3176
Vol. 71
CARLB. KRETSCHMER AND RICHARD WIEBE
cose was dissolved in 100 ml. of warm, absolute methanol. p-L-Glucopyranose Pentaacetate.-a-L-Glucosea (0.20 Crystallization was effected on slow solvent removal in a g.) was heated on a steam-bath with freshly fused sodium desiccator over sulfuric acid. Pure material was obtained acetate (0.12 g.) and acetic anhydride (2.0 ml.) and heaton recrystallization (or elution) from methanol or ethanol ing was maintained for two hours after solution, which and drying was effected ;rider reduced pressure over phos- required about thirty minutes. The crystalline material phorus pentoxide at 56 ; m. p. 112-113.5", [ ~ ] ' O D 0 " (0.38 g., SS%, m. p. 130-131") obtained on pouring the (no mutarotation). It crystallized as fine needles. reaction mixture into 20 ml. of ice and water, was reAnal. Calcd. for C6H1206: C, 40.00; H, 6.66. Found: crystallized from 95% ethanol; m. p. 131-132", [cY.]*~D -3.9" (c 2.9, chloroform) in agreement (opposite sign) C, 39.90; HI 6.72. with those cited by Hudson and Dales for P-D-glucose a-DpL-Glucopyranose Pentaacetate .--a-D,a-L-G1ucose (0.200 8.) and 2 ml. of anhydrous pyridine were cooled pentaacetate. Anal. Calcd. for ClsH22011: C, 49.22; H, 5.68. to 0" and 1.4 ml. of acetic anhydride was added slowly under mechanical stirring while maintaining the same Found: C, 49.30; H, 5.32. temperature. The suspension was stirred a t 0' for twenty p-D,p-L-GlucOpyranose Pentaacetate.-Equal amounts hours after which time the solution was poured slowly (0.1000 9.) of p-D- and @,-glucose pentaacetate were reinto 10 ml. of ice and water. The crystalline material crystallized from 95% ethanol; m. p. 125-126", needles. was removed by filtration, washed with water and reAnal. Calcd. for C16H22011: C, 49.22; H, 5.68. crystallized from 95% ethanol; yield 0.258 g. (6l'%), Found: C, 49.60; H , 5.43. m. p. 140-140.5". The substance formed complex Summary rhombic crystals. Anal. Calcd. for C16H22011: C, 49.22; H, 5.68. 1. A crystalline form of racemic glucose has Found: C, 49.21; H, 5.75. been prepared. This product showed the same X-ray powder diffraction 2. The crystalline pyranoid pentaacetates of diagram (Table I ) as that obtained by recrystallizing equal amounts (0.2000 9.) of the OI-D and the below-de- a-L-, p-L-, a-D,a-~-and p-D,P-L-glucose are described a-L-glucose pentaacetate from 95% ethanol; m. p. scribed. 141-141.5", mixed melting point unchanged; m. p. 1133. X-Ray powder diffraction diagrams of the 115" on admixture with theo below-described P - D , ~ - L - above compounds have been obtained. pentaacetate of m. p. 125-126 . 4. Mild acetylation of the racemic glucose a-L-Glucopyranose Pentaacetate.--a-~-Glucose~(0.40 9.) was acetylated as described above for the CY-D,~-Lproduces a-D,a-L-glucopyranose pentaacetate. isomer and the crude crystalline product (0.77 g., 89%, 5. The D,L-glucose (m. p. 112-113.5') dem. p. 108-110°0) was recrystallized from ethanol-water; scribed is a racemic mixture of the CY-Dand a-L m. p. 112-113 , [cY]*~D -101' ( c 4.4, chloroform) in agreement (opposite sign) with those cited by Hudson and forms of glucose, probably pyranoid. 6. The described pyranoid pentaacetates of Dales for a-D-glucose pentaacetate. Anal. Calcd. for ClsHz2Oll: C, 49.22; H , 5.68. ~-D,cY-Land @-D,P-L-glucoseare true racemic Found: C, 49.29; H , 5.30. compounds. (8) C. S. Hudson and J. K. Dale, THIS JOURNAL, 37, 1264 (1916).
[CONTRIBUTION FROM
THE
COLUMBUS, OHIO
RECEIVED APRIL 29, 1949
NORTHERN REGIONAL RESEARCH LABORATORY']
Liquid-Vapor Equilibrium of Ethanol-Methylcyclohexane Solutions BY CARLB. KRETSCHMER AND RICHARD WIEBE gel. Facilities for a determination of the freezing point were not available but the physical proper~ are in satisfactory ties, dZ540.76496, n Z 51.42059, agreement with the values given by Forziati5 for a sample containing 0.10 mole per cent. impurity, v ~ z . : d"4 0.76501, ZZ5D1.42056. Since the densities of the two liquids differ by only 0.02, solutions were analyzed by means of Experimental their refractive indices for the green mercury line The apparatus, method and purification of the (5461 A.). A Bausch and Lomb precision refracethanol have been described fully in previous ar- tometer was used. In order to establish the relat i c l e ~ . * , ~A commercial grade of methylcyclo- tion between composition and index of refraction, hexane was fractionally distilled in an efficient solutions were made up by weight as previously column and percolated through n column of silica described.2 Because of the theoretical and practical interest in the change of volume on mixing, (1) One of t h e laboratories of t h e Buredu of Agricultural and I n the densities of these solutions were measured as dustrial Chemistry, Agricultural Research Administration, U. s. well as their refractive indices. Table I gives the Department of Agriculture. Article not copyrighted. (2) C. B. Kretschmer, J. Xawakowska and R. Wiebe, THISJOIJRresultingvalues. The volume changes are all posiNAL, 70, 1785 (1948). tive and are remarkably close to the values for (3) C. B. Kretschmer and R. Wiebe, i6id., 71, 1793 (1949).
In two recent publications results on the liquidvapor equilibria of ethanol-isooctane2 and ethanol-toluene3 solutions were described. The work has now been extended to solutions of ethanol in methylcyclohexane. The only previous measurements on this system are those of IsiiI4who measured the total vapor pressure a t 0 to 30'.
(4) N. Isii, J. SOC. Chcm. I n d . J a b a n , 38. Sup. Binding, 639 (1935).
( 5 ) A. F. Forziati, A. R. Glasgow, C. B. Willingham and F. D Kossini, J . Res. Nor. Bur. Standards. 36, 129 (1946).
EQUILIBRIUM OF ETHANOL-METHYLCYCLOHEXANE SOLUTIONS
Sept., 1949
ethanol-isooctane solutions a t the same temperature. Compositions of the liquid and vapor samples were calculated from their indices of refraction with the aid of the indices listed in Table I. The accuracy of such analyses is believed to be about 0.1 unit in the percentage by weight.
the vapor pressure curves for this system are similar to those for the other systems studied, they are not reproduced here.
Discussion The vapor compositions were fitted by the empirical equation previously used3
-
TABLE I y(1 - %)/[%(I - y)] = a = ( A Bx)/ DENSITIES AND REFRACTIVEINDICESOF ETHANOL[(x C)(1 - 2 c METHYLCYCLOHEXANE SOLUTIONSAT25°
+
1W Av,
Ethanol Wt. fract. Mole fract.
0,0000 .0327 ,0882 ,1586 ,3366 .5376 .5639 .5662 ,6610 .8018 .8518 1.0000
0.0000 .0671 .I709 ,2866 .5196 .7125 .7337 .7356 .8060 .8961 .9245 1.0000
n%61
d'64
ml./g.
1.42240 1.41987 1.41605 1.41133 1.40002 1.38755 1.38600 1.38587 1.38021 1.37195 1.36904 1.36073
0.76496 .76457 .76484 .76562 .76836 .77226 .77284 ,77288 .77508 .77877 ,78024 .78505
0 176 316 418 547 562 553 554 504 364 290 0
The results of the liquid-vapor equilibrium measurements a t 35 and 55' are given in Table 11. The vapor pressures for pure ethanol were taken from our previous work3; those for methylcyclohexane were calculated from the equation published by Willingham and co-workers.6 Since
0.0000 .0526 .1446 .2878 .4052 .5403 ,6914 .8450 .9676 i.ooo0
p,
Aya
Log yi
Log
+
400
i
i
1
I
300
9 E
200
B
100
o
YZ
35" 73.62 . .. . . . . . 0.0000 135.40 0.0027 1.0634 .0120 146.97 - ,0008 0.7016 .0509 151.27 .0024 ,4353 .1204 152.36 .0018 .2986 .I913 152.93 .0025 .1834 .2947 .0927 .4414 152.22 .0042 145.73 - .0042 ,0299 .6541 .0024 120.04 .0048 .9106 1.0000 103.14 ...,.. .oonn . . . .
0.0000 .4645 .5118 .5362 ,5471 .5575 .5817 ,6423 .8369
(1)
+
-3
mm.
+ CX)l
where x and y are the mole fractions of ethanol in liquid and vapor, respectively. The following values of the constants were found to represent the data for ethanol-methylcyclohexane solutions: a t 35O, A = 1.131, B = 1.003, C = 0.020; at'5S0, A = 1.339, B = 1.179, C = 0.028. That the fit is satisfactory is shown by the deviations listed in Table 11. Fig. 1 illustrates the method of fitting this equation. The quantity q = a ( x C) ( 1 - 2C Cx) is plotted against x for various values of C, and that value of C is chosen which makes the points lie closest to a straight line, bearing in mind that the points must be given weights proportional to y (1 - y ) / q if all the values of y are assumed equally accurate.
TABLE I1 .* LIQUID-VAPOR EQUILIBRIUM OF ETHANOL-METHYLCYCLO8 HEXANE SOLUTIONS AT 35 AND 55" Mole fract. ethanol Liquid. z Vapor, y
3177
..
-
- 100 0.2 0.4 0.6 0.8 1 I) Mole fraction of ethanol. Fig. 1.-Relative volatilities of ethanol-methylcyclohexane solutions: upper curve, 55O, C = 0.028; lower curve, 35", C = 0.020.
Equation (1) was substituted into the GibbsDuhem relation in the form given as equation (3) of our previous paper,3 and used to calculate the 55" vapor pressure a t mole fractions of 0.5 and 1, for 0.0000 0.0000 168.10 .. ... . . . . 0.0000 both temperatures. The liquid molal volumes and .0528 .4835 319.83 0.0012 1.0189 ,0072 second virial coefficients for ethanol were the same .1251 .5375 352.80 - .0026 0.7322 .0345 as used p r e v i ~ u s l y . For ~ methylcyclohexane, the ,2205 .5645 368.00 .0008 .5255 .0758 critical constants given by Kay,? were used with .3621 ,5846 376.34 .0028 .3347 .I517 the equation of state of Keyes, Smith and Gerry in .5071 .5988 379.83 ,0009 .2028 .2524 the form used by Scatchard and Raymond,8 to .0918 ,4160 give the following values of the second virial coef,6832 .6244 380.06 - .0024 .0492 ,5340 ficient: a t 35') -3361; a t 55', -2592 cc./mole. .7792 .6528 375.78 - .0027 .9347 .7879 337.52 ,0038 .0060 .SO46 Liquid molal volumes were calculated from the 1,0000 1.0000 279.89 . . . . . . .0000 . . . . density equation of M a ~ s a r t . ~The calculated Observed values minus those calculated from equation vapor pressures differ from observed values by
(1).
(6) C. B . Willingham, W. J. Taylor, J. M. Pignocco and F. D. Rossini. J . Res. N u l . Bur. Standards. 35, 219 (1945).
(7) W. B Kay, THISJOURNAL, 69, 1273 (1947). (8) G Scatchard and C. L. Raymond, ibid, 60, 1278 (1938). (9) L. Massart. Bull SOC. chim B d a . , 45, 76 (193fi).
3178
CARLB. KRETSCHMER AND RICHARD WIEBE
0.8% or less. This result may be taken to indicate a satisfactory degree of thermodynamic consistency in our data. Redlich and Kisterl" have developed the following equation to represent the behavior of solutions containing an associating component such as an alcohol. tog ( n / T a )
p
A'(K,x)
+ B'(1 - C'[6~(1 2%)+ - X) - 11
(2)
Here y1 and yo are the activity coefficients of alcohol and hydrocarbon, respectively. We have used primes to distinguish the notation from that of equation (1). A' is a function of x and of K , the equilibrium constant for the association reaction ROH 4- (R0H)n F1 (ROH)n+l. The form of the function A'(K,x) is given in reference (lob), equation (20). In equation (2) above, the terms in A' and B' are odd functions of (1 - 2x), while the term in C' is, of course, an even function. To test the applicability of equation (2) to the present data, the values of log y1 and log yz listed in Table I1 were calculated from the experimental data by use of the relation log 71 = log
(Yifi/Xifii)
+ (1/2.303RT)(i% vi)(#
- Pi)
(3)
Here Xi and y i are the mole fractions of component i in liquid and vapor; p i , pi and Vi are the vapor pressure, second virial coefficient, and liquid molal volume of component i, and p is the vapor pressure of the solution. Values of log (y1/ yz) were plotted against x,and the resulting curve was resolved into odd and even functions of (1 2%)by means of the relation f(x) = I/z[f(X)
- f ( l- 4 1 + '/2[f(x)
+ f(l- 4 1
(4)
It was found that the odd function could be represented quite accurately by the odd terms of equation ( 2 ) . However, the even function is represented poorly by the term Cf[6x(1 - x) - 11. This is shown in Fig. 2, where the plotted points have been taken from the experimental curve of 0.2
-0.08
1
'
I
I
I
I
Vol. 71
log (yl/yz) vs. x, and the curve is calculated for C' = -0.111. The deviations represent errors in y of up to 0.015 mole fraction. The following values of the parameters were found for ethanolmethylcyclohexane solutions: a t 35', K = 16, B' = 1.350, C' = -0.111; a t 55', K = 16, B' = 1.331, C' = -0.111. The ethanol-toluene systern3 a t 35' can be represented by K = 6, B' = 0.994, C' = -0.130, with deviations very similar to those shown in Fig. 2. The above values of K are anomalous in that one would reasonably expect K to decrease appreciably between 35 and 55'. Furthermore, the theory behind equation (2) requires that K have the same value for solutions of the same alcohol in different hydrocarbons a t the same temperature. However, it has been emphasized'O that K is very sensitive to experimental errors, and also, we might add, to deviations from the simplifying assumptions necessarily introduced in deriving equation ( 2 ) . TABLE 111 EXCESS THERMODYNAMIC FUNCTIONSOF ETHANOLMETHYLCYCLOHEXANE SOLUTIONS IN CALORIES/MOLE Mole fract. ethanol
GS, 35'
GB, 5 5 O
TSB, 35'
HZ
0.0528 .1251 .2205 .3621 .5071 .6832 ,7792 .9347
95 186 263 323 334 287 23 1 86
91 183 263 327 341 292 235 87
79 42 - 9 62 92 95 79 29
174 228 254 261 242 192 152 57
-
The excess thermodynamic functions in Table I11 and Fig. 3 were calculated from the experimental values of y and p by the usual m e t h ~ d . ~ The behavior is similar to that of the ethanoltoluene system, as well as various systems containing methanol which have been studied and discussed by Scatchard and co-workers. l 1 Little I
I I
1
I
I
I
I
I
I I
0.1 0.2 0.3 0.4 0.5 I, mole fraction of ethanol. Fig. 2.-Ethanol-rnethylcyclohexane solutions a t 55' : points, even component of log(yJy2); curve, 0.111[1 6x(1 I)].
I
I
I
I
-
0.2 0.4 0.6 0.8 1.0 I, mole fraction of ethanol. Fig. 3.-Excess thermodynamic functions of ethanolmethylcyclohexane solutions at 35'.
0. Redlich and A. T. Kister, J . Chcm. Phrs.. 10, 849 (1947): (b) Ind. Eng. Chem., 40, 345 (1948).
(11) G. Scatchard, S. E. Wood and J. M. Mochel, THIS JOURNAL, 68, 1957, 1960 (1946); S. E. Wood, {bid., 68, 1963 (1946); J . Chem. Phys., 10, 358 (1947).
-
1 (lo) (a)
0
Sept., 1949
3 179 SEPBRATION OF ZIRCONIUMAND HAFNIUM WITH THENOYLTRIFLUOROACETONE
can be added to their discussion except to say that a completely satisfying theory of the behavior of such solutions is not yet possible, and to emphasize again the fact that interactions between the unlike molecules are involved, in addition to the type of association of the alcohol postulated by Redlich and Kister.l0
Summary For the system ethanol-methylcyclohexane, densities and refractive indices have been measured a t 25'. Total and partial vapor pressures have been determined a t 35 and 55'. Thermodynamic con-
sistency of the data was found to be satisfactory. Vapor compositions are represented by the empirical equation previously developed by the authors, to within 0.005 mole fraction. They can be represented by an equation developed by Redlich and Kister, with somewhat less accuracy, and with rather anomalous values of the equilibrium constant involved. The excess free energy, entropy and enthalpy have been calculated from the data, and exhibit a similar behavior to that of other alcohol-hydrocarbon systems. PEORIA 5, ILLINOIS
RECEIVED MARCH24, 1949
[CONTRIBUTION FROM THE RADIATION LABORATORY, UNIWRSITY OF CALIFORNIA]
The Fractional Separation of Zirconium and Hafnium by Extraction with Thenoyltrifluoroacetone BY E. H. HUFFMAN AND L. J. BEAUFAIT Values for R and K have previously been determined3for zirconium using trace amounts of radioCO-CH,-CO-CF,, has recently been re- active ZrgS. The close similarity of zirconium and hafnium ported by Reid and Calvin2and a study of the ex- in chemical behavior makes it logical to assume traction of its zirconium chelate into benzene has that the above equations apply to both elements. been reported by Connick and M C V ~ ~The . ~ The determination of values for R and K for both work reported in this paper on the separation of zirconium and hafnium, using macro quantities zirconium and hafnium by means of this &dike- and concentrations for which the zirconium ions tone was undertaken a t the suggestion of M. Cal- has been shown to be a monomer in 2.0 M pervin and G. T. Seaborg. A study has been made of chloric acidI4has been carried out in this work in some differences in the extractions of zirconium the thenoyltrifluoroacetone concentration range of to 3 X for zirconium and to and hafnium by benzene solutions of thenoyltri- 4 X for hafnium. fluoroacetone (referred to as H T in equations and 7 X tables) from aqueous 2 M perchloric acid and the Experimental results applied to separations of these elements. Materials.-Thenoyltrifluoroacetone of 99.5% purity Zirconium is generally regarded as being in the was obtained from M. W. Davis and H. R. Lehman of this form of ZrO++ in acid solutions. The complexing Laboratory. Weighed amounts were dissolved in reagent of zirconium by the diketone is negligible in the thiophene-free benzene t o give solutions of the desired aqueous phase and the only important species in concentrations. Zirconium solutions were prepared in a manner similar ~ reaction then is the benzene phase is Z I - T ~ .The t o that of Connick and McVey.a Recrystallized ZrOCl?. expressed by the equation 8Hz0 was dissolved in 2.0 M perchloric acid to prepare a ZrO+*+ 4HT = ZrT, H20 + 2H+ stock solution of approximately 0.09 M zirconium. Tracer zirconium solutions were prepared from carrierand, assuming constant acid concentration, the free Zr96 in 5% oxalic acid which was received from the equilibrium constant by Isotopes Branch, United States Atomic Energy Commission. This solution was made about 10 M in nitric acid -log K = log [ZrT,]/[ZrOfe] - 4 log [HT] = and carrier zirconium added from the above stock solulog R 4 log [HT] tion. The solution was purified by the following treatA small amount of 0.1 N potassium permangawhere R is the extraction coefficient. If the zir- ment.* nate was added to precipitate manganese dioxide which conium is in the form of the simple Zr+4 ion in 2 M cames the radioactive columbium but not zirconium. perchloric acid, as recently reported,' the expres- This precipitation was carried out three times to remove sion for the equilibrium constant is not changed, both oxalic acid and columbium. The zirconium soluwas then diluted t o about 1 M nitric acid and exproviding that the zirconium ion is monomeric. tion tracted with 0.02 M thenoyltrifluoroacetone in benzene, (1) This paper is based on work done under the auspices of the and the benzene phase washed twice with 2 M perchloric The
p
synthesis
of
thenoyltrifluoroacetone,
+
-
Atomic Energy Commission. (2) J. C. Reid and M. Calvin, MDDC-1405(BC-75), August 13, 1947. (3) R. E. Connick and W. H. McVey, The Aqueous Chemistry of Zirconium, UCRL-101, March 1, 1948. (4) W.H.Reas, Thesis, University of California, 1948.
acid. The benzene solution was removed to a platinum dish and fumed with sulfuric acid with the addition of small amounts of 30% hydrogen peroxide to destroy organic matter. It was then diluted and zirconium hydroxide precipitated with ammonium hydroxide. The precipitate was dissolved in hydrochloric acid and the
