Vapor-Liquid Equilibrium Data for the System Acetone-Methanol Saturated with Salts"

VAPOR-LIQUID EQUILIBRIUM DATA FOR THE SYSTEM ACETONE-METHANOL SATURATED WITH SALTS"

SHUZO OHE, KIMIHIKO YOKOYAMA, AND SHOICHI NAKAMURA Ishikazvajima-Harima Heavy Industries Co., Ltd. Research Institute, Yokohama

Vapor-liquid equilibrium data at atmospheric pressure of the system : acetone-methanol-salt are studied. The Salts, Kl, NaCI, MgCI2, CaCI2/ LiCI, and CaBr2 are examined to observe the salt effect on the acetone-methanol system. Effective salts are CaQ2, LiC! and CaBr2, which are more soluble in methanol than Kl, NaCI and MgCh. CaCI2, LiCI and CaBr2 are observed to shift the azeotropic composition from 8O.I to 88.6, 9I-O and 94.O mole /# of acetone, respectively. The salt effect at each infinite dilute concentration of acetone and methanol increases with the increasing solubility of each salt in the rich concentration component.

In general, salts shift azeotropic compositions or 2) Experimental method eliminate azeotropes. For example, sodium chloride The salt, being completely non-volatile, appears only saturated in ethanol-water system shifts the azeotropic in the liquid, hence yielding a system consisting of composition from 87 to over 90 mole %ethanol45 and a two-componentvapor phase and a three-component calcium chloride saturated in the ethanol-water system liquid phase. The concentrations of acetone and eliminates the azeotrope25. This salt effect may be used for the separation of azeotropic mixtures. Systems which contain water as one component have been well studied5>9>10), but studies on non-aqueous systems are scarce from the point of salt effect. L. Belcku studied the effect of calcium chloride on the acetone-methanol system and reported on a constant concentration of 2.3 moles of salt/mole of solution. J. Proszt and G. Kollar6) also studied the effect of calcium-chloride and lithium-chloride and reported on a constant concentration of 1 mole of salt/liter of solution. In this study, isobaric vapor liquid equilibrium data at atmospheric pressure are reported for the six systems ; acetone-methanol-KI, NaCl, MgCl2 CaCl2, LiCl and CaBr2. L. Belckl:> suggested the possibility of the elimination of the azeotrope, if the data observed from 0 to 95 mole %acetone were extrapolated to 100 mole %acetone. But, the authors' data show the fact that azeotropic composition is only shifted from 80.1 to

Apparatus and Method 1) Equilibrium still The authors modified further the improved Othmer recirculation still presented by Johnson and Furter4) for salt effect studies. Fig. 1 shows the equilibrium still employed. Heating is done by a wall electric heater adjusted by a transformer. Two high quality standard thermo- meters are used for measurement of the boiling-liquid and vapor-phase temperatures. Received on December 2, 1967 Fig. Equilibrium still

VOL.2 NO.1 1969 1 methanol in the equilibrium liquid phase were calculated by mass balance, using the concentration of acetone and methanol in the original charge, and the analyzed concentrations of acetone and methanol in the equilibrium vapor condensate samples. The hold- up in the vapor phase chamber and condenser was neglected. Only the equilibrium vapor condensate samples were analyzed. Salt concentrations were calculated by the original charge of each component, which had been weighed. The saturation with salt was attained with slight excess of solid salt persisting in the still. The excess solid of salt in the liquid phase was observed from the windowof the still. Twelve runs of measurement were made using the binary system acetone-methanol at atmospheric pressure, in order to check the accuracy of data obtained from the still. The results without salt were compared with the literature data8) and found to be consistent with the data, within the maximumerror of 1%. Fig. 2 x-ycurvesofacetone (I)-methanol (2)- Thermodynamic consistency of the data was tested by CaCh system at I atm. Herington's3) method, and the data of the system without salt were shown to be consistent thermodynami- cally. The method, however, cannot be applied to the system containing salt, thus the consistency was not tested. To avoid change of salt concentration in the boiling chamber owing to the deposit of salt on the inner wall of the still, the revolution rate of the magnetic stirrer and the distillation rate were adjusted carefully by continuous observation from the window of the still. 3) Materials Acetone, methanol and salts used for experiments were guaranteed reagents. 4) Analysis Analysis of the vapor condensate samples was made Fig. 3 x-y curves of acetone (l)-methanol (2) by refractive indices. The refractive indices of acetone saturated with LiCI, CaBr2 and CaCh at I atm. and methanol at 20°C are, respectively, 1.3587 and 1.3920. The difference in value of these components is 0.0333, which is sufficient to determine the concentrations. Compositions were calculated from the tables of refractive index for the acetone-methanol system published in Timmermans8) data-book, with the tabu- lated data plotted on a large scale. The refractometer employed was an Abbe type.

Results All data are reported on a salt-free basis. Vapor liquid equilibrium data at atmospheric pressure are shown in Figs.2, 3, 4, and Tables1,2,3. The salt effect of six salts on the azeotropic composition are listed in Table 15). The data for acetone-methanol-calcium-chloride are plotted in Fig. 2 and listed as smoothed values in Table 2. Relative volatility of acetone to methanol increases with increasing CaCL concentrations at liquid phase concentrations from 0 to 90.0 mole % acetone, but decreases from 90.0 to 100 mole %acetone. (Fig. 4) Fig. 4 Relation of relative volatility of acetone CaCl2, LiCl and CaBr2 are observed to shift the azeo- (I)-methanol (2)-salts system at I atm. tropic composition from 80.1 to 88.6, 91.0 and 94.0

JOURNAL OF CHEMICAL ENGINEERING OF JAPAN Table I Salt effect of six saturated salts on acetone (l)-methanol (2) system and solubilities" (l atm) Salt effect* Solubility at 55°C** in Salts Solubility ratio - Acetone (A) Methanol (B) D KI 0.809 0.811 NaCl 0. 809 0.818 3.00X10" MgCl2 0. 802 0.839 55.6 CaCl2 0. 802 0. 840 56.2 0.0005 0.0027 LiCl 0. 806 0. 908 57.0 0.0083 0.0328 CaBr2 0. 800 0. 938 57.1 0.0098 0.0775 * Salt saturated ** Mole fraction

Table 2 The smoothed data of vopor-Iiquid equilibrium in acetone (l)-methanol (2j-CaC!2 (l atm) Concentrations of calcium chloride Xi 5 wt.: 10 wt. % 15 wt. % 20 wt. % Saturated

t[°C] y\t[°C] *_rc] t[°C] yit[°C] ^i 0.050 0.133 63.6 64.4 64.9 0.148 64.8 0.175 70.6 0.100 0.227 61.9 62.8 63.0 0.251 63.0 0.312 65.8 0.150 0.302 60.8 61.3 61.7 0.373 61.6 0.422 63.4 0.200 0.365 59.9 60.1 60.5 0.459 60.5 0.501 61.7 0.300 0.467 58.6 58.6 58.7 0.580 58.8 0.600 59.4 0.400 ' 0.550 57.6 57.6 57.6 - - 0.659 57.9 0.500 0.628 56.7 56.6 __ __ _ 0. 704 56.9 0.600 0.694 56.0 0.745 56.2 0.700 0.765 55.6 0.790 55.8 0.800 0.835 55.9 0.834 55.9 0.850 - 0.860 55.9 0.900 - 0.883 56.1 0.950 - 0.921 56.2

Table 3 Vapor-Liquid equilibrium data for Acetone (l)-Methanol (2) with saturated LiCI and CaBr2 (l atm)

X\yi yi tC°c] yi t [°C] t re] LiCl LiCl CaBr2 0.025 0.117 95.2 0.494 0.868 63.1 0.050 0.164 82.8 0.050 0.232 92.1 0.600 0.883 60.5 0.096 0.274 86.2 0.075 0.321 89.2 0.703 0.890 57.8 0.245 0.596 77.9 0.096 0.433 85.1 0.805 0.899 57.4 0.351 0.718 69.0 0.129 0.515 84.0 0.890 0.907 57.8 0.486 0.852 62.8 0.179 0.610 78.5 0.940 0.931 56.5 0.800 0.930 57.1 0.245 0.718 73.1 0.950 0.932 56.6 0.902 0.938 56.6 0.359 0.818 67.8 0.982 0.958 56.5 0.950 0.948 56.4 mole % acetone, respectively. (Figs. 2, 3) CaCL, LiCl concentrations, the effect is LiCl>CaCl2>CaBr2. At and CaBnare the most effective at about 20, 50, and acetone-rich concentrations, the effect is CaBr2>LiCl 60 mole %acetone, respectively, when the salts are >CaCl2. Saturated salt concentrations were not deter- saturated. (Fig. 4) mined directly. Approximate solubilities, however, are available from x-y curves of each concentration of Discussion of Results salt. In Fig. 2, the x-y curves of constant salt concentrations: 5, 10, 15wt. % intersect the x-y curve of Generally, the salt effect may be predicted by the salt saturated. The salt concentration at each inter- solubility of salt in each component. If the salt is sected point must be the same as that of the respective moresoluble in a less volatile component, then the constant salt concentration x-y curve. Therefore, relative volatility will be raised, because of the lowered solubilities are able to be determined, graphically at vapor pressure of the less volatile component. In this each intersected liquid concentration. Solubilities thus case, the salts are moresoluble in methanol, the less obtained are as follows : volatile component, thus increasing relative volatility. mole % acetone solubility (Table 1) (salt free 45.8basis) (weight15.0 %) On the other hand, the salt effect increases with 63.3 10.0 increasing solubility ratio of salt in acetone to methanol 78.3 5.0 at the concentration from 60 to 100 mole %acetone. The effect is CaBr2>LiCl>CaCl2. In addition to this, Acknowledgment the salt effect at each infinite dilute concentration of acetone and methanol increases with increasing solu- The authors acknowledge the continuing guidance of Professor bility of salt in the rich component. At methanol-rich Mitsuho Hirata (Tokyo Metropolitan University) and the sugges-

VOL.2 NO.1 1969 tions on experiments of Mr. Motoyoshi Hashitani (Tokyo Met- Literature cited ropolitan University, Dr. Course) 1) Belck, L.: Chem. Ingr. Techn., 23, 90 (1951) 2) Hashitani, M., M. Hirata, & Y. Hirose: Kagaku Kogaku Nomenclature {Chem. Eng. Japan), 32, 182 (1968) 3) Hala, E., et al. : "Vapour Liquid Equilibrium", Pergamon D = percentage deviation: Press (1958) = the value of 4) Johnson, A. L, & W. F. Furter: Can. J. Technol., 34, 413 log (71/V2) dxi (1957) where?i and j2 are activity coeff. 5) Landolt-Bornstein Zahlenwerte und Funktionen aus Physik. J a function of boiling points Chemie. Astronomic Geophysik und Technik 2. Teil b (1962) 6) Meranda, D. & W. F. Furter: Can. J. Technol., 44, 298 Tnim = the lowest measured temperature [°K] (1966) x\, xi = mole fractions of acetone & methanol in liquid phase, respectively (salt free basis) 7) Proszt, J. & G. Kollar: Roczen. Chem., 32, 611 (1958) yi, j/2 = mole fractions of acetone & methanol in vapor phase, 8) Timmermans, J. : "The Physico-Chemical Constants of Binary respectively (salt free basis) Systems in Concentrated Solutions" Interscience Pub. (1959) 9) Uchida, S., S. Ogawa, & M. Yamaguchi: Jap. Set. Rev. a = relative volatility of acetone to methanol Eng. Sci.y 1, 41 (1950) as = relative volatility salt free basis including salt 10) Yamamoto, Y. et al. : Kagaku Kikai {Chem. Eng. Japan), 0 = a function of boiling points 16, 166 (1952) ll) Yoshida, F. et al. : Kagaku Kogaku {Chem. Eng. Japan), = a function of log (7-1/7*2)

VAPOR-LIQUID EQUILIBRIA OF BINARY SYSTEMS CONTAINING ALCOHOLS : ETHANOL WITH NITROMETHANE AND DIETHYLAMINE*

KOICHIRO NAKANISHl*2 RITSUJI TOBA*3 AND HIDEKO SHIRAl*4 Department of Industrial Chemistry, Shinshu University, Wakasato, Nagano-shi

Vapor-liquid equilibrium data are reported for the binary systems ethanol-nitromethane (MeNO2) and ethanol-diethylamine (Et2NH) at 73OmmHg.As expected from a strong hydrogen-bond interaction between hydroxyl group and amino base, the ethanol-Et2NH system shows negative deviation from the idea! solution law and no azeotrope can be found. The ethano!-MeNO2 system shows a positive deviation. MeNO2forms an azeotrope at 76.4°C and 75.0 mole % of ethanol. Based on these results and other activity coefficient data available, the prediction of azeotrope3 formation in binary ethanol solutions by the previously presented correlation and azeotrope diagram is discussed.

In a series of studies on the vapor-liquid equilibria direction of the deviation from the ideal solution law, of binary solutions,^12'1" we have obtained the equi- is primarily dependent on the proton accepting ability librium data for ten binary methanol solutions. We of the molecule in alcohol solution. However, the have also proposed a correlation between the limiting value of the activity coefficient is also affected by the value of the activity coefficient of various liquids in difference in size and shape of component molecules. a large excess of methanol, log?"0, and the "hydrogen- It is thus necessary to obtain detailed information on bond shift" of the stretching vibration of the O-H such size effect for the purpose of establishing a gene- bond of methanol in these liquids, Jvs.12^ It was ralized activity coefficient correlation in alcoholic and demonstrated in these studies that the sign and magni- other associated solutions. Although a survey of the tude of the activity coefficient, i. e., the degree and literature indicates that, in contrast with binary methanol systems, both isobaric and isothermal vapor-liquid Received on April 30, 1968 equilibrium data are widely available for binary ethanol Presented in part at the 1st autumn meeting, the Society of Chemical Engineers, Japan, Toyonaka, Osaka, Nov. solutions, some important information as to the activity 1967 coefficient behavior in the ethanol systems is still To whomcorrespondence should be addressed to Depart- missing. ment of Industrial Chemistry, Kyoto University, Kyoto Present address : Hitachi Seisakusho Co. Ltd., Tokyo. In this paper, wewill report the barometric vapor- Present address: The Tokyo College of Pharmacy, liquid equilibrium data of ethanol (EtOH) with nitro- Shinjuku, Tokyo methane (MeNCh) and with diethylamine (Et2NH). 4 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN