Fourth Method

In the course of his experiments with a still-head kept at a constant temperature in a bath of liquid (p. 149), Brown observed that the composition of the distillate was independent of that of the mixture in the still, and depended only on the temperature of the still-head. Mixtures of carbon disulphide with carbon tetrachloride and with benzene were specially examined and, for these two pairs of liquids, Brown had previously determined the relations between the composition of liquid and vapour, and also the relations between boiling point and composition of liquid by Method I. From the curves representing these relations he read: 1. The composition of the mixtures which would evolve vapour of the same composition as the distillates obtained with the still-head of constant temperature; temperature again rose to 77.1°, when No. 11 was put in its place. At 844° a new receiver, No. 13, was substituted for No. 11 and the temperature was allowed to rise to 91.7°. The last fraction from I. was then added to the residue in the flask and the distillate collected in No. 13 up to 91.7°, after which fractions were collected in No. 14 from 91.7° to 984° and in No. 15 from 984° to 101.5°, when the distillation was stopped and the residue was poured into No. 17.

2. The boiling points of these mixtures.

He found that these boiling points, read from the curves, agreed closely with the temperatures of the still-head. The actual numbers are given in Table 20.

Table 20

In the last four determinations with carbon disulphide and carbon tetrachloride, the distillation was continued until not more than an occasional drop of distillate fell; this may probably account for the larger differences between the two temperatures.

The agreement is sufficiently close with both pairs of liquids to allow of the statement being made that the temperature of the still-head is, approximately at any rate, the same as the boiling point of a mixture which, when distilled as in Method I, would give a distillate of the same composition as that actually collected.

Special experiments further to test the truth of this statement were made with mixtures of ethyl alcohol and water, but in this case the comparison was made by determinin :1. The composition of the distillates when the still-head was kept at 81.8° and 86.5° respectively;

2. The composition of the distillate from mixtures which boiled at the same two temperatures. The results obtained are shown in Table 21.

Table 21

It will be seen that the agreement is very satisfactory, and it is evident that this method may be used to determine the relation between the composition of liquid and of vapour if the boiling point-composition curve has previously been constructed. Further experiments to test its applicability and the accuracy attainable have been carried out by Rosanoff, Lamb and Breithut,1 and Rosanoff and Bacon.2 The authors find that a double- walled metallic cylinder, open at both ends and immersed in liquid kept at constant temperature and vigorously stirred, gives better results than the ordinary spiral form of still-head.

The annular space a a, Fig. 31, between the two walls of the cylinder is bounded by a very large condensing surface ; the greatest height of the cylinder is 76 cm., the smallest height 70 cm. ; the mean diameter is 25 cm., and the width of the annular space is 0.95 cm.

A mixture of carbon disulphide and carbon tetrachloride containing at first 26 per cent of the sulphide was distilled through this still-head, the temperature of which was 59.82° throughout the experiment. The distillate was collected in nine fractions,which contained respectively 60.8, 60.7, 60.7, 60.5, 60.7, 60.5, 60.7, 60.7 and 60.7 per cent of carbon disulphide, mean 60.67 per cent. On the other hand the distillate from a mixture which boiled at 59.82° contained 60.8 per cent of the sulphide. The two values are in very good agreement.

Rosanoff, Schulze and Dunphy 3 further investigated the distillation of a ternary mixture through the same still-head. They arrived at the following general conclusions, based on Brown's results and those of Rosanoff and his co-workers. In distillations with a still-head maintained at a constant temperature, the composition of the distillate is at every instant identical with that of the vapour evolved by a mixture whose boiling point equals the temperature of the still-head. If the mixture is binary, the composition of the distillate is, in the course of a single distillation, constant. In those cases in which the binary boiling point curve passes through a maximum or a minimum, the composition of the distillate, although constant during the distillation, depends on that of the mixture originally placed in the still: it may have either of two values according to whether the mixture in the still is richer or poorer in one component than the mixture of constant boiling point. If the number of substances in the mixture is three or more, the composition of the distillate not only depends on that of the original mixture, but varies in the course of a single distillation. This variation, however, is moderate, and the nearer the constant temperature of the still-head is to the boiling point of the most volatile component, the more nearly constant is the composition of the distillate.

1 Loc. cit.

2 " Fractional Distillation with Regulated Still-heads," J. Amer. Chem. Soc, 1915, 37, 301.

3 " Fractional Distillation with Regulated Still-heads," J. Amer. Chem. Soc, 1915, 37, 1072.

For binary mixtures the method certainly possesses the following advantages: 1. The first part of the distillate may be altogether rejected, and the errors already referred to thus avoided ;

2. It is not necessary to determine the composition of the mixture in the still.

On the other hand, this method would not be suitable for mixtures the boiling points of which vary very slightly with change of composition, such, for example, as mixtures of normal hexane and benzene containing, say, from 1 to 20 per cent of benzene.

Method employed by Carveth.1 - The relation between the boiling points and the composition of mixtures of the two liquids is first determined under constant pressure, and the curve is constructed. A distillation is then carried out in a special apparatus with the object of determining simultaneously the boiling points of the liquid and of the distillate. If this could be done satisfactorily the composition of the mixture in the still and of the distillate at different stages of the distillation could then be ascertained from the curve.

The apparatus is ingenious, but there are errors involved in the method which, although they partially compensate each other, leave the results somewhat doubtful.

Carveth's actual results differ somewhat widely from Brown's, and the conclusion arrived at by Rosanoff and Easley (loc. cit.), who compared their own results with those of Brown, Zawidski and Carveth, is that Carveth's method is not reliable.

In the opinion of the author the methods most strongly to be recommended are those of Rosanoff and his co-workers.

1 Carveth, "The Composition of Mixed Vapours," Journ. Phys. Chem., 1899, 3, 193.