect of the temperature on the characteristic parameters in
KCI 0.856 AZ 0.002 0.830 e 0.001 the model of Pitzer (1973). It was noteworthy that the predicted
BaCI, 0.910 f 0.001 0.895 k 0.002
KNOB 0.944 k 0.001 0.907 f 0.002
a, values at 15°C and 35°C were in very good agreement with
K&L, 0.979 e 0.000 0.972 55 0.000 the values measured with the Novasina Thermoconstanter pre-
a Pitzer (1973) viously calibrated with unsaturated NaCl and LiCl solutions;
b Fluctuations in the predicted aw values arise from the standard deviation of the differences between theoretical and experimental values were
solubility data. only about 0.002 a,. The exception was KN03 for which a
larger difference was observed, these differences being 0.003
Table *The a, of selected saturated salt solutions between 15°C and
at 15°C and 0.009 at 35°C. This may be attributed to the
37-c relatively large values of the temperatures derivatives (Table
Temp
5) which make more uncertain the accurate theoretical predic-
(“a NaCl NL,LSOz, KCI BaCI, WOa tion of the effect of temperature. KN03 also exhibited a large
15 0.753 0.808 0.859 0.910 0.979 increase of solubility with increasing temperature; extrapola-
17 0.753 0.806 0.856 0.909 0.978 tion of Pitzer’s (1973) model to the saturation concentration at
19 0.752 0.805 0.852 0.907 0.977 35°C may be somewhat less reliable.
21 0.752 0.804 0.849 0.906 0.977 Labuza et al. (1985) used the vapor pressure manometer to
23 0.751 0.803 0.846 0.905 0.976
25 0.751 0.803 0.842 0.903 0.975 measure the a, of saturated NaCl and KC1 solutions and re-
27 0.750 0.802 0.840 0.902 0.975 ported values of 0.765, 0.748 and 0.727 for NaCl and 0.846,
29 0.750 0.801 0.836 0.900 0.974 0.841 and 0.786 for KC1 at 25”C, 30°C and 45”C, respectively.
31 0.750 0.800 0.833 0.899 0.973 The values for NaCl were not in good agreement with the
33 0.749 0.799 0.830 0.898 0.973
35 0.749 0.798 0.827 0.895 0.972 theoretical predictions or with present experimental determi-
37 0.748 0.797 0.823 0.894 0.971 nations; the values for KC1 at 25°C and 30°C are in better
agreement but the value for 45°C seems to be somewhat low.
A simplified theoretical approach was also tested to predict
atives were the order of 1O-4 to 10-3. Hence, a temperature the effect of a temperature change (from 25°C to 15°C or 35°C)
change from 25°C of lO-20°C will cause little change in these on the a, of salt slurries studied in this work. For this purpose
parameters and thus in a,. A different picture was observed it was assumedthat the effect of temperature may be accounted
for KN03 where the temperature derivatives were relatively for by only considering the effect on solubility and neglecting
more important. the intrinsic variation of a, with temperature (i.e., the tem-
An important prerequisite for the prediction of the a, of perature derivatives of ion-interaction parameters are negligi-
saturated salts at different temperatures is precise knowledge ble). Thus, Eq. (2) with parameters p(O), p(l), CMx
of the salt concentration at saturation. For this purpose a com- corresponding to 25°C was used to predict the a, values at
pilation of solubility data for NaBr, NaCl, (NH4)2S04, KCl, saturation concentration at 15°C or 35°C. The Debye-Hiickel
BaC12, KN03, and K2S04 at various temperatures, was made. parameters were those at the corresponding temperature 15°C
Table 6 gives the results of the literature review on the solu- or 35°C (Ananthaswamy and Atkinson, 1984). The results are
bility of the salts examined at 15”C, 25°C and 35°C (sources shown in Table 8. Predicted values for NaBr, NaCl, KCl,
of data are not reported here due to the very large number of BaCl* and K2S04 agree quite well with the values obtained
references considered). with the more rigorous approach (Table 7). This method al-
lowed the prediction of a, at 15°C and 35°C of (NH&S04;
The solubility of salts studied showed a significant increase there is a good agreement with measured values (Table 7). As
with increasing temperature, with KNO, showing the largest found before, KN03 at a higher temperature (35°C) exhibited
increase. The exception was NaCl where solubility increased the greatest deviation. These results indicated that for the salts
only slightly with increasing temperature (1% change between studied (with the exception of KN03) the increase in solubility
15°C and 35°C). Considering the small temperature depend- was the dominant factor which regulated the change in a, when
ence of the parameters for the ion-interaction model (Table 5), temperature shifted from 25°C to 15°C or 35°C.
it could be anticipated that the a, of saturated NaCl solutions Having demonstrated the reliability of the theoretical pre-
will be only slightly affected by temperature (in the range dictions, the different equations and solubility data were used
studied). to calculate the a, of saturated NaCl, (NH&S04, KCl, BaCl*
It is now possible to combine the data of solubilities at and K2S04 every 2°C between 15°C and 37°C (a slight ex-
various temperatures with the concentration dependence of the trapolation above 35°C was done because of the microbio-
osmotic coefficient for each salt (calculated from the ion-in- logical interest in the temperature 37°C). The results are shown
teraction model) at different temperatures and to predict the in Table 9; it is believed that these data are accurate enough
a, of saturated salt solutions at 15°C and 35°C. For NaCl and for almost all needs in food microbiology research or quality
KC1 the ion-interaction model of Pitzer with the rigorous tem- control. NaBr and KN03 were not included here because the
perature dependenceof parametersp(O), /3(l) and Cm [as given predictions were somewhat less accurate than those for the
by Eq. (4), (5), (6) and (7)] were used. For NaBr, BaCl,, others. This theoretical approach may help to find ways to
KN03 and K2S04 the temperature derivatives (Eq. 8) were reach a general agreement on the a, of standards used to cal-
used. No information is available on the temperature depend- librate a, measuring devices.
ence of ion-interaction parameters for (NH4)$S04. The results
of the prediction are shown in Table 7 which compares pre-
dieted values at 15°C and 35°C with those experimentally de- REFERENCES
termined using the Novasina Thermoconst$nter previously
calibrated as described. Ananthaswamy, J. and Atkinson, G. 1984. Thermodynamics of concen-
For all salts studied, the a, (either predicted or measured) trated electrolyte mixtures. 4. Pitzer-Debye-Hiickel limiting slopes for
water from 0 to 100°C and from 1 atm to 1 kbar. J. Chem. Eng. Data 29
decreased with increasing temperature; this effect has been (1): 81.

104O-JOURNAL Of FOOD SCIENCE-Volume 51, No. 4, 1986

Bromley, L. 1973. Thermodynamic properties of strong electrolytes in
aqueous solutions. AIChE J. 19: 313.
Chirife, J., Favetto, G., Ferro Font@ C., and Resnik, S.L. 1983. The water
activity of standard saturated salt solutions in the range of intermediate
moisture foods. Lebensmittel.-Wissenschaft. Technol. 16: 36.
Chirife, J. and Resnik, S.L. 1984. Unsaturated solutions of sodium chloride
as reference sources of water activity at various temperatures. J. Food
Sci. 49: 1486.
EEC. 1976. Meat Products, Regulation Nr. 77/99, Dec. 21. Economic Eu-
ropean Community, Bruxelles, Belgium.
FAOWHO. 1979. Recommended International Code of Hygienic Practice
for Ground Nuts (Peanuts), CACRCP 22-1979. Food and Agriculture
Fa ;m.zation of the United Nations, World Health Orgamzation, Rome,
‘i
FDA. 1973. Federal Register 38(16), Part III, 2398. Food & Drug Admin-
istration, Washington, DC.
Gal, S. 1981. Recent development in techniques for obtaining complete
sorption isotherms. In “Water Activity: Inlluences on Food Quality.” (Ed.)
L.B. Rockland and G.F. Stewart, p. 89. Academic Press, New York.
Greenspan, L. 1977. Humidit fixed points of binary saturated aqueous
solutions. J. Res. Nat. Bur. Qtand. A, Phys. Chem. 81A: 89.
Hamed, H.S. and Owen, B.B. 1958. “Physical Chemistry of Electrolyte
Solutions.” 3rd ed. Reinhold Publishing Co., New York.
Holmes, H.F. and Mesmer, R.E. 1983. Thermodynamic roperties of aqueous
solutions of the alkali metal chlorides to 250°C. J. P Rys. Chem. 87: 1242.
International Critical Tables of Numerical Data: Physics, Chemistry and
Technology. 1926. MC Graw Hill Co., New York.
Kitic, D., Resnik, S.L., and Chirife, J. 1986. Solutions of lithium chloride
as reference sources of water activity above 0.6 and between 5” and 45°C.
Lebensmittel. Wiss. Technol. (in press).
Labusa, T.P., Ascott, K., Tatini, S.R., Lee, R.Y., Flink, J.M., and MC Call,
W. 1976. Water activity determination: A collaborative study of different
methods. J. Food Sci. 41: 910.
Labuza, T.P., Kaanane, A., and Chen, J.Y. 1985. Effect of tern erature on
the moisture sorption isotherms and water activity shift oP two dehy-
drated foods. J. Food Sci. 50: 385.
Pitzer, K.S. 1973. Thermodynamics of Electrolytes. I. Theoretical basis and
general equations. J. Phys. Chem. 77: 268.
Pitzer, K.S. and Mayorga, G. 1973. Thermodynamics of electrolytes. II.
PERSIMMONS IN MODIFIED ATMOSPHERE. . .From page 1016
sealed lyethylene-lined boxes of apples, oranges and lettuce with hy-
drated p”ime mserta. Proc. Am. Sot. Hort. Sci. 82: 83.
Hardenburg, R.E. 1971. Effect of in-package environment on keeping qual-
ity of fruits and vegetables. HortScience 6: 198.
Honda, Y. and Ishiguro, 0. 1968. Studies on the storage of fruits and
vegetables. II. On the keeping of some vegetables packaged with plastic
bags. J. Jap. Sot. Hart. Sci. 37: 77.
Kawada, K. 1982. Use of polymeric films to extend postharvest life and
improve marketability of fruits and vegetables-Unipack: individually
wrapped storage of tomatoes, oriental rsimmons and grapefruit. In
“Controlled Atmosphere for Storage an 8”Transport of Perishable Agri-
cultural Commodities”, (Ed.1 D.G. Richardson and M. Meheriuk, p. 87.
Timber Press. Beaverton. OR.
Marcellin, P. 1961. La conservation de pommes ou de poires en atmosphere
controlee au maven d’amballaees de matiere .dastiaue.* Int. Inst. B&rig.
Bul. Annex 1: 361.
Matauo, T. and Itoo, S. 1977. On mechanisms of removing astringency in
persimmon fruits by carbon dioxide treatment. I. Some properties of the
two processes in the deastringency. Plant Cell Physiol. 18: 17.
Matauo, T. and Itoo, S. 1982. A model experiment for destringency of per-
Volume
Activity and osmotic coefficients for strong electrolytes with one or both
ions univalent. J. Phys. Chem. 77: 2300.
Resnik, S.L., Favettq, G., Chirife, J., and Ferro Font&, C. 1984. A world
survey of water activity of selected saturated salt solutions used as stan-
dards at 25°C. J. Food Sci. 49: 510.
Robinson. R.A. and Stokes. R.M. 1965. “Electrolvte Solutions.” 2nd. ed.
Reinhold Publishin Co., ‘New York.
Rockland, L.B. 1960. Baturated salt solutions for static control of humidity
between 5” and 40°C. Anal. Chem. 32: 1375.
Scatchard, G. 1968. The excess free energy and related properties of so-
lutions containin electrolytes. J. Am. Chem. Sot. 90: 3124.
Scatchard, G., Rus fi , R.M., and Johnson, J.S. 1970. Osmotic and activity
coefficients for binary mixtures of sodium chloride, sodium sulphate,
magnesium sulphate, and magnesium chloride in water at 25”C.fII. J.
Phvs. Chem. 74: 3786.
Scott, V.N. and Bernard, D.J. 1983. Influence of temperature on the mea-
surement of water activity of foods and salt systems. J. Food Sci. 48: 552.
Silvester, L.F. and Pitser, K.S. 1978. Thermodynamics of electrolytes. X.
Entalphy and the effect of temperature on the activity coefficients. J.
Solution Chem. 7: 324.
Stokes, R.H. and Robinson, R.A. 1949. Standard solutions for humidity
control at 25°C. Ind. Eng. Chem. 41: 2013.
Stoloff, L. 1978. Calibration of water activity measuring instruments and
devices: collaborative study. J. Assoc. Off. Anal. Chem. 61: 1166.
Tran, T.T. and Lenzi, F. 1974. Methods of estimating the water activity of
su ersaturated a ueous solutions. Can. J. Chem. Eng. 52: 798.
Trol Per. J.A. and C1. n&an. J.H.B. 1978. “Water Activitv and Food.” Ac-
ademic Press, New York.’
Wexler, A. and Hasegawa, S. 1954. Relative humidity-temperature range
of 0°C to 50°C. J. Res. Nat. Bureau Standards 53: 19.
Young, J.F. 1967. Humidity control in the laboratory using salt solutions-
a review. J. Appl. Chem. 17: 24.
MS received g/13/85; revised 3/15/86; accepted 3/15/86.
The authors acknowledgethe financial supportfrom Secretariade Cienciay T&z-
nica de la RcptiblicaArgentina (ProgramaNational de Tecnologtade Alimentos).One
of us (D.C.P.J.)acknowledgesthe financial support from CONICET (Argentine)and
CNPa (Brazil).
simmon fruit with high carbon dioxide treatment; in vitro gelation of
kaki-tannin by reacting with acetaldehyde. A ‘c. Biol. Chem. 46: 683.
Per+ E. and Ng, T.G. 1984. The role of anaeroT ic respiration in germi-
natmg muskmelon seeds. II. Effect of anoxia treatment and alcohol de-
hydrogenase activity. J. Expt. Bot. 35: 366.
Pesis, E. and Ng, T.G. 1984. The role of anaerobic respiration in germi-
nating muskmelon seeds. II. Effect of anoxia treatment and alcohol de-
hydmgenase activity. J. Ex th. Bot. 35: 366.
Saguy I. and Mannheim C.!I 1975 The effect of selected lastic films
and rhemical dius on th;! shelf-life of Marmande tomatoes. B Food Tech-
nol. 10: 547. _
Sugiura, A., Yonemori, K., Harada, H., and Tomana, T. 1979. Changes of
ethanol and acetaldehyde contents in Japanese persimmon fruits and
~~;~relatlon to natural deaatringency. Stud. Inst. Ho&, Kyoto Univ.
Tolle, W.E. 1971. Vegetables affecting film permeability requirements for
~4$df’ed atmosphere storage of apples. Tech. Bull. U.S. Dep. Agnc. No.
MS received 112186;revised 3126186;accepted 3128186.
51, No. 4, 1986-JOURNAL OF FOOD SCIENCE-1041