THE SYSTEM BENZENE-ETHYL ATLX3HOTr-\VATRR ETC. 1397 CLII .-The System Benzene-Et hy 2 A I cohol- Water between c 25" and - 5". By NEVIL VINCENT SIDGWICI~ and WILLIAM JAMES SPURRELL SEVERAL investigations have been made on parts of this system at various temperatures but no complete study has yet been published of the conditions of separation of solid benzene from a mixture of the three components. This question is of general theoretical interest and also of some practical importance in view of the use of mixtures of benzene and alcohol as motor fuel. Considering first the two-component systems involved the system water-benzene is soon disposed of. The mutual solubilities a t any temperatures with which we are concerned are very small. A t 23O the solubility of water in benzene is 0.061 per cent.* (Groschuff, Zeitsch.Elektrochem. 1911 17 348); a t 1 5 O that of benzene in water is 0.15 per cent. (Moore and Rolaf Proc. Roy. SOC. 1905 [B], 77 96). If we assume that the solubilities change1 with temperature in the same proportion as the vapour pressures their values near t'he triple point (5-4O) will be water in benzene 0.082 benzene in water 0.092 per cent. The temperature of the triple point solid benzeneliquid benzene-water has been found to be 5.39O (Sidg-w i d tlhis voll. p. 1340)-0ne-tenth of a degree below the freezing point of benzene; that of the triple point ice-water-benzene is by calculation -O*OIGo the solubility of the benzene a t this temperature being 0.069 per cent. The freez-ing-poinh curve of solutions of water in alcohol has been measured (for example by Pickering T.1893 63 1015); a 10 per cent. solution freezes a t about -4*6O but the solubility of benzene in such a solution is so small that we did not investigate the ice curve. The system benzeneethyl alcohol is of much greater importance. The freezing-point curve has been examined by several observers. Our own results (extrapolated for anhydrous alcohol) are nearest to those of Piclcei-ing (Zoc. cit. p. 1019) with whom Viala (BUZZ. SOC. chim. 1914 [iv] 15 5) and R6zsa (Zeitsch. Elektrochem., 1911 17 934) also agree fairly closely. McIntosh ( J . Physical Chem. 1896 I 4-80) who also measured this curve states that his * All solubilities in this paper are expressed in grams of solute per 106 gram5 of solution.The system water-alcohol also scarcely concerns us alcohol may have contained water; from our results i t seems that it must have contained about 4 per cent. The three-component system has been examined by McIntosh (Zoc. cit.) Taylor ( J . PJhysicaZ C’Jwm. 1896 1 301 461) Lincoln (ibid. 1900 4 176) Bonner (ibid. 1910 14 779) and R6zsa (Zoc. c i t . ) . McIntosh investigated the freezing point of the benzene and drew attention to the remarkable fact that the freez-ing point of benzene containing alcohol is raised by the addition of water; he correctly explained this as being due to the increase of the vapour pressure of benzene on addition of a component in which it is insoluble but his numerical results are vitiated by the presence of water in his original alcohol.Taylor Lincoln and Bonner examined the two-liquid equilibrium between loo and 2 5 O , and determined the composition of the conjugate solutions (tie-lines). R6zsa investigated the system both above and below the freezing point; his results are difficult to understand as he does not define his concentrations but on the most probable hypothesis they are roughly in agreement with ours. These results taken together give an incomplete account of the behaviour of the system in the neighbourhood of the freezing point of benzene. We have endeavoured to extend and complete them. E X P E R I M E N T A L . The method adopted was to prepare mixtures of alcohol and water of known strength; these were mixed with various known proportions of benzene and the temperature was determined a t which the liquid separated into two layers or the benzene crystal-lised out.The concentrations are always expressed in grams of component per 100 grams of solution. The alcohol was obtained by distilling ordinary “ absolute ” alcohol with lime; the percentage of water in it was determined from the density using the tables given in the last edition of Beilstein’s “ Handbuch der Organischen Chemie.” From this the requisite aqueous mixtures were prepared gravimetrically . The purest alcohol used (99.5 per cent.) was obtained from this by redistillation after treatment with anhydrous copper sulphate. The benzene was freed from thiophen by sulphuric acid and frozen out seven times; it was then distilled over sodium. The apparatus consisted of an ordinary Beckmann tube with air-jacket thermometer and stirrer.A known weight of one of the liquids was placed in it and successive amounts (usually 1 c.c.) of the other added through the side-tube from an accurate pipette, which had been carefully graduated by weight with the particula ALCOIIOL-WATER BETWEEN -1 25" AND - 5'. 1399 liquid to be used in it. The tube was warmed or cooled by a bath of water ice or ice and salt. The thermometers had been compared with an instrument standardised a t the Reichsanstalt and were checked by redeter-minations of the ice-point. The readings were corrected when necessary for the emergent stem. At high concentrations of benzene the method of supercooling could be used as the amount of solid benzene separating did not seriously affect the proportions.At lower concentrations the point was observed at which the last TABLE I. Alcohol 99.5 per cent. Benzene. Per cent. 92-42 80.29 75-32 70.92 67.05 62.05 57.85 54-16 50.95 Alcohol, 95.20 93.09 91.12 86-86 81-82 77.15 69.23 65.56 58.76 53.30 50.55 45.30 Tenip. + 2-95' 1.26 0.50 -0.14 - 0.86 - 1-94 - 2.94 -4.16 - 5.42 98.0 per cent. + 3.50 3-10 2-76 2-22 1-62 1-08 0.03 - 0.72 -2.16 - 3.90 - 4.62 -5.72 Alcohol 95.76 per cent. 94-50 + 3.48 89.60 2.84 81.15 1.86 77.51 1.48 65.70 0-12 59.35 - 1.04 55-46 - 1.92 52.04 - 2.75 49-02 - 3.62 46.34 - 4.42 43.94 - 5-34 Alcohol 90.06 per cent. 89.18 +30*7 (A) 87.55 26.1 (L) 86.88 23.9 (L) Alcohol 90.06 per cent.-contcl. Benzene. Per cent. Temp. 85.24 16.1" (L) 84-30 11.0 (L) 83.15 5.2 (L) 82.54 2.0 (L) 81-19 2.51 74.22 2-1.3 68-33 1.78 59-42 0.95 55.60 0-42 49.31 - 0.51 46.62 - 0.91 42.09 - 1.97 37.65 - 3.54 35.76 - 4.30 34.05 - 5.14 Alcohol, 73-89 72-81 71-63 70.35 65.93 67-38 Alcohol, 63-52 61.66 57-41 50.55 47-14 44.19 44-19 39-24 33-03 30.64 26.78 86.0 per cent. +28-3 (L) 22.7 (L) 17.6 (L) 12.9 (L) 8.5 (L) 4.5 (L) 80.10 per cent. $39-2 (L) 32.1 (L) 19.6 (L) 9.2 (L) 5.2 (L) 1.6 (L) 2-45 3-00 0.96 0.30 - 1.12 Alcohol 74-45 per cent. Benzene. Per cent. 44.43 35-23 35.12 3 1-02 28.90 27.04 27-04 24-03 18.64 17-65 15-96 Temp.+29-5" (L) 20.3 (L) 15.2 (L) 3.00 (L) 7.75 (L) - 1.25 (L) $- 2.30 + 1.20 - 2.25 - 3.10 - 4.85. Alcohol 69.08 per cent'. 27-67 +25*3 (L) 24.31 17.3 ( L ) 25.88 21.2 (L) 22.91 21.67 '20.55 19.06 18-20 17.15 15.37 13.72 -12.37 11-81 Alcohol, 10.53 10.05 9.25 8-56 8-27 8-27 6-66 6.34 13.5 (Li 6.5 (L) 3-30 1-52 0.02 1.34 9.9 (L) 1-95 (L) - 3.02 - 3.70 57-66 per cent. +154 (L) 12.15 ( L ) 7-1 (L) 2.3 (L) - 0.40 (L) -+ 2-15 - 2.92 - 4.20 25.18 - 2.02 23.76 - 2.97 Alcohol 39 p e ~ cent 32-50 - 3-75 Less than 21.36 - 4.67 1.0 f-350 ( L 1400 SIDGWICK AND SPURRELL THE SYSTEM BENZENE-ICTIIWTJ crystals remained in equilibrium with the liquid. For the two-liquid equilibria the point was taken a t which the liquid on slow cooling became sufficiently turbid to obscure a bright object placed behind it.The temperature is in every case the mean of two or three concordant observations; these in the solutions containing more benzene did not differ by more than 0.lo but in the weaker solutions differences of as much as 0 - 2 5 O could not always be avoided. Each series in the table was carried out with alcohol of the strength stated a t the head; the percentage of benzene given in the first column is the number of grams in 100 grams of the result-ing mixture. The points marked ( L ) are those in which the separation was into two liquids; in all other cases they are the freezing points of the benzene. One or two solutions gave points of both kinds a higher (stable) solid and a lower (metastable) liquid point.From the curves the values for a series of round temperatures were interpolated and these are collected in table 11. The first column gives the temperature the second the freezing points of mixtures of benzene and anhydrous alcohol from the results of Pickering (Zoc. cit.); the other columns are from our own experi-ments. The two-liquid points are distinguished by an asterisk. The total composition of any liquid is of course obtained thus: the percentage of benzene is that given in the table; the per-centage of water or alcohol is obtained by multiplying the difference between this and 100 by the proportion of water or alcohol (from the figure a t the top of the column) in the aqueous alcohol used.At the foot of each column is given the percentage of benzene and the temperature a t the triple point solution I-solution 11-solid benzene. The results are plotted for temperatures from +5" t o - 5 O in Fig. 1 the full lines representing two-liquid and the dotted lines liquid-solid equilibria. The results are given in table I (p. 1399). Discussion of Results. The equilibrium of a three-component system can be expressed only by a solid diagram; this can take the form of a triangular prism the isothermal curves being represented on a series of triangles. Each of the curves obtained by adding varying quanti-ties of benzene to aqueous alcohol of a definite strength (which are plotted in Fig. 1) represents a section of this solid diagram by a plane which contains the temperature axis passing through th TABLE 11.the percentage of benzene in the liquid a t equilibrium. The first line gives the percentage of alcohol in the mixture of alcohol Alcohol per cent. 100 Pic kering - 25" .................. 15 .................. -5 .................. 99.48 4 .................. 97-57 3 .................. 93.82 3 .................. 2 .................. 87-00 2 .................. 1 .................. 80.89 1 .................. 0 .................. 74.08 0 .................. - 1 .................. 67.79 - 2 .................. 63-87 - 3 .................. 59.10 - 4 .................. 55.56 - 5 .................. 52.60 --I -............... -99.5 -__ --92-65 85.28 78-36 71.87 66.45 61.70 57.54 54-49 51-85 -----98.0 95.76 -_ _-02-52 91.22 85-06 82.64 76.58 72.78 69.25 65.03 63.93 59.48 59.62 55.07 56.00 51.17 52.71 47-82 49.90 44-81 96-76 - -- -- -- -- { +4.10" 90.06 87-22" 85.05" 83.05" 82-95" 82.76" 71-90 82*55* 59.80 82*35* 52.01 82*10* 46-39 42.00 39.00 37.48 34.45 82.62 f-2.55" I 80.10 69.41 55.00" 46.92" 46-45-39.24 44.50" 33.15 43* 29.85 43.08" 27.34 25.40 23.52 22.21 21.00 44.95 + 2.50" 1402 SIDGWICR AND SPURRELL THE SYSTEM BENZENE-ETHYL benzene edge of the prism and cuts the opposite side along a vertical line corresponding with the particular mixture of alcohol and water employed.The isothermals for three temperatures, namely +25O +3O and -so are plotted on the triangle in Fig.2. A t 2 5 O there is of course only the two-liquid equilibrium represented by the innermost curve. The dotted tie-lines giving 5" 3" . I " B 2 0" u F c -11" - 3" - 5" FIG. 1. 100 8 0 60 40 20 Benzene. Weight per cent. Full lines Liquid-liquid equilibria. Dotted lines liquid-solid ,, the compositions of the two liquids in equilibrium with one another are taken from Taylor's measurements (Zoc. c i t . ) . At telmpelratares below 5*49O sollid benzeaa is prelsent at B (pure, benzene) and extends further along BA as the temperature falls. At + 3 O for example the solid is in equilibrium with a mixture of benzene and alcohol containing 6.18 per cent.of the latter. I f water is added to this mixture the freezing point of the benzene is raised because the water being insoluble in the benzene ATL'OTTOL-WATER BETWEEN -k 25" AND - 5". 1403 increases its vapour pressure. I n order therefore to remain in the 3 O isotherm if water is added alcohol must also be added; the solid-liquid curve (the dotted line) will thus incline to the right. Our results show that when it has reached the point a a second FIG. 2. Water. W Benzene. Alcohol. $25" Isotherm -0- + 3" Isotherm Solid-liquid - - - x - - -Liquid-liquid -@--$ Isotherm - - - . _ - -Dotted tie lines Taylor 25". Full tie lines Sidgwick and Spurrell at 3". I 1 liquid phase appears; we have now reached the triple line solid benzene-solut'ion I-solution I1 (conveniently written S-L,-L,) .This new liquid is the conjugate solution the composition of which is given by the point b which is a t the other end of the tie-line ab ; being in equilibrium with the first liquid it must have th 1404 TITE SYSTEM BENZENE-ETHYL ALCOHOL-WATER ETC. same pressure of benzene vapour and so must also freeze a t 3 O . Thus at 3 O the L,-L2 part of the curve is confined to the part ccb; a t b the solid curve begins again passing between the two-liquid curve and the side AW. The triple line that is the curve joining the whole series o€ triple points S-L,Tl for all mixtures of alcohol and water on addition of benzene will begin on the line BTV (no alcohol) a t 5*39O (tlriple point, watler-benzene liquid-benzene solid) atl tmo point,s corresponding re'spectivdy with 99.97 and 0.092 per celnt.of benzene. As successive quantities of alcohol are added the temperature of the triple points will fall but a t any given tempera-ture (not too low) there will be two triple points corresponding with two conjugate solutions joined by a tie-line; these points will move towards one another with a fall in temperature as the alcohol increases but more rapidly from the water than from the benzene side (owing to the position of the tie-lines) until they finally meet a t the plait-point where the tie-lines vanish. The temperature of this point is the lowest a t which two liquid phases made up of water alcohol and benzene can co-exist. Our results show that this temperature is +2-50° and the composition of the liquid about 80 per cent.of benzene and 2 per cent. of water; a t 2 5 O Taylor found a t the plait-point 80.2 per cent. of benzene. The solid benzene curve has the form of the outer dotted line; somewhere near the water end it must meet the ice-curve which a t - 5 O starts on A'CV a t 91 per cent. of water but the solubility of benzene in this region is so small that it could not be investigated. One point of practical importance may be mentioned in view of the use of these mixtures for motor fuel. I f a fourth com-ponent (another hydrocarbon) were added the freezing point of the benzene would be proportionately lowered (some 5O for 10 per cent.) but; if a paraffin were used for this purpose it would certainly increase the tendency of the liquid to separate into two layers. The solubility of paraffins in alcohol containing water is extraordinarily small. The system hexane-alcohol-water has been examined a t Oo by Bonner (loc. cit. p. 777); the heterogeneous area occupies nearly the whole triangle. For example whilst 90 pelr clelntl. aqueloas alcolhoJ will dissollve up to1 f o a r timels itls weight of benzene it will only dissolve a third of its weight of hexane. At -5O no second liquid can exist. ORUANIC CHRXISTRY LABORATORY, OXFORD. [Received Septembcr 6th 1920.