ethanol- thamaolmAeetone -LIQUID EQUILIB H. H. AMER’ AR”) R . R . PAXTOS? Stanford Einiversity, Stanford, Calif.
IIIATTHEW VAN WINKLE University of Texas, Austin, Tex.
HERE are several methods of predicting vapor-liquid equilibria of a ternary system using the vapor-liquid equilibria of the three related binary systems (2, 6, 19), but there are relatively few consistent sets of data suitable for testing these methods. The purpose of this investigation Tyas to furnish a consistent set of data for the ternary system methanol-ethanolacetone.
pressure in this system was adjusted to a value determined from the barometric pressure, so that the absolute pressure in the still was 760.0 mm. of mercury. Temperature variations of as little as 0.05’ C. in the equilibrium chamber were readily detectable by means of an iron-constantan thermocouple that was installed therein. Equilibrium wa,s indicated by a constant rate of feed into the vaporizer, a constant volume of unevaporated liquid (one small drop) in the low portion of the vaporizer tube, and constant temperature in the equilibrium chamber. The still was run for a t least 1 hour a t these conditions, then the stopcock between the chambers was closed and the heaters were shut off. Samples were quickly withdrawn into 50-ml. Erlenmeyer flasks, which were immediately stoppered and cooled. The samples were analyzed hy determining their refractive index and bubble point; the relation between these properties and composition was already known ( 1 ) .
PURITY OF COMPONENTS
BINARY SYSTEMS
Table I compares the specific gravities, refractive indices, and bubble points of the components used in this investigation with those reported by earlier workers. The methods used to ascertain these physical constants have been reported ( 1 ) .
Table I1 presents the equilibrium temperatures and vaporliquid compositions for three binary systems: acetone-methanol,
istillatton process designers
will use the consistent set of data for the system methanol-ethanol-
acetone
PROCEDURE
The vapor-liquid equilibrium was established using a modified Colburn ( I S ) still, provided with an enlarged vapor condensate chamber to give a larger “vapor” sample. The vapor condensate chamber and the still had a volume of 30 and 55 cc., respectively. Figure 1 is a drawing of the still stripped of its heating elements. The still Kith its external wapping of heating wires was enclosed in a nindowed boy t o minimize the effects of air currents upon its performance, The heaters on the vaporizer tube and the equilibrium chamber were separately controlled. The pressure-control system on the still utilized dried air. The
-
Table I.
Properties of Pure Compounds Used Literature Values
Compound
Property
AIethanol (m. w. 32.04)
Spec. gr., dz5 Ref. index, nl,o Boilingpoint, O C.
0.7866 ( I O ) 0.7865 1.3290 (12) 1.32904 64.65 (18) 64.6
Ethanol (ni. w. 46.07)
Spec.gr., d:j , Ref. index, n Z o Boilingpoint,DoC.
Acetone (in.
Spec. gr., dz5 Ref. index, %go Boiling point, O C.
0,78505 1.36155 78.27 0.78508 1,35880 56.20
1 2
w.
58.08)
Exptl.
C3NDENSbTi C u 0hl BE R
(4) 0.7850 (4) 1.36152 (4) 7 8 . 3 (4) 0.7840 (4) 1.36878 (4) 5 6 . 1
Prwent address, 12 Talaat St., Cairo, Egypt. Present address, General Electric Co., 1 Plastics Rd., Pittsfield, Mass.
Figure 1. Modified Colburn equilibi.iuiti still
142
INDUSTRIAL AND ENGINEERING CHEMISTRY
January 1956
143
Table 11. Constant Pressure Vapor-Liquid Equilibrium Data for Acetone-Methanol, Acetone-Ethanol, and Methanol-Ethanol Solutions (Pressure = 760 mm. H g ) Acetone-Methanol Mole Fraction Acetone T~~~ Liquid Vapor (2.”
Acetone-Ethanol Mole Fraction Acetone T~~~ Liquid Vapor O (2.”
Methanol-Ethanol Mole Fraction Methanol T ~ ~ ~ . , Liquid Vapor C.
64.6 63.5 62.2 60.7 59.4 58.1 56.9 56.2 55.9 55.8 55.8 55.8 55.8 56.1
22 2c
I
1.8
+ z
w 0 k w
1.6
1.4
0 0
Table 111. Vapor Pressures of Acetone, Methanol, and Ethanol Vapor Pressures, M m . H g Acetone Methanol Ethanol
Temperature, O
c.
g
1.2
z
c 0 a 1.0
0
02
04
0.f
0.8
IO
MOLE FRACTION ACETONE I N LIQUID
Figure 2.
Activity coefficients and bubble points
at 760 mm. for acetone-methanol solutions
0 0
02
04
06
0.8
02
04
06
OF
IO
10
MOLE FRACTION ACETONE IN LIQUID
Figure 3. Activity coefficients and bubble points at 760 mm. for acetone-ethanol solutions
MOLE FRACTION METHANOL IN L I Q U I D Figure 4. Activity coefficients and bubble points at 760 mm. for methanol-ethanol solutions
144
INDUSTRIAL AND ENGINEERING CHEMISTRY
Figure 5.
vapor pressure data ale used, a wrong activity coefficient ai11 be computed. The liteiature m s searched to obtain the "best" values for the vapor pressure of acetone, methanol, and ethanol betneen 50" and 80' C To aid in interpolation between the data points, logarithms of the vapor pressure values vr ere plottedagainst temperature and a smooth curve was drawn. Zmaczyntki's ( 2 1 ) data v ere chosen to locate the acetone cuive For cthanol, both Kretschiner arid Wiebe's (14) and Llerriman's (16) data !%ere used to f i the ~ vapor piessure curve. There was a plethoia of data foi methanol (3, 11, I?'), the set obtained by Young ( 2 0 ) seeming to give the best curve for temperature6 beti? e m 50" and 80' C. Table I11 report? selected values of vapor pressures read from these curves. All activity coefficients herein reported used these vapor pressure values. Acetone-Methanol. -4ctivity Loefficients and bubble point temperatures for the acetone-methanol binary are plotted in Figure 2. I n addition to the authors' data, Figure 2 includes the isobaric equilibrium data of Griswold and Buford ( 7 ) and of Othmtxr, Fiiedland, and Schiebel ( 1 6 ) The authois'
ETHANOL
ME TH ANG L
Activity coefficients of acetone in acetone-methanolethanol ternary boiling at 760 mm. of mercury
8 8 - 99 99-111 111-127 -
0 0
88 90
9
IOU
9
I02
+
0
9
I04
0
9
0
92 94 96
9
I06 I O 8
-c)
8
98
9
I I O
9
0
-0
-0
I12
I14 116 118 120 125
128-165
b b
165-22
CP
I 7
b
150 1.35 140
0-
b
145
0-
I 8 19 2 0
6
150 155
0 21 6- 2 2
b
Vol. 48, No. 1
*
acetone-ethanol, and methanol-ethanol. All of these data were obtained with an equilibrium still pressure of 760 mm. of mercury. Binary vapor-liquid equilibrium data are often presented as z - y plots. Such a plot, \\bile useful for giving a picture of the ease of separation of the components, is not a very severe test for consistency of the data. A more exacting test of the data may be obtained by plotting the activity coefficients against the mole fraction of the more volatile component in the liquid. Activity Coefficients. The activity coefficient used in this paper is defined as follows:
7% =
7rYZ/PZrCl
where
T~ = activity coefficient for the ath coin-
ponent total pressure, 760 0 mm. of mercury in this case ut = mole fraction of z in the equilibrium vapor P , = vapor pressure of pure i at the equiiibrium temperature xi = mole fraction of i in the equilibiiuni liquid T
=
METHANOL
8 8 - 99 ___
0
o
88 EiO 92 94 96
Q
98
0
The activity coefficient thus calculated depends upon E',, the vapor pressure of the pure component. If the equilibrium bubble point temperature is wrong, or if erroneous
CTHANOL
Figure 6. Activity coefficients of acetone in acetone-methunol-ethanol ternar) boiling at 760 mm. of mercury
0 0
39-111 ___
9 9
loci I 0 2
i-~ i i - 1 2 1
128-165
a 1 1 2 d l 1 4
(5 b
9
I C 4
0 1 1 6
b
9
I C 6
e l 1 8
9 9
I
b L
oa
-0
120
I I O
0
I25
t
1 6 5 - 2 2
120
CP
I35
e-
140 145 150 155
a-
17 I8
e
19 20 21
0
22
3
145
146
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 48, No. 1
the other components. The Hughes and bIaloney data have so much scatter that It is not possible to tell whether this small amount of water affects the activity coefficients of methanol and ethanol. The water would, of course, affect the bubble point of the system; therefore these points are not plotted in Figure 4. Bergstrom’s data amplified and reported by Hausbrand ( 8 ) could not be used because no temperatures were reported. TERh ARY D4TA
Table 11- reports vapor-liquid equilibi id data for the acetone-methanol-ethanol ternary. The behavior of acetone in this ternary IS illustrated by Figure 5, which shows lines of constant acetone activity coefficient and lines of constant bubble point. Similar plots for methanol and ethanol are shown in Figures 6 and 7. LITERATURE CITED
(1) Amer, H. H., Paxton, R. R , and Van Winkle, Matthew, Anal. Chem. 25, 1204 (1953). (2) Benedict, M., Johnson, C A , Solomon, E., METHANOL ETHANOL and Rubin, L. C., Trans. Am Inst Chem. Figure 7. Activity coefficients of ethanol in acetone-methanolEngrs. 41, 371 (1945) ethanol ternary boiling at 760 mm. of mercury (3) Dorochevsky, A. G., and Poliansky R J Russ. Chem. Soc. 42, 109 (1910). .88-.99 99-Ill 111-127 1.28-1.65 1.65-2.2 - (4) Dreisbach. R. R., and Martin, It. A , IXD. EXG.CHEM.41, 2875 (1949). b 1.50 cr 1.7 9 I.0U -3 1.12 0 .88 ( 5 ) Duffey, private communication reported in e .go 7 1.02 - 0 1 . 1 4 i 1.35 * 1.8 “Chemical Engineer’s Handbook,” 3rd ed., b 1.40 I, 1.9 0 .92 9 1.04 0 1.16 p. 573, iMcGraw-Hill, S e w York, 1950. b 1.45 e- 2.c 9 1.06 -3 1.18 0 .94 (6) Gerster, J. A., Mertes, T. S..and Colburn, 0 .96 9 1.08 -0 1.20 b 1.50 0 2.1 A. P., I X D . ENG.CHEM. 39, 787. 1520 (1947). 1.55 c- 2.2 9 1.10 9 1.25 0 .98 (7) Griswold, John, and Buford, C . E., Ibid., 41, 2347 (1949). (8) Hausbrand, E., “Principles and Practice of Industrial Distillation,” 4th ed., tr. by E. H. Tripp, p. 232, data indicate that a t 760-mm. pressure the acetone-methanol azeoWiley, New York, 1926. trope exists a t about 75 mole acetone and has a boiling point (9) Hughes, H. E., and Maloney, J. D., Chem. Eng. Progr. 48, 192 between 55.7” and 85.8’ C. Otherwise all three sets of data are (1962). substantially in agreement, except for the activity coefficient of (10) “International Critical Tables,” vol. 3 , p. 27, 3IcGram-Hill, acetone a t low acetone concentrations. Kew York, 1928. The departure of these correlations from the shape predicted (11) Ibid., pp. 216, 237. (12) Ibid., vol. 7, p. 80. from the Duhem equation is believed to be due to nonideal (13) Jones, C. A , Schoenborn, E. >I., and Colburn, A . P., ISD. ENG. behavior of the mixed vapors. CHEM.35, 666 (1943). Acetone-Ethanol. Figure 3 s h o w bubble points and activity (14) Kretschmer, C. B., and Wiebe, R., J . Am. Cham. Soc. 71, 1793, coefficients for the acetone-ethanol system. The author’s data 3176 (1949). agree with Duffey’s ( 5 ) . (15) Merriman, R. W., J . C h m . SOC.(London) 103, 628 (1913). Methanol-Ethanol. Perhaps the most interesting of these (16) Othmer, D. F., Friedland, D., and Schiehel, E. G., unpublished dat,a reported by J. C. Chu in “Distillation Equilibrium three binaries is the methanol-ethanol system, which is generally Data,” p. 19, Reinhold, New York, 1950. assumed to be ideal. Figure 4 shows that the components \?-hen (17) Timmermans, J., “Physico-Chemical Constants of Pure Organic present in low concentration do not behave ideally. The activity Compounds,” p. 303, Elsevier, New York, 1950. coefficients of methanol at low methanol concentrations are less (18) Timmermans, J., and Hennault-Roland. J . chim. phys. 27, than 1.00, and the coefficients for ethanol a t low ethanol con401 (1930). (19) Wohl, Kurt, Trans. Am. Inst. Chem. Engrs. 42, 215 (1946); centrations are greater than 1. Chem. Eng. Progr. 49, 218 (1953). The only modern data y i t h which these could be compared (20) Young, S., Sei. Proc. Roy. Dublin Soc., N.S. XII, 374 (1909-10). were those reported by Hughes and Maloney (9). The Hughes (21) Zrnaczgnslii, h.,J . chim. phys. 27, 503 (1930). and Maloney system contained up to 2 weight % of water, but A C C E P T EJune D 3, 1955. R E C E I V Efor D review August 12, 19;54. this does not prevent calculation of the activity coefficients of ~~