Table 18. Result of Distillation

Table 19. Tabulation of Results

The combined weights of distillate were in each case plotted against the combined composition (Fig. 28) and curves were drawn through the points. Only the point corresponding to the first fraction was badly off the curve which, however, was well defined by the remaining points. Possibly there was a trace of moisture in the mixture which would interfere with the first distillate.

Extrapolating the curves to zero weight of distillate in each case, the following results are obtained for the composition of the vapour in contact with (a) the original mixture, (b) the residue :Molar percentage of Cs2 in Liquid. Vapour.

Fig. 28.

Original mixture . . 36.77 60.35

Residue . . . 22.18 44.85

The results of this and other distillations agreed well with those obtained by the method of Rosanoff and Easley (p. 72).

Sources Of Error

The following sources of error should be noted and guarded against as far as possible:

1. When a still is heated from below in the ordinary way, the vapour which is evolved before the temperature of the upper portion of the still has reached the boiling point of the liquid will be partially condensed, and the residual vapour will contain an excessive amount of the more volatile component. Partial fractionation will, in fact, go on, and the first portions of distillate will be too rich in the more volatile component.

2. If, while the distillation is progressing, the upper part of the still or delivery tube is exposed to the cooling action of the air, partial condensation of vapour will occur and a similar error to the first will be produced.

3. The air which is in the still before the liquid is heated will become saturated with vapour, and as it passes through the condenser part of this vapour will be condensed. In Brown's case, as already mentioned, this premature condensation of liquid was considerable in amount. The error introduced by the incomplete condensation of the vapour carried over by the air will partially compensate that referred to under No. 1, and, if the top of the apparatus is heated before the liquid in the still is boiled, may possibly more than counterbalance it.

4. If a thoroughly dehydrated liquid is exposed to the air even for a short time, especially if it is poured from one vessel to another so that a large surface of it is so exposed, a certain amount of moisture will almost invariably be absorbed. It is not only liquids which are regarded as hygroscopic which thus absorb moisture, but even substances, like benzene or the paraffins, which are classed as non-miscible with water. As a matter of fact, it is probable that no two liquids are absolutely non-miscible, and certainly all commonly occurring liquid organic compounds can dissolve appreciable quantities of water.

When such a liquid as benzene or carbon tetrachloride, containing a minute amount of dissolved water, is heated, a mixture of minimum boiling point is first formed, and the first small portion of distillate will contain the whole or at least the greater part of the water, and may probably be turbid. Suppose now that a mixture of benzene and carbon tetrachloride is being examined and that a minute amount of water has been absorbed during the preparation of the mixture or in pouring it into the still. The first portion of distillate will then contain the greater part of this water and the specific gravity, refractive index, boiling point and other physical properties of the distillate will be appreciably altered. A considerable error may thus be introduced in the estimation of the composition of the first fraction.

Brown rejected the data derived from the first distillate, and that is probably the safest plan to adopt. Zawidski rejected the first small portion of distillate and avoided the first and second sources of error, partially at any rate, by heating the delivery tube in the manner described, but in spite of this, the first results in some of his series of experiments appear to be less accurate than the later ones. Lehfeldt guarded against the first and second sources of error by heating the whole of the upper part of his apparatus by means of an incandescent lamp, as described.

The fourth source of error is probably of greater influence than is generally recognised. All that can usually be done is to dehydrate the liquids, to dry the apparatus as completely as possible, and to keep the liquids, as far as possible, out of contact with moist air; but Rosanoff, Bacon and White avoided all the errors to which the first fraction is liable by the method just described.