2866 DOBSON : CCCXCVI1.-The Partial Pressures of Aqueous Ethyl Alcohol. By H~YARD Jom EQLINTON DOBSON. IN connexion with work now in progress it became necessary to have accurate data for the partial vapour pressures of ethyl alcohol at 25" for aqueous mixtures containing up to 90% of alcohol. The system is in itself of obvious importance yet the existing data, for these properties seem t o be defective. Foote and Schole THE PARTIAL PRESSURES OF AQUEOUS ETHYL ALCOHOL. 2867 ( J . Amr. Chm. Soc. 1911 33 1309) obtained figures for the aqueous partial pressure which exhibit a curious break in sequence; this they ascribe to experimental error. Wrewski working at 39.7" (2. physikul. Chem. 1912 81 l) obtained similar discon-tinuities in the same region between 20 and 40% alcohol.The system has therefore been newly studied by the method indicated below; the results prove to be free from the anomalies hitherto prevalent and are believed to have considerable accuracy. Method.-The partial pressures were measured by the d p m i c or ''bubblmg" method. This method consists essenthlly in meaewhg the maw and composition of the mixed vapour which at 25" saturates a measured volume of a chemically inert gas and FIG. 1. FIG. la. thence calculating with the aid of the simple gas laws the partial pressures or more accurately the vapour concentrations of the two componenfs alcohol and water. With some of the higher strengths of alcohol the mass and the composition of the vapour were m d si.multaneously but it was generally found more convenient fo determine these separately for a given aqueous alcohol mixture.The determination of vapour composition is first described and secondly the determination of the mass of vapour saturating a measured volume of gas. Vapur Cmposition.-Nitrogen gas was saturated with the vapour of aqueous alcohol contained in a system of bubblers immersed in a thermostat electrically regulated at 25" & 0.02". The saturated vapour was then carried from the thermostat without condensation through an electrically heated tube E (Fig. l) to a U-tube immersed in a powerful refrigemat and in this U-tube the vapour was quanti 2868 DOBSON : tatively condensed. The composition of the distillate was deter-mined pyknometrically by the special means mentioned below. Since for this purpose as much as 3 grams of liquid had to be distdled it was essential to emure that the change in composition of the final saturating bubbler was negligible.This was achieved by placing a slightly stronger alcohol mixture in the first bubbler, A in which most of the evaporation occurred; the chilling due to latent heat was taken up in the long glass tube B; whiLst the remaining concentration changes were taken up by the second saturator C containing the same liquid as was in the k a l saturator, D. This was specially designed to remove all danger of spray being carried over from bursting bubbles; the absence of spray from this bubbler was proved in special experiments with highly coloured dyes. The exposed tube E was covered with asbestos on which Nichrome wire was wound and a temperature of 80-100" was maintained electrically throughout the experiment.An external glass sheath was sealed to E so that the heating coil exfended below the level of the water in the thermostat. The U-tube proved very efficient in condensing the vapour quanti-tatively; and liquid air used in many of the experiments appeared to have little advantage over a mixture of carbon dioxide snow and ether as a refrigerant for this purpose. The U-tube was 2 feet in total length and was placed in a correspondingly large Dewar flask filled with refrigerant and the distillation was made (generally over-night) in a current of nitrogen flowing a t the rate of 1 to 3 litres per hour. Special precautions were taken and mechanical indicators were devised which showed when any mishap had occurred.All determinations were in duplicate. When sufficient con-densate had been obtained the U-tube was removed from the Dewar flask and the pyknometer (Fig. la) was fitted by the ground glass joint J to the sampling tube F. When the condensate had thawed, and had been thoroughly shaken in the U-tube (to nullify fractiona-tion) it was forced by air pressure into the bulb of the pyknometer to the mark H. The tap was then turned so as to close the bulb, and to connect the joint J to the side tube I. A rubber cap was slipped over J and the pipette was immersed in the thermostat a t 25". When equilibrium was attained excess of alcohol was absorbed by spills of filter-paper till the meniscus stood a t the mark H.The pyknometer was now dried and before weighing it the passage through the tap was washed out from I to J by sucking 30% alcohol through and drying the passage with a current of air. In this way weighings reproducible to 0-2 mg. were obtained, giving densities agreeing within less than 0.0o01 unit. Densities The method of sampling was as follows THE PARTIAL PBESSURES OF AQUEOUS ETRYL ALCOHOL. 2869 a t 25" were used throughout for the estimation of percentage com-position. This waa read from the smooth curve obtained by p l o w on a large scale the accurate data of Osborne and McICelvy (Landolt-Bornstein-Roth " Tabellen," 1923 Ed. p. 448). TABLE I. The Vapow of Aqueous Alcohol. Liquid. @5' 0-986 18 0-97701 0.96572 0.96337 0.94360 0.93285 0.90890 0.89512 0.86068 0.84392 0.81337 4- ' Weight yo alcohol in liquid.6.21 12.36 20-51 28.40 33-90 39.32 50-46 56-50 71.09 78-07 90.12 in vapour. 35.80 54.20 66-17 72.9 1 75-38 77-10 80.03 81.23 84.40 86.20 91.90 Condensed vapour. d y . 0.93990 0.90045 0-87241 0.85634 0.85038 0-84634 0-83903 0.83614 0.828 14 0-82341 0.80860 Error of weight yo 0.03 0-04 0.03 0.01 0.04 0.06 0.02 0-04 0.01 0.03 0.02 figures. Temp. of condensetion of vapour. - 190° - 190 - 190 - 80 - 190 - 190 - 80 - 190 - 80 - 80 - 190 The results of these determinations are in Table I and each is the mean of two or more experiments. The maximum deviation from the mean figure for each concentration is expressed in the fifth column and shows the uncertainty in the percentage compositions.That these figures give a smooth vapour composition curve is shown in Fig. 2 in which these results a*re compared with the previous data of Foote and Scholes which are here represented by crosses on a broken line. Partial Pressures.-The apparatus used was a modification of that employed by Dobson and Masson (J. 1924,125 673) in measuring the aqueous vapour pressure of hydrochloric acid. A current of nitrogen gas saturated with the vapour of the aqueous alcohol at 25" passed through a weighed bubbler containing about 50 g. of concentrated sulphuric acid in which both the aqueous and the alcoholic vapours were quantitatively condensed (Masson and McEwan J .SOC. Chem. Id. 1921 40 2 9 ~ ) . The device used for collecting the issuing gas a t atmospheric pressures has been cGLrettdy described (Dobson J. 1924 125 1968) ; it materially increased the accuracy of the experiments since it maintained in the saturator a constant total pressure which was measured by a manometer. The importance of measuring this quantity has been emphasised by Berkeley (Nature 1918 95 54). Foote and Scholes were apparently able to neglect this through the arrangement of their apparatus and on this account the method of calculation materially differs from theirs. The volume of nitrogen v saturated at 25" with the vapour i 2870 DOBSON : related by equation (1) to the volume V collected in the aspirator and measured at a temperature To absolute and a pressure.B - h, where B is the barometric height and h the vapour pressure of water a t the temperature of the aspirator To. where B is the total pressure in the saturator measured by the manometer and p = p + p& is the total vapour pressure of the mixture which must be estimated by successive approximations. FIG. 2. The uapour compo8ition of aqueoua alcohol at 25'. v = 298*1V(B - h)/{T(B - p ) > . . . . (1) 20 40 60 80 100 Weight yo of alcohol in Zipid. 0 Expe&wntaZ points here determined. x Found by Foote and Scholes (1911 loc. cit.). I n no m e were d u p l h t e experhenb auficiently divergent to be shown by sepamte points on the scale of this figure. The volume v is saturated a t 25" by m grams of vapour having a composition x% alcohol as given in Table I.Taking the molecular weights 18.02 and 46.06 for water and ethyl alcohol respectively we obtain from the simple gas laws the following equations for the partial pressures : (2) V(B - h) loo . . * 3462 mT(B - p ) 100 - x x p* THE PARTIAL PRESSURES OE' AQUEOUS ETHYL ALCOHOL. 2871 Kendall ( J . Amer. Chern. Sw. 1920 42 2481) has shown that saturated steam a t 100" obeys van der Waals's equation when the comtanb a and b are deduced from the critical data. Calculation then shows that the deviation of saturated water vapour from the perfect gas laws amounts to only 0.03% a t 25". A similar calcu-lation for saturated ethyl alcohol vapour a t the same temperature shows that the gas laws hold within 0.1%. The accuracy of the equations is therefore as great as that of the experimental data.The experimental figures a m given in colnmns 2 to 7 of Table 11. With the aid o€ equations 2 and 3 the partial pressures of aqueous and alcoholic vapour pw and p, have been calculated and are given in columns 8 and 9. TABLE 11. Experimental partial pressure results. Pressure in millimetres of mercnry. Wt. % alcohol. lldass c . c-'- in Vol. ** Absolute." " Relative.'' Liquid. z. rn. 7. !.?Oabs. B - h . B,. PI. p d ~ . pa. pr. me. Vapoar. grams. litres. - 0 0 0.3449 15.10 293.9' 738.4 778.4 23.75 - - - -0.4638 20.31 293.9 739.5 779.5 23.75 - - - -12.36 54.20 0.6107 12-24 292.5 742.2 764.4 22.80 10.55 - - -0-7888 16.00 293-2 740.3 763.2 22.60 10.46 23-80 22-55 10.44 0.7454 15.10 293.9 738-4 762-1 22-71 10.51 23-75 22.71 10.51 20.51 66.17 0.9739 15.41 297.1 736.8 766.0 21-72 16-61 23.77 21.70 16.60 14427 16-10 294.1 732-0 755.9 21-87 16.73 23.78 21-84 16.n 1.0366 16.14 293.8 731-3 754.9 21.65 16-56 23.51 21.87 16-73 38-40 72-91 1.6580 20.30 285.8 7494 768.3 21.17 22.29 23-74 21.18 22.30 1.6316 20.33 286.7 735-8 763.3 21.11 22-22 23-n 21-15 '22.26 33-90 75.38 1.4092 16-49 294.3 737.8 762.7 20.83 24.95 23.77 20.81 24.93 1-3859 16.35 594.3 73543 762.2 20.71 2479 23.68 20.77 24.86 1.2826 15-00 294.5 738.2 762.4 20.84 24-95 - - -39.32 77-10 1.4590 16-37 296.7 '132.2 754.1 20-24 26-65 23-45 20.50 26.99 14367 16.10 297.5 731.9 754.7 20-34 26.79 - - -2.7841 31-10 294-7 741.9 761-1 20.13 26-50 23-43 20.39 26-84 1.4935 16.15 292.0 746.5 763-0 20-55 27.07 23.73 20.56 27.09 50.46 80.03 2.0296 20.31 293.9 739.5 765.4 19.63 30.77 23-75 19-63 30.77 2.0229 20-33 293.7 735.6 760-5 19.50 30-57 23.59 19-63 30-78 2-0348 20.33 393.6 7374 762.8 19-62 30.76 23-80 19-58 30.70 56.50 81.23 1.5886 15.02 590.5 747.2 760.7 1846 32.09 23.85 19.04 32.22 1-7458 16.36 289-0 749.7 763.3 19-04 32.22 23-86 18-95 32-07 2.9705 28.26 292.4 732.0 7484 1944 32-22 - - -71-09 84-40 2-3037 20.31 293.7 738.8 763-6 17-28 36.56 23-74 17.29 36-58 2.3741 20.32 288.4 741.0 761.5 17.37 36-77 - - -78.07 86.20 3.2033 26.18 291.3 745-5 764.5 16.19 39.56 23.63 16.27 39.76 1.8569 15.30 292-7 728.2 145.4 16.08 39.28 - - -90.12 91-90 1.9786 14.45 293-7 728-4 746.7 10.66 47-31 23-66 10.70 47-49 2.1785 16.15 296.7 730-1 752.1 10-67 4735 W70 1069 4745 100 100 4 .n ~ ~ 29.70 292.3 746.0 759.3 - 69-01 - - -~~~~ 1) (2) (3) (4) (5) (6) (7) (s> (9) (10) 01) P I As in the experiments of Dobson and Masson (loc. cit.) a simul-taneous determination was made of the vapour pressure of pure water figures being obtained for the mass of water saturating the =me volume of gas a t a total pressure B, measured by a second manometer. Using an equation similar to equation 2 (in which 2 of come becomes zero) figures were obtained for po the vapou 2872 DOBSON pressure of pure water and these constitute column (10) of Table 11. These figures provide not only a measure of the accuracy of the experiment but taking the standard mean value po = 23.75 mm. at FIG. 3. l ’ h e partial pressurea of aqueow alcohol at 25’. 25” they yield an independent measure of the vapour pressures relative to that of water at the same temperature.The “ relative ” partial prwures so obtained are given in columns 11 and 12 of Table 11. In calculating the mean values of these experiments both ’’ rela THE PARTIAL PRESSURES OF AQUEOUS ETHYL ALCOHOL. 2873 tive ” and “ absolute ” figures were considered to be of equal value, and these mean partial pressures together with the total vapour pressures and the vapour compositiom are given in Table 111. The vapour pressure resulb have been ,plotted graphically in Fig. 3 in which they are compared with Foote and &holes’s resulb. TABLE 111. Weight yo alcohol T-T-liquid. vapour. 5. 0 0 6-21 35-80 12-36 54.20 20.51 66-17 2840 72.19 33-90 75-38 39.32 77.10 Vapour pressures in millimetres.7-Aque- A~co-ous. holic. Total. 23.75 - 23-75 22.67 10.50 33.17 21-78 16-66 38-44 21.15 22.27 43.42 20.79 24.90 45.69 20-36 26.85 47-21 - - -Weight yo alcohol in in liquid. vapour. 50-46 80.03 56-50 81.23 71-09 84.40 78.07 86.20 90.12 91-90 X. 100 100 Vapour pressures in millim0tl-M. -7 Aque- A~co- ow. holic. Total, 19.60 30.73 50-33 19.01 32-16 51-17 17-31 38-64 53.95 16.18 39.63 55.71 10.68 47-40 68.08 - 59-01 69.01 Discussion. It appears that these experimental results difEer essentially from those of Foote and Scholes in the measurement of the composition of the vapour. For when these authors’ figures for partial pressures are recalculated using vapour compositions interpolated from the present data instead of from their own the resulting figures for the partial pressures fall in almost every case on the smooth curves of Fig. 3. Foote and Scholes (h. cit. p. 1317) used an analytical method in which the alcohol content of the vapours was determined by combustion over copper oxide the percentage of water in the vapour being estimated by subtracting from the total water pro-duced 3/2 g.-mols. for each g.-mol. of carbon dioxide weighed. In spite of the analytical care displayed by these workers it appears that inaccuracies were liable to occur especially where the water content of the vapour was low. The author wishes to express his indebtedness to Professor Irvine Masson for the interest he has taken in this work and for much helpful criticism and advice. THE UNIVERSITY DUR,E~AM. [Received October 318& 1926.