tluoride was studied in the following manner. A small copper reaction vessel with a Hoke valve was provided, Iriaititxined a t constant temperalure, zml aliquot aii~ourith of the gas phase were ptimped ofl a t \ .irious interv,tls, the total pressure of the systeiii a n t i the 105, ili lreighr .)E tlic reaction Lwsel being dctcrniinetl . T h e teiii1)er'itii:c kept constant at 15.0'. The o\-cr-.i11 I-I:VIIIS ol)t:ii:!(~d , r v shown in Fig. 2. 301
,
+-MOLES
CrOiF2:KF.HF
Fig. ?.--System CrCJ..I;:
+ K;I:,HF a t 13.0'.
KHF1. Chroniyl fluoride was condensed over this pure bifluoride in a molar ratio 2.2CrOnFn:lKHFp. In the initial part of the curve the vapor pressure was appreciably above the vapor pressure of pure chromyl fluoride, indicating that :t more volatile ccnstituent was being pumped off simultaneously with the chrviiiyl fluoride. This volatile constituent proveti to he hydrogeil fluoride. This quantity correspoiitleti to '1 decrease of ahout 0.2 t o 0.25 mole of hydrogen fluoride per mole of the original KHF? as determined by titration. The pre.jsure o l the system decreased rapidly a t a rno1:rr ratio of about I : 1 to about 0.8: 1 and reached a value of litlow 4 inm. a t 0.75: 1. The composition of the yellow cleterinined by titrating the chromyl fluoride the total acidity by titration with alkali. €IF = 1 :0.97. Thus the final mixture i)ouiitl I;F.HF'.CrO?F2 and aliout 23 1 rnole (,;, o1 K F due t t i lob\ of €IF indicated above. !. --?'he following procedure was found iiple 7 ~ : ~sublimed s into a dry and fresh sealed by rneiins of hot tongs ( z200°), : ~ i i ( l \ri.ighetl. All the chroniyl fluoride was then sublimed i n t o oiie t.iitl of the tube iiy cooling in Dry Ice and the other ciid cut off with a razor I)lacfe atid the sample tube quickly i n ~ c r t v dinto :i glass stoppered crlenmeyer flask containing :ti1 cxccss oi it:tud:trd itlkali. After complete hydrolysis I i ; d taken pl;icc, ;.e., lrheii d l :he red crystals had disal)pe:rred, fmiiing :I pdlc yellow solution of chromic acid, [ h e Kel-1; t i i l l ? ~ r a athordughly rinsed out with distilled irtitrr, ; i l l \ r a ~ i i i i ~ g:ititled s to thc hydrolysate, diluted to a i!e,irid rolutiic a i i t l tiirateti iti the usual manner. The Kel-F t ~ 1 1 ) e1 ~ : ~dried s :inil ircighed with t h e tip, tlius giving the
The HF content of the potassium bifluoride was tletcr:L viinpli.: it provet1 t o I)e 9 i 5 5
iiiined h y alkali titration of
A Study of the Solubility Limits at 35" in the System NaBr-NaC1 BY EIIWARD I,. SIMONS,' CILLRLES '-1.ORLICK.\.xi) P,rrr,rp A. V.\ITGII.~N IZ1:csrvEI)
17, I !J-)?
JAYL.AR\.
A study of the ternary system sodium bromide-sodium chloride-6.C in0131 sodium hydroxide at 35" shows t h a t the two salts form a continuous series of solid solutions with no miscibility gap. This experimental result is in disagreement with the calculations of Fineman and ii'allace nnrf of Hovi :ind Hyviinen which indicate limited miscibility below 50 ' in the binary system sodium bromide-sotliuin chloritlr
Recent thermodynamic and statistical calculations by Fineman and Wallace2 and by Hovi and Hyvonen3 have indicated limited riiiscibility below 50" in the binary system sodium bromide-sodium chloride. The former investigators report a n upper consolute point a t .'341°K. and 0.45 mole fraction of sodium bromide, while the latter estimate a critical solution temperature of :32:3"E;. Both estiniates are based upon the nature of the free energy US. composition isotherms for the system. To establish experimentally the existence of any miscibility limits in this system below the reported critical solution temperatures, a study has been ~ n a d eof the ternary system sodium broiii;cle-sndiuni chloricfe-(i.G molal sodium hydroxide at :$.?". An exainination of the phases present :it equilibrium in this system showed the absence of any solid phase scilu-bility limits.
Experimental Choice of Solvent.-W'ater, the most convenient solvent for ii study of this kind, could not be used because below
3)2' sodium bromide cr from water as the dihyents were carried out in drate. Some solubility I absolute Iiiethanol," h i t the rate of attainment of equilibrium in the solid phase was prohibitively slow. X-Ray examination5 of the solids present after the pure salts had Ixen agitated for several months in contact with methanol 5howet1, in most cases, the presence of two phases, one a bromide-rich solid solution and the other pure sodium chloride. These results are in qualitative agreement with the observations of Matsen and Beach,6 who found t h a t a t elevated temperatures in the binary system the solution process occurs by the incorporation of chloride ions into the sodium bromide lattice anti not zlice versa. Soluliility data of Xkolajew and Rawitch7 show that i o d i u i r i bromide crystallizes in the anhydrous condition a t room teinprrature from concentrated sodium hydroxide Thew data :md those of Schreinemakers for Iilori~l? iiirlicate that a t 35' the soluhilitits of ~
.
. ..
( 4 j lI?:~.;nrernentsm a d e by Dr. S. IC. Blum, formerly of this Labora-
trrry.
( 5 ) \Ye are indehted to I l r . S.Weissman and the late Dr. A . J. Reis of t h e 15iiginrering Iiaperiment Station, Rutgers University, for their assistance i i i t l i v S - r ~ i ye.x:rrnirmtiun of the solids crystallized froni methanol (6) F. A . Matsen a n d J. Y. Reach, THIS J O G R N A L , 63, 3170 (19411,
( 7 ) V. t. Nikolajew and hf. I . Rawitch, Z h w . Obs. Khim., 1, 780 ( l W i ) , reported in Seidell, "Soliibilities of Inorganic and Metal Orgonic Ciimpoun~ls." \'I, I , 11, V a n Sostrand Co., Inc., New 'I'ork, s.\-,, 1940, p. 1 l . x .
SOLUBILITY LIMITSIN
Nov. 5 , 195%
these salts in 6-7 molal sodium hydroxide should be about 35 and 12 weight per cent., respectively. Experimental Procedure.-C.P. NaCl was used without further purification. Baker C.P. analyzed grade NaBr, found t o contain 0.08% NzCl, was dried by fusion and then The solvent, 6.611 molal Naround and stored at 120 H, was prepared from C.P. carbonate-free material. Mixtures of the components were brought to equilibrium by rotation a t 34.95 f 0.05' in rubber-stoppered polyethylene bottles containing stainless steel balls t o keep the solids well ground. The sampling and analysis of the saturated liquid solutions were carried out as described pre~ i o u s l y . ~The solutions were neutralized with nitric acid before analysis. The compositions of the equilibrium solid phases were obtained by an examination of their X-ray powder diffraction patterns. A slurry of the solid material was removed from a solubility bottle by pipet, filtered rapidly through a small buchner funnel, washed with 10 nil. of petroleum ether and dried in an oven a t 12i)o.13 The dried, solid agglomerate was broken up in a heated mortar and then added t o a warm, thin-walled Pyrex capillary tube which was sealed in the flame of a micro-burner. The filling was carried out in a n atmosphere of carbon dioxide. X-Ray diffraction patterns were made with a DebyeScherrer camera (360 mm. circumference) using nickelfiltered C u K a radiation. Exposure times of &8 hours were used. Calculations of the lattice constants from the resulting photographs were based upon the extrapolation method of Bradley and Jay." Compositions were obtained from lattice constants using the data of Nickels, Fineman and tVallace.12 Six weeks of stirring was usually suffrcierit for the establishment of equilibrium, as indicated by constancy of liquid and, in some cases, solid analyses. As a further check on the attainment of equilibrium, several bottles were prcpared by the method of duplicate c o m p l ~ x e s . ~ ~ Results.-The data are shown in Table I . tVheii they are plotted on a triangular diagram, the solubility curve runs smoothly between the values for the pure salts, showing no isothermally invariant point. At equilibrium, all solids consisted of a single anhydrous phase. The direction of approach t o equilibrium in the solid phase in this system was the same as that observed in the methanol study: the gradual incorporation of chloride ions into the bromide lattice. The distribution of the salts between liquid and solid phases is shown in Fig. 1 in which the mole fraction of sodium bromide in the dissolved salts ( y ) is plotted against its mole fraction in the solid solution (XI. The system represents a markedly non-ideal distribution of the Roozeboom Type 114 which a t lower temperatures will probably display the limited miscibility of Types I V or V . T o show that the 35' miscibility limits calculated from the data of Fineman and Wallace2 (30.3 and 81.8 weight per cent. sodium bromide) do not represent phases which are stable relative to solid solutions of intermediate coinpositions, the following experiment was performed. Two homogeneous solid solutions containing 28.2 and 83.7 weight per cent. sodium bromide were prepared by the method of Fineman anti IVallace .z An intimate mechanical mixture of these two solid solutioiis \vas made having the over-all composition of 40.6 weight per cent. sodium bromide, the composition of the equilibrium solid phase i n complex nuniber 52. A sample of the wet solid from this complex was rcnioved and replaced by a n approximately equal weight of the mechanical mixture. The equilibrium relationships in bottle number 52 were thus disturbed by effectively converting one solid phase into three without appreciably changing the over-all composition ratios. After one month of stirring the solid was removed. X-Ray examination showed the presence of only one phase with a lattice con-
8
.
(9) E. L. Simons and S. E. Blum, THISJ O U R K A L , 73, 5717 (1951). (10) Experiments with one solid solution and one intimate mixture of sodium bromide and sodium chloride showed t h a t 24 hours of heating a t 120' produced no change in t h e X-ray diffraction patterns of the solids. (11) A . J. Bradley a n d A. XI. Jay, Proc. Phys. S O L .( L o n d o n ) , 44, 563 (1932). (12) J . E. Nickels, 11. A. Fineman and W. E. Wallace, J. P h y s . Colloid Chem., 63, 6 2 5 (1949). (13) A. E. Hill and N. Kaplan, T H I S J O U R N A L , 6 0 , 550 (1938). (14) 13. Roozeboom, Z. p h y ~ i h C . h i n . , 8, 5 2 1 (1891).
THE SYSTEM
5265
NaBr-NaCIl TABLE I
SYSTEM NaBr-NaC1-6.611 MOLALNaOH AT 35' Complex
no.
Liquid solution
Wt. % NaCl
Wt' NaBr % 37.03 35.86 34.72 34.75 34.15 33.73 33.68 33.39 32.85 30.82 30.68 29.00 29.04 27.11 25.30 21.43 18.23 13.94 8.40 3 . 60
Solid solution Lattice constants W t . c/o Kx. units NaBr
5.959 100,00 5.928 93.2 0.70 5,904 88.0 1.16 70" 5.890 84.9 1.23 67b 5.862 78.7 68 1.46 5.815 74,s 1.61 69 5,813 66.2 71 1.68 5,752 48.9 1.60 72 5.726 40.6 1.75 52 5,679 22.9 63" 2.06 5.677 22.0 2.13 5sb 5.662 15.9 2.46 64" 5.658 13.9 2.41 54b 5.650 10.4 2.86 55 5.645 8.2 3.23 56 5.639 3.6 4.27 57 5.6:36 3.5 5.11 58 5,630 0.88 6.49 59 5.628 .o 60 8.58 5.628 .u 10.49 61 5,628 .o 12.07 62 NaCl initial solid phase. (In complex 70 the NaBr content exceeded the 2.5' solubility of this salt so that in preparing the complex a t room temperature a small amount of NaBr remained undissolved a t the time the NaCl was added.) NaBr initial solid phase. 50 66
stant of 5.715 Kx. units, compared with a value of 5.726 for the solid originally present in complex number 52. This corresponds to a solid phase composition change from 40.6 t o 36.6 weight per cent. sodium bromide.
,
1.0
0.8
0.6
? 0.4
0.2
0
0.2
0.4
0.6
0.8
1.0
X.
Fig. l.--lDistribution
of XaBr and S a c 1 between liquid and solid solutions at 35'.
Discussion.-Although the results of this investigation, showing complete miscibility a t 35" in the system sodium bromide-sodium chloride, are in disagreement with the predictions of Fineman and Wallace? and of Hovi and H y ~ o n e nthey , ~ do not point explicitly to the errors in the calculations of these workers. They indicate simply that the proposed free energy isotherms a t 33' are incor-
5266
JACK
HINE AND MILDREDHINE
Vol. 74
rect, but they do not permit one to assess individually the accuracy of the entropy and enthalpy terms used in obtaining the free energy values. Fineinan and Wallace' assumed that the entropy of formation of the solid solution between 0 and 350°K is the entropy of random mixing, while Hovi and Hyvonen3 based their entropy calculations upon a postulated local order for the solid solution lattice. The enthalpy values used by the latter investigators were also calculated on the basis of their postulated lattice structure, while those used by Finernan and 15-allace were determined experimentally on solids prepared by fusion of mixtures of the pure salts. Such experimental enthalpies may not be characteristic of the solid solutions in their equilibrium state a t room temperature. In the potassium broniide--potassiuiri chloride system Fontell and co-workcrs'j have shown that the enthalpy determined calorirnetricrtlly a t 25" for a 50 mole per cent. solid solution prepared by fusion is 10 cal. per mole higher than the value determined in the same manner for a preparation obtained by isothermal, isobaric crystallization from water (and therefore more nearly in a state of thermodynamic equilibrium a t room t e n ~ierature'~).The sensitivity of the free energy iso-
therms to small changes in the enthalpy values is illustrated by Wallace and Fineman's recalculation" of the phase diagram for the potassium bromide-potassium chloride system originally proposed by Fontell's on the basis of his measurements of the heats of formation in this system. By choosing an analytical expression for the data which was statistically more satisfactory than the one used by Fontell (thus effectively altering Fontell's enthalpy values by amounts ranging up to 25 cal.) they obtained a critical solution temperature a t 240°K. compared to Fontell's calculation of 398°K. The calculations of both pairs of investigators are probably correct in predicting incomplete miscibility below about 23OoK., but the uncertainties in the free energy values (arising from the enthalpy and entropy assumptions) make it difficult to establish with confidence a critical solution temperature. The results of this investigation indicate that, while partial miscibility is to be expected in this system, the critical solution temperature must lie below 30S"K. Acknowledgment.-The authors acknowledge with appreciation the support given to this investigation by the Rutgers University Research Council.
(13) N. Fontell, V. Hovi and 1.hlikkola, d i i i z . Ariid. Sei. Fciiiticu, .lloih.-Phys., Ser. A , 54 (19'49). (16) J. A. TVasastjerna, Soc. .Sei. I - e i i i z i r o Coiiiiiie,ifiilioiies f'hys.. .lJdth., 15, Xo. 3 (19419~i.
(17) E '. E Wallace and XI A Fineman, z b z d . , 14, No. G (19.48) (18) PIT. F o n t e l l , r b t d , 10, No. 12 (1939).
[COsTRIBVI'IOX FROM
THE SCHOOL O F
SCHENECTADY. KEW YORK
CHEMISTRY O F
THE
GEORGIA ISSTITUTE OF TECHNOLOGY]
The Relative Acidity of Water, Methanol and Other Weak Acids in Isopropyl Alcohol Solution1 BY J.4cK HINE .4ND MILDREDHINE RECEIVED APRIL 30, 1952 The relative acidity of five indicators is determined in isopropyl alcohol solution. By use of one, 4-nitrodiphenylamine, the acidity of about thirty-five alcohols and amides is determined. Methanol is found t o be more acidic than water. This, and the decrease in basicity resulting from the successive substitution of alkyl groups for the hydrogen atoms of water is explained in terms of "B-strain" and electronic effects. The relative effect of various R groups on the acidity of several functional groups and the effect of solvent changes on the relative strength of acids are also discussed.
Introduction In relation to an investigation of the mechanism of the basic hydrolysis of chloroform,3it was found that the reaction proceeded much more slowly in rnethanolic solution than in water or aqueous dioxane. There are several other reactions proceeding by the mechanism HA
+ I3 A
fast
A--
-
+ HB
slow -
--+products
(where B- is the conjugate base of the solvent) described in the literature which are also slowed by the addition of methanol to an aqueous reaction mixture. These include the formation of ethylene (11 Presented, in a preliminary form, before the Section on Physical and Inorganic Chemistry of the S I I t h International Congress of Pure ond Applied Chemistry, X e w York. S . I' , Sept. 10-13, 19.j1. (7, J. Wine, T H I S J O K R N . ~ ~ . 72, . 2-138 11g.jo)
oxide from ethylene halohydrins3 and the dealdolization of diacetone alcoh01.~ I n none of the three cases mentioned is the effect due to the change in dielectric constant, since i t is not shared by other low dielectric solvents such as dioxane, t-butyl alcohol and isopropyl alcohol. On the other hand, compounds such as ethylene glycol and glycerol, known to be somewhat more acidic than wdter,j have an effect similar to but somewhat larger than that of methanol. Stephens, hlcCabe and Warner3 have suggested that their data may be explained by the postulate that methanol is a stronger acid than water. By comparing the effect of 1 M methanol on the solubility of calcium hydroxide with its effect on 1 3 1 J I C i e \ c , n i C I XIrC ihr dnrl 1 C Warner i h i , TO, 2149 (1'248) ( 4 ) G .%herlof, t b t d , SO, 1272 11928) ii) L XIichaelis, B e ? , 4 6 , 3685 (1913) I \Iichaelis arid 1' Kona z:I) / i E ? ) i / 49, 232 (191.3)
