The composition of a mixture of liquids which are not difficult to separate may, as a rule, be ascertained with a fair degree of accuracy from the results of a single distillation with an efficient still-head, or, if the components are more difficult to separate, from the results of two or three distillations.

It will be well to consider first the case of mixtures which tend, on distillation, to separate normally into the original components.

Taking first the simplest case, that of a mixture of two liquids, it is found that the weight of distillate that comes over below the " middle point" is, as a rule, almost exactly equal to that of the more volatile component, even when the separation is very far from complete.

By "middle point" is to be understood in all cases the temperature midway between the boiling points of the two components, whether single substances or mixtures of constant boiling point, into which the original mixture tends to separate ; or, in the case of more complex mixtures, the temperature midway between the boiling points of any two consecutive fractions of constant boiling point.

If the original mixture tends to separate on distillation into more than two, say n, components, the weights of these components will be very nearly equal respectively to (No. 1) the weight of distillate below the first middle point, (No. 2 to n-1) the weights of distillate between the successive middle points, (No. n) the weight above the last middle point.

Loss By Evaporation

It is obvious that there must be some loss by evaporation, which always makes the weight of distillate somewhat too low. This loss will be greater as the initial boiling point of the liquid is lower and as the temperature of the room is higher. It is not, however, proportional to the amount of liquid distilled, for a great part of it is caused by the saturation of the air in the flask and still-head with vapour when the liquid is first heated ; since this vapour is mixed with much air its partial pressure is low, and a large proportion of it escapes condensation when cooled in the condenser. Under otherwise similar conditions the loss is, therefore, roughly proportional to the volume of air in the still and still-head, and it is advantageous to use as small a flask as possible and to employ a still-head of as small a capacity as is consistent with efficiency.

1 Young, " Experiments on Fractional Distillation," Journ. Soc. Chem. Industry, 1900, 19, 1072. Young and Fortey, "Fractional Distillation as a Method of Quantitative Analysis," Trans. Chem. Soc, 1902, 81, 752.

Choice Of Still-Head

A plain wide tube or one with spherical bulbs is the least satisfactory, but the "pear" still-head, owing to the diminished capacity of the bulbs and the increased efficiency, gives much better results. Of all forms, the "evaporator" is the best, because the capacity is very small relatively to the efficiency, and the amount of condensed liquid in it is smaller than in any other equally efficient still-head; moreover, almost the whole of the condensed liquid returns to the still at the end of the distillation. With a liquid of low viscosity, like one of the lower paraffins,, the quantity left in the still-head is almost inappreciable, and in other cases it may be reduced to a very small amount by disconnecting the apparatus, while hot, from the condenser, tilting the tube from side to side to facilitate the flow of liquid back to the still, and, if the original form of "evaporator" still-head is used, shaking out any liquid remaining in the funnels.

Estimation Of Loss By Evaporation

The following may be taken as an example of the estimation of loss by evaporation. Mixtures of benzene and methyl alcohol, one with benzene, the other with methyl alcohol in excess (these liquids form a mixture of minimum boiling point), were distilled through an " evaporator " still-head of five sections of the original form, the distillation being stopped in each case when the middle point was reached. The following results were obtained :Table 59

When the benzene was in excess, it is certain that the amount of it left in the still-head did not exceed 0.1 gram, and the loss by evaporation was therefore estimated as 0.5 gram; in calculating the composition this amount was added to the observed weight of distillate. With methyl alcohol in excess the total loss was greater, but this more viscous liquid does not flow back so completely to the still, and the loss by evaporation was taken to be the same, 0.5 gram. With an "evaporator" apparatus of five sections the loss by evaporation is usually from 0.3 to 0.5 gram.

Mixtures Containing Two Components

The following are examples of the distillation of mixtures of two liquids which separate normally into the original components.

Table 60

Methyl Alcohol and Water Boiling points - Methyl alcohol, 64.7° ; water, 100° ; middle point, 82.35°.

I

Methyl alcohol in large excess

The first result is quite satisfactory, while the second is not, but it must be remembered that it is always difficult to separate the more volatile component of a mixture when present in relatively small quantity, and, in such a case, a second distillation is usually necessary. The first distillation was therefore continued until the temperature reached 100°, and the whole of the distillate, weighing 66.8 grams, was then redistilled and the double correction for loss by evaporation was applied.

The weight below the middle point was now 38.9, corrected 39.5, giving the percentage composition.

It will thus be seen that, by repeating the distillation, the result was as satisfactory as that given by a single distillation when the alcohol was in excess. Even without correcting for loss by evaporation the agreement is fairly good, but it is much improved by introducing the correction.