SALT EFFECT ON VAPOR-LIQUID EQUILIBRIA FOR ACETONE-WATER SYSTEM'

E1ZO SADA, TOSHIO OHNO** AND SHIGEHARU KITO Department of Chemical Engineering, Nagoya University, Nagoya, Japan Vapor-liquid equilibrium data of acetone-water system saturated with sodium chloride, potassium chloride, sodium nitrate, and calcium chloride dihydrate are determined under atmospheric pressure. These vapor-liquid equilibrium data are correlated by a method which gives approximately the behavior of salt in the liquid. The standard deviation of correlated results is 2.32%. To carry out this correlation, the vapor pressures of aqueous solutions saturated with salts are also determined.

Investigations concerned with the effect of salt the portion between Cx and Hj is brought to the addition on the vapor-liquid equilibria of binary- vapor pressure of sample solution. Whenthe system systems are of theoretical and industrial importance, attains equilibrium (about one hour is needed), and not a little knowledge of vapor-liquid equilibrium the heights of Hl5 H2, and H3 aremeasured, and the data in this field is already available2}. However, vapor pressure is calculated from the following most investigations are limited to presentation of equation.

P '---(Hi-H,W+(Hg-H,W experimental data with no attempt to develop a cor- LO -% r relation. Among these, only a few correlations />*" (1) are proposed, for example, by Johnson and Furter5), As the density of sample solution is usually an order and Hashitani and Hirata3). In this paper, the vapor-liquid equilibrium data of magnitude lower than for mercury, it does not decrease the accuracy of measurementto use the for the acetone-water system saturated with each density of pure water at measuring temperature in of four kinds of salt, i.e. sodium chloride, potassium chloride, sodium nitrate, and calcium chloride place of psl. To check the reliability of the experimental ap- dihydrate, are presented and these experimental results are correlated by a new method which ap- paratus, the vapor pressures of pure water weremea- proximates the mechanism of the effect of salt addi- sured at various temperatures of 50° to 70°C and are tion on vapor-liquid equilibrium. Also, the vapor compared with the values in the literature4) in Fig. 2. Fromthis figure, it is found that the vapor pressure pressures of aqueous solutions saturated with salt data obtained here agree well with those in the are observed in order to carry out this correlation. literature4), and the average deviation is ± 0.6%. Experimental Apparatus and Procedure Vapor-liquid equilibria In this work, all vapor-liquid equilibrium data Vaporpressure were obtained at atmospheric pressure of ca. 750 to The schematic diagram of the experimental ap- 770 mmHg, and therefore the boiling points should paratus is shown in Fig. 1. The measuring procedure be corrected to those at the pressure under considera- is as follows. The entire system, is held at a desired tion. However, in the case of salt addition this temperature within ^ 0.1°C by circulating tempera- operation is difficult, and the following discussion ture-controlled water through a water jacket. With is based upon the reference state of 760mmHg, C2 opened and C2 closed, mercury is filled to Cx and allowing the entrance of a little error in the activity then about 5 cc of the degassed sample saturated with coefficient. salt at the measuring temperature is introduced An improved Othmer type still which was modified into the portion above mercury level by manipulating for salt effect studies was used. This still was devised the mercury bulb. During this operation, care to avoid the partial condensation of vapor and to should be paid to avoid the entrance of air. Then be suitable for the dissolution of salt. Details were Ci is closed. After a while, C2 is opened to drain off described in a previous paper6). Some equilibrium a proper amount of mercury from the system so that liquid phase samples were separated quantitatively * Received on September 8, 1971 into salt and acetone-water fractions by a technique ** Plant Division, Nippon Sharyo Kaisha, Ltd., Nagoya, of evaporation to dryness. The acetone-water frac- Japan. tion was analyzed by refractive index determination VOL, 5 NO. 3 1972 (1) 215 S A 2 . 6 ¥ K C W M ¥ ¥ C| s s. ¥ -n o s a lt 2. 4 ^ ^ ^ ^ ^ H ^ ^ a _ H i | i ^ ^ B H I ^ ¥ I ^ ^ ^ ^ M I I IIlIlIIIHHHH ^H I W Q T

I ^ ^ ^ ^ ^ ^ B ^ S a B S B I ^ ^ ^ ^ ^ ^ I C J> N a C I ¥ H 2 o ¥ V ¥ 2 .2 ¥ N J N N ¥ s.N V N a N O 3 s.o n 2 .0 N ^ s. H 3 2 . 8 2 . 9 3 . 0 I/ T x /O * -- pure water (4) Fig. 2 Vapor pressure of water saturated with salt w * c ; of calcium chloride dihydrate, on the other hand, c 2 the data are obtained fromthe literature4\ since this solution is too viscous to deal with in the apparatus. These results werewell correlated by the Antoine A : pressure regulation J : waterjacket equation with the standard deviation of 0.017, and tube M : mercury bulb the values ofA, B, and C in Eq.(2) are listed in Cl5C2,C3 : stop cocks O : mercury outlet Table 2. Hx: upper level of liquid P : circulating pump

sample S : sample liquid inlet logP2s=A-t+C H2: upper level of mercury W: constant tempera- (2) H3 : lower level of mercury ture bath Thedegreeof vaporpressuredepressionfor acetone Fig. 1 Schematic diagram of experimental apparatus for causedby salt addition wasnegligible, as predicted measuring vapor pressure fromthe very low solubilities of these salts in it. Vapor-liquid equilibrium data for acetone-water for acetone-water ratio and thus the composition of system saturated with sodium chloride, potassium the liquid phase sample was established. Equilib- chloride, sodium nitrate, and calcium chloride di- rium vapor condensate samples were also analyzed hydrate are shownin Figs. 3 and 4 as well as Table by refracvtie index determination. Experimentally, 3. Since saturation was assumed to exist experi- mentally when a slight excess of solid salt persisted saturation was assumed to exist whena slight excess of solid salt persisted over a period of time. over a period of time, as described previously, it First of all, some vapor-liquid equilibrium data can be concludedfromthese figures that the addition of calcium chloride dihydrate and sodium chloride for this system without salt were obtained in the range of 0.20solubility of this salt in acetone-watersystem. The additions of sodium chloride and potassium Experimental Results and Discussion chloridecausesimilar salt effects, whilethe addition Vapor pressures of aqueous solutions saturated of sodiumnitrate results in the smallest effect of the with sodium chloride, potassium chloride, and four kinds of salts studied in this work. sodium nitrate were measured and the results are The system of acetone-water-saturated salt is shown in Table 1 and Fig. 2. For aqueous solution considered to be a highly non-ideal system, and 216 (2) JOURNALOF CHEMICALENGINEERINGOF JAPAN COS- s*~ /

C / /

- / '£ 0.6 -/' / / / / / J 0.4t' / I 1 / / | ! / ONaCI I O CaCI2'2H20 S02[ / QKCI O NaNOs CD /^ O IX , i j 0 0.2 0.4 0.6 0.8 0 0.2 04 0,6 0.8 1.0 acetone mole fraction in licfuid, acetone mole fraction in liquid, X| 1.0 (salt-free basis) Xi (salt-free basis) nosalt nosalt Fig. 3 Vapor-liquid equilibria of acetone-water saturat- Fig. 4 Vapor-liquidequilibria of acetone-water saturated ed with salt under atmospheric pressure with salt under atmospheric pressure

Table 1 Vapor pressures of water saturated with salts Table 3 Vapor-liquid equilibrium data for acetone-water systems saturated with salts under atmospheric pressure Salt Temperature[°C] Vapor[mmpressureHg] t C°C] xx yhobs y1>cal A [%] NaCl 54.3 89. 1 NaCl 64.1 138.8 56.2 0.960 0.967 0.973 0.64 69.2 173.9 57.5 0.756 0.894 0.91 1 1.86 77.2 245.5 57.9 0.682 0.887 0.900 1.47 81.9 293.1 58.2 0.630 0.894 0.893 0.15 KC1 52.2 89.8 58.3 0.535 0.894 0.885 -0.98 58.3 0.389 0.888 0.874 - 1.56 60.3 128.1 65.7 163.9 58.6 0.253 0.869 0.854 - 1.73 58.4 0.101 0.869 0.797 -9.07 72.6 222.2 83.8 348.1 CaCl2-2H2O 84.2 357.8 55.7 0.941 0.961 0.989 5.10 56.7 0.831 0.927 0.982 5.95 NaNO3 56.1 91.7 56.8 0.754 0.941 0.977 3.86 64.8 131.2 56.7 0.635 0.947 0.973 2.77 76.0 218.9 56.8 0.533 0.937 0.971 3.95 84.6 308.6 56.6 0.345 0.927 0.966 4.2 1 56.6 0.299 0.934 0.964 3.25 Table 2 Values of A, B andCin theAntoineequations 57.0 0.258 0.941 0.963 2.29 68.4 0.031 0.854 0.837 -2.03 for aqueous solutions saturated with salts KC1 Salt A B C 57.1 0.829 0.920 0.927 0.74 NaCl 6.75 1081.6 170.7 57.6 0.783 0.907 0.904 -0.39 KC1 6.ll 781.2 135.0 58.1 0.695 0.894 0.885 - 1.03 NaNO3 7.83 1694.3 232.8 58.8 0.552 0.875 0.879 0.43 CaCl2-2H2O* 6.69 1312.9 188.9 58.5 0.330 0.868 0.855 - 1.56 58.9 0.150 0.868 0.825 -5.17 Calculated by using the data in the literature4) NaNO3 59.0 0.502 0.874 0.883 0.98 60.9 0.260 0.855 0.862 0.76 various properties of this system maybe very com- 60. 1 0.235 0.853 0.856 0.39 plicated. Therefore, for the purpose of simplifi- 64.8 0.093 0.795 0.792 -0.36 cation in this investigation, the salt addition in liquid (subscript 1 refers to acetone) phase is assumed to have a noticeable effect only on the depression of vapor pressure for each liquid component due to dissolution of the salts and to Therefore, by assuming vapor phase to be ideal, have no effect on the activity coefficients. the following equations are considered for a binary

217 VOL. 5 NO.972 3 (3) system saturated with salt. acetone mole fraction, where the solubility of cal- Vi= TlPlSXl (3) cium chloride dihydrate becomes low. However, TV this effect does not discredit the experimental results y - ^2^2S^2 (A) because the liquid composition is determined by the evaporating method. 2/1+2/2-1 (5) Conclusion where T1 and T2 are the activity coefficients of acetone and water in a binary system without salt, respec- Vapor-liquid equilibrium data for acetone-water tively. In addition, P^ is regarded as P^ since system saturated with each of four kinds of salts were the vapor pressure depression for acetone caused obtained under atmospheric pressure. These ex- by salt addition is negligibly small, and Eq.(3) be- perimental results were correlated by a new ap- comes proximate method which requires the activity coef- ficient data of each volatile component for a system 1Z (6) without salt and vapor pressure data for an aqueous Here, Eq.(6) seems to show that acetone mole frac- solution saturated with salt. Predicted results of tion in vapor for the system with salt does not differ vapor composition agreed with the experimental from that for the system without salt. In fact, data within a standard deviation, of 2.32%. however, the change of acetone mole fraction caused by salt addition is considered in Eq.(6) because the Nomenclature boiling point of the system is elevated by salt addi- tion and this results in a rise of the vapor pressure of Pts = vapor pressure of pure componenti saturated with salt [mm Hg] pure acetone, P^. Pi0 = vapor pressure of pure componenti [mmHg] By use of Eqs.(4), (5), and (6), the prediction of t = temperature [°C] vapor-liquid equilibria for acetone-water system T = absolute temperature [°K] %i = mole fraction of component i in liquid phase saturated with salt is carried out in the following (salt-free basis) [-] manner. For a liquid composition, a vapor pres- yi = mole fraction of component i in vapor phase [-] sure is calculated from the Antoine equation by activity coefficient of component i for binary system without salt [-] position is calculated from Eqs.(4) and (6) by sub- A = relativeerror [%] stitution of these vapor pressure data. The opera- pM = density of mercury [gm/cm3] tion mentioned above is repeated by the trial-and- ^s = density of aqueous solution saturated with salt[gm/cm3] error methoduntil vapor phase composition satisfies Eq.(5). In these calculations, total pressure is taken tz å ==total pressure [mmHg] to be one atmosphere. 1 = acetone The predicted results of vapor-phase compositions 2 å = water obs = observed value are listed in Table 3 with the experimental results. cal = calculated value From this table, it is found that the predictions are Superscripts) good. Their average deviations are 2.18%, 1.55%, 0 = at0°C 0.62%, and 3.71%, respectively, for the additions t =at*°C of sodium chloride, potassium chloride, sodium nitrate, and calcium chloride dihydrate. In these Literature Cited numerical deviations, that for the addition of cal- 1) Brunjes, A. S. and M.J. P. Bogart: Ind. Eng. Chem., 35, cium chloride dihydrate is to some extent higher than 255 (1943) the others. This may be explained as follows. 2) Ciparis, J. N.: "Data of Salt Effect in Vapor-Liquid Equilibrium", Ed. Linthnanian Agr. Acad., Kaunas, When calcium chloride dihydrate is added to liquid, USSR (1966) hydrated water becomes free from salt, and as a 3) Hashitani, M. and M. Hirata: J. Chem. Eng. Japan, result water fraction in the liquid is increased. This 2, 149 (1969) effect on liquid composition is not negligible in this 4) "International Critical Tables", Vol. Ill, pp. 368, case because solubility of this salt is high and there- McGraw-Hill, New York, N.Y. (1928) 5) Johnson, A. I. and W.F. Furter: Can. J. Chem. Eng., fore the predicted results may be somewhat wrong. 38, 78 (1960) This explanation seems to be justified by the fact 6) Sada, E., T. Ohno and S. Kito: Kagaku Kogaku, 35, that the deviation becomes small in the region of high 372 (1971)

218 (4) JOURNAL OF CHEMICAL ENGINEERING OF JAPAN