THERMODYNAMIC STUDIESOF
Nov., 1958
A few exchange runs were made by Moore and Shim with cuprous oxide consisting of individual crystalline grains cut from the polycrystalline strips. The diffusion coefficients obtained were in fairly good agreement with those from the polycrystalline oxide. These results tend to support
THE
SYSTEMACETONE-CHLOROFORM
1441
the theoretical argument that the observed oxygen diffusion is not governed by grain-boundary processes. Further experimental tests of this question are planned, with single crystals of cuprous oxide prepared by oxidation of single crystals of copper.
THERMOD~YNAMIC STUDIES OF THE SYSTEM ACETONE AND CHLOROFORM1 BY CHARLES R. MUELLERAND EDWARD R. KEARNS~ Department of Chemistry, Purdue University, Lafayette, Indiana Received July 7, 1968
An equilibrium still of the Scatchard type has been used for a determination of isothermal vapor-li uid equilibrium relations in the acetone-chloroform system a t 25, 35 and 50’. Purity of the components was guarantee3 b careful handling and evaluation of several physical constants, including the freezing point and cryoscopic constant. Varues of excess free energy of mixing are believed accurate to about 0.5 cal./mole. A tern erature-dependent relation between free energy and mole fraction is suggested as having certain advantages over the $an Laar Equation. Heats of mixing calculated from a derivative function agree with the calorimetric values at various temperatures to about 20%.
+
The acetone chloroform system has been of interest since Zawidski2 first noted the great deviation from Raoult’s law. Of the 6287 azeotropic systems listed by Lecat,3 only 459 exhibit negative deviations. In most of the systems of this type, strong interaction between molecules of the components is indicated. Result& of ultraviolet spectroscopy14measurement of supersonic waves6 and proton magnetic resonance6have been cited as evidence for strong hydrogen bonds in the acetonechloroform system. Any simple considerations of this system are precluded by evidence of association of acetone molecule^.^ It has been argueds that chloroform is also associated in solution. Barker and Smithg and MunsterlO have been among those attempting to predict the thermodynamic mixing functions by consideration of ordering or orientation forces. Most of the theories have been compared quantitatively with values given by Kirejew,” based on the isothermal vapor-liquid measurements of Zawidski2 and the calorimetric measurements of Hirobe. l 2 In this paper isothermal vapor-liquid equilibrium measurements over a temperature range are reported. There is reason to believe the data obtained are more reliable than the recent values given by Rock and S~hr0der.l~Compounds (1) Based on part of a thesis by Edward R. Kearns, submitted to the faculty of Purdue University, in partial fulfillment of the requirements for the degree of Master of Science, August 2, 1957. (2) J. von Zawidski, 2. physik. Chem., 35, 129 (1900). (3) M. Leoat, “Tables Azeotropiques,” Vol. 1, 2nd Ed., Chez I’Auteur, Uccle-Bruxelles, 1949. (4) R. Krernann, Monatsh., 61, 351 (1932). (5) R. Parshad. Indian J . Phys., 16, 307 (1942). (6) C. M. Huggins, G. C. Pimentel and J. N. Shoolery, J . Chem. Phys., 23, 1244 (1955). (7) W. Hers and M. Lewy, Chem. Zentr., 77, I , 1728 (1906). (8) A. Nikuradse and R. Ulbrich, 2. physik. Chsm. (Frankfurt), 2 9 (1954). (9) J. A. Barker and F. Smith, J . Chem. Phye., 2 2 , 375 (1954). (10) A. Mlinster, Trans. Faraday Soc., 46, 165 (1950). (11) V. Kirejew, Acta Phyeicochim. U R S S , 13,531 (1940). (12) H. Hirobe, J . Faculty Sci., I m p . Uniu. Tokyo, Sect. I , 1, Part 4, 155 (1926). (13) H. Rock and W. SchrGder, Z . physik. Chem. (Frankfurt), 11, 41 (1957).
of known high purity are utilized in connection with a circulation still of proven reliability. Temperatures and pressures have been measured to 0.001” and to a few hundredths of a millimeter. This precision is extremely important in calculation of enthalpies and entropies. Purification of Components Acetone.-It has been demonstrated that the properties of acetone vary considerably with the absorption of moisture.14J6 In the present work unusual precautions were taken to remove both non-aqueous and aqueous impurities from the acetone used. Accidental exposure of the dry acetone to moist laboratory air demonstrated the remarkable effect of moisture on refractive index. Since the equilibrium still had a condenser of proven inability to condense water vapor, refluxing for about an hour in the equilibrium still produced a sample of constant refractive index, corresponding to the value observed for acetone of the purity herein designated (I). Technical grade acetone from Commercial Solvents Corp. was fractionally distilled at partial take-off in a helicespacked, 4.5 ft. column, rejecting the first and last 10%. The distillate was subjected to further purification by the Livingston’s modification of the method of Shipsey and Werner,17with fractional distillation utilized in the last step. Fractional distillation of the product from hosphorus pentoxide’s in an air atmosphere gave samples, ¬ed (I), which were subjected to cooling curve analysis19*20 and ultraviolet ray absorption analysis. Acetone (11) was obtained by tractional distillation of acetone (I) from phosphorus pentoxide in an atmosphere of nitrogen dried by passage through calcium chloride. A bulb having a single stopcock was used for collection of the distillate in an atmosphere of dry nitrogen. All samples were removed from this bulb only in an atmosphere of high purity dry nitrogen. All joints and stopcocks were lubricated with Apiezon N. The physical constants observed for acetone (I) were: nPD 1.35881, n Z 61.35598; ~ dad 0.77909 g./ml: By cooling curve measurements, the sample was determined to freeze 1
(14) P. Thirion and E. C. Craven, J . A p p l . Chem. (London), 2 , 210 (1952). (15) K. T. Thomas and R. A. McAllister, A . I . Ch. E. J . , 3, 161 (1957). (16) R. Livingston, J . Am. Chem. Soc., 69, 1220 (1947). (17) K. Shipsey and E. Werner, J . Chem. Soc. (London), 103, 1256 (1913). (18) J. Timmermans, Bull. 8oc. chim. Belg., 24, 259 (1910). (19) IP,XB f TVxM
(7)
The second term usually is ignored because it is ordinarily small. Since estimation of slopes accurately requires data over a considerable temperature range, a modification of equation 7 has been often used
From graphical plots of the data, the slopes used in eq. 7 were calculated. The results may be seen in Fig. 2. In computing the average per cent. deviations, the values for X B = 0.1 and 0.9 have been ignored, since they are beyond the range of even the observed values. Over the remaining range, calculations from equations 7 and 8 agree reasonably well with the calorimetric data of Hirobe12 may be in error because of the vapor space above the liquid, which could cause a lowering in the calorimetric (49) K. T. Y u and J. Coull, Chew. Eng. Prom., Symposium Bar. No. 2, 48, 38 (1952). (50) H. C. Carlson and A. P. Colburn, Ind. Eng. Chem., 84, 581 (1942). (51) J. H.Ashley and G. M. Brown, Chem. Eng. Progr., Symposium Ser. No. 10,60,129 (1954).
..
THERMODYNAMIC STUDIESOF
Nov., 1958
SYSTEM ACETONE-CHLOROFORM
THE
1445
0 values of - H x E . Cheesman and Whitaker68have noted that these corrections may be estimated only - 60 to about =klO%. This problem has not been removed by the calorimetric measurements of Brown W -J -120 and F ~ c k , but ~ ? the essential agreement with the 0 values of Hirobe12confirms the results are probably f 180 not greatly in error. Mathematical Expression of Results.-Examina4 -240 a tion of the constants in equation 6 suggested rela- 9 tions of the type used by Barker and Smith9 u & -300 I
a. = b,
+ c.T + d, log, T
I-
(9)
360
This relation seemed to hold rather well, with evaluation as lo4 a. = 2341.7 = -6516.7 lo4 a2 = 5408.3
lo4 al
+ 1.6306T + 491.50 log, T - 4.25997' + 1358.0 log, T + 3.46752' - 1126.6 log, T
XAXB
XA
(bo
01
02
03
04
05
06
*
2101
\\
+ (bz + czT + dz 1% T P B ' (10) X A [ ( ~+Odo) + 2~07'+ do loge T + {(bl + dl) + 2clT + dl loge T]XB + ( ( b z + Q ) + di log. T)XB
+ COT+ (di + clT)XB + ++cZT)XBa] - '[bo + COT + do log. T + (bl + di loge T)XB + (bz + C2T + dz loge T)XB*l2 XA[do
TSp" XAXB=
(d2
ClT
08
09
I
entropy of mixing of acetone and chloroform.
+ COT + do log, T ) + (bl + ClT +
(11)
07
MOLE FRACTION CHLOROFORM.
Fig. 3.-Excess
Substitution of equation 9 into the related equations, (6) and (7), yielded general relations of the form
-GXE - - -
0
VAPOR
I6OO
0I
0.2
q b , 0.3 0.4
LIQUID
A I
0.:
MOLE FRACTION OF CHLOROFORI
Fig. 4.-Vapor
pressure-composition diagram of acetone and chloroform at 25'.
experimental data for only one mixing function. These equations seem to a t least approach a useful (12) form, but because of the errors involved in evaluaAccumulating errors arising from uncertainties tion of the constants they may not be applied to acintroduced in the process of smoothing values of GxE curate quantitative predictions at present. Imat three temperatures and in calculations of the proved weighting of experimental quantities and constants in equation 9 have, naturally, affected smoothing of data would enable one to use better the values calculated from equation 11. It is not quantitative tests. Possibly the evaluation of the surprising, then, that the average deviation from constants could be better approached a t the present calorimetric values a t 25" over the range X B = time by using equation 11 for estimation of the con0.2 to 0.8 is about 20%. Comparisons with the stants from calorimetric data. From the experimental viewpoint there remain scanty calorimetric data a t other temperatures (of questionable validity) gives about the same yo de- large discrepancies between the calorimetric and viation. A plot of HxE against XBusing values cal- vapor pressure values for the enthalpy. We believe culated from equation l l has a minimum displaced that the present values represent the useful limit for t o a higher mole fraction and to lower values of HxE, the equiIibrium still. If we had accurate calorimetbut otherwise the shape of the curve seems to be ric measurement, without vapor space, we could about the same as that obtained from Hirobe's13 easily establish the limits of the Scatchard apparatus. Discrepancies of this order of magnitude are calorimetric values. This is illustrated in Fig. 2. Equations 10, 11 and 12 represent an attempt to frequent and it seeys important to know what facpredict all thermodynamic mixing functions from tors are a t fault. We should like to express our appreciation for the (63) G. H. Cheesman and A. M. B. Whitaker, Proc. Roy. Soc. support given by the Research Corporation to thia L ~ ~ ~ OA T aLi a ) , 406 (1952). work. (53) I. Brown and W. Fock, Australian J . Chem., 8, 361 (1955).
~