Method Of Calculation

In calculating the composition of the ternary mixture it is assumed, as before, that the corrected weights of the two distillates are equal to those of the ternary and binary mixtures respectively which would be obtained if the separation were perfect. That being so, in case I., the weight of benzene in the ternary mixture is simply that in the original mixture ; the weight of alcohol is that taken, less the amount in the binary mixture, which can be calculated; the weight of water is given by difference.

The composition of the ternary mixture was also directly determined, and it will be seen that the agreement with the mean value obtained by the distillation method is very satisfactory.

On the other hand, some of the individual values, notably those of alcohol and water in I., differ somewhat widely from the mean. The explanation of the rather large errors in this distillation is given below.

Cases to which the Distillation Method is not applicable.When small quantities of alcohol are successively added to water, the boiling point is rapidly lowered ; the middle temperature between the boiling points of the pure components is, in fact, reached when the mixture contains 6.5 molecules per cent of ethyl alcohol. On the other hand, water must be added to alcohol until the mixture contains 25 molecules per cent before the boiling point rises 0.1° above that of pure alcohol, and with equal molecular proportions the rise of temperature is only 1.50.1

Very similar results are obtained with normal hexane and benzene ; a mixture containing 16 molecules per cent of benzene boils only 0.1° higher than normal hexane, and the mixture which has the boiling point 74.6°, midway between those of the pure components, contains 79 molecules per cent of benzene.2

In both cases mixtures of minimum boiling point, very rich in the more volatile component, are formed so that the separation would be that of the mixture of constant boiling point from that component which is in excess.

Form Of Boiling Point Composition Curve

It is, however, found practically to be impossible in either case to separate the mixture of minimum boiling point even from the less volatile component, although the difference between their boiling points is considerable. In both cases the boiling point composition curve is very flat where the more volatile component is in large excess, and it is in such cases - when the curve is very flat at either one end or the other - that one at least of the components is exceedingly difficult to separate, and that the distillation method cannot be relied on for the determination of composition.

Ethyl Alcohol And Water

Thus on distilling ethyl-alcohol-water mixtures, containing from 15 to 25 per cent by weight of water, through an 18-column dephlegmator and calculating the percentage of water in the mixture of constant boiling point in the usual way from the weight of distillate below the middle point, values from 7.6 to 8.0 instead of 443 per cent were obtained. Referring back to the calculation of the composition of the ternary ethyl-alcohol-benzene-water mixture from the first distillation (p. 180), if we take 7.8 as the percentage of water in the binary A.W. mixture, the calculated composition of the ternary mixture becomes:

1 Noyes and Warfel, "The Boiling Point Curves of Mixtures of Ethyl Alcohol and Water," Journ. Amer. Chem. Soc, 1901, 23, 463.

2 Jackson and Young, " Specific Gravities and Boiling Points of Mixtures of Benzene and Normal Hexane," Trans. Chem. Soc. 1898, 73, 923. which agrees very well indeed with that observed.

General Conclusions

In the great majority of cases the distillation method may be safely employed for the determination of the composition of a mixture which separates normally into its components, provided that a very efficient still-head be employed and that the distillation be carried out slowly. But it must be borne in mind that from a mixture of two liquids it is almost always more difficult to separate the more volatile component than the other, and therefore, if the original mixture contain a relatively very small amount of that component, a second distillation may be necessary, and a large quantity of the original mixture will be required in order to give a sufficient amount of distillate for a second operation. As regards the separation of three or more components from a mixture, it must be remembered that, as a general rule, the least volatile component is the easiest, while the intermediate components are the most difficult, to separate.

If a binary mixture of constant boiling point is formed, the composition of the original mixture may be determined if that of the mixture of constant boiling point is known; or if the composition of the original mixture is known, that of the mixture of constant boiling point may be determined. The method may even be applied to the determination of the composition of a ternary mixture of constant boiling point.

It appears to be only when the separation of the components (either simple substances or mixtures of constant boiling point) by distillation is exceedingly difficult that the method is inapplicable.

The "middle point" method was carefully investigated by Mlle J. Reudler1 with good results; it has been used for the determination of the composition of binary and ternary mixtures of constant boiling point by Wade,2 Atkins and Wallace,3 Hill,4 and Merriman 5; it has frequently been employed by Atkins, and Lecat6 made use of it in the case of more than 350 out of the 1100 azeotropic mixtures examined by him. Lecat states on p. 56 that it is only when the separation of the fractions by distillation is very difficult that the method becomes inapplicable.

1 "Eenige Opmerkingen over Sydney Young's Distillatieregel," Versl. Amst., 1903-1904, 12, 968 [Eng. trans. Proc. Amst., 6, 807].

2 Trans. Chem. Soc, 1905, 87, 1656. 3 Ibid., 1912, 101, 1179 and 1958. 4 Ibid., 1912, 101, 2467. 5 Ibid., 1913, 103, 1790.

6 f La Tension de vapeur des melanges de liquides, Pazeotropisme," Brussels, 1918.

The method has also been found useful for the analysis of commercial products such as crude benzene, toluene, etc.

In the case of very complex mixtures, such as petrol, no attempt is made to estimate the relative amounts of individual hydrocarbons present, but for commercial purposes 100 c.c. of the "spirit" is slowly distilled under specified conditions usually without an efficient still-head. The initial and final temperatures are noted and the weights or volumes of distillate coming over between these limits, the range of temperature for each fraction being usually 10°. Some chemists, however, consider that an improved still-head should be used, and Washburn recommends a more complete separation of light oils by distillation through a combined Hempel and constant temperature still-head.1

1 Rittman and Dean, "The Analytical Distillation of Petroleum," U.S. Bureau of Mines, Washington, 1916, Bull, 125, Petroleum Technology, 34; Lomax, "Testing and Standardisation of Motor Fuel," J. Inst. Pet. Tech., 1917-18, p. 6; Anfilogoff, "Distillation Test of Petrol," J. Soc. Chem. Ind., 1918, 37, 21 T; Phillips, " Some Laboratory Tests on Mineral Oils," J. Inst. Pet. Tech., 1919 ; J. Soc. Chem. Ind., 1919, 38, 393 R; Dean, " Motor Gasoline Testing," U.S. Bureau of Mines Technical Paper, 214, Petroleum Technology, 52; Luynas-Bordas Apparatus as used in France, J. Soc. Chem. Ind., 1920, 39, 220 A.; F. M. Washburn, "Constant Temperature Still-head for Light Oil Fractionation," J. Ind. and Eng. Chem., 1920, 12, 73.