V. P. Markov.

Senior Assistant of the Siberian Technological Institute.

Distillation of Acetic Acid Solutions

The question of the distillation of acetic acid solutions has more than once attracted the attention of investigators. The earliest work, that of Roscoe¹), is of only historical interest, since its results do not lie on a smooth curve, which must be attributed to insufficient protection of the liquid being distilled against partial condensation due to heat loss through radiation.

In the next work, in chronological order, by Rayleigh²), the indicated error was eliminated. With the aid of a special steam jacket the vapors were subjected to a certain superheating, and partial condensation of the distillate was not observed. Unfortunately, there is another weak point in the work. The collected samples of distillate were relatively very large (50 cc). During their distillation the composition of the liquid in the flask had time to change appreciably; and since the composition of the vapor and the composition of the liquid are not in a linear dependence on one another, there is no basis for considering the average compositions of the distillate and of the liquid in the flask to be in equilibrium. The influence of this factor may be especially marked at high concentrations, at which, during boiling, the composition changes particularly rapidly.

The postwar period includes the works of Pascal and Mariller. I was unable to obtain the original articles, but in the latter’s work on the dry distillation of wood³) there is a table and a diagram giving the final results of the determinations both of Mariller himself and of Pascal. For low and medium concentrations they agree well both with one another and with Rayleigh’s data; at higher ones \((>70\% \mathrm{CH_3COOH})\), however, there is a considerable discrepancy. The curve of vapor composition according to Mariller’s data has a noticeable downward bend (in comparison with Rayleigh’s data), whereas the curve according to Pascal’s data deviates upward.

The absence of precise data on this question, which is of great importance for the calculation of rectifying columns for acetic acid, led me, in 1919–1920, when Mariller’s work was not yet known in Russia, to undertake an investigation in this direction. The work was carried out in the laboratory for the investigation of the dry distillation of wood under the Tomsk Provincial Council of National Economy, under the general direction of Prof. G. G. Povarnin.

Only a preliminary report on this work was published in the Journal of the Russian Physico-Chemical Society.

The abnormal working conditions in 1919–20, when the temperature in the laboratory fluctuated from \(4\) to \(18^\circ\mathrm{C}\), and in midwinter even below zero, which forced the work to be interrupted until summer, were the reason why I was able to begin checking the observations only in the summer of 1925.

As in 1919–1920, the work was carried out by Rayleigh’s method, modified only in that the collected samples of distillate were so small that during their distillation no appreciable change could occur in the composition of the boiling liquid.

¹) Roscoe, Ann. 125. 325.
²) Rayleigh Phil. Mag. 1902, 528.
³) Ch. Mariller. La carbonisation des bois, lignites et tourbes. Paris 192[[unclear: final digit]].

A flask, of about 1 liter capacity, with 600 cc of an acetic-acid solution, was heated on a wire gauze. To prevent condensation of vapor, the upper part of the flask was surrounded by a copper ring and a cup with an opening for the neck of the flask to pass through. By raising or lowering the ring with the cup, it was possible to intercept, to a greater or lesser degree, the stream of hot combustion products which, washing the upper part of the flask, could not only prevent condensation but, if desired, superheat the vapor.

A checking experiment with boiling water showed that superheating of 3–5° could easily be obtained.

The first series of experiments was carried out with superheated vapor. The steam-delivery tube remained completely dry; on the walls of the flask likewise no streams of liquid were observed, except those arising from splashes due to vigorous boiling.

At the beginning about 50 cc of liquid was distilled off, until the apparatus had become fully heated; then a sample of distillate, about 3 cc, was collected in a bottle with a ground-glass stopper (the volume of the bottle was about 15 cc). At the end of the collection the burner was extinguished and, simultaneously, a new receiver was put in place, into which the last portions of the distillate (2–3 cc) flowed, running back into the flask.

From the distillate sample taken, two samples of 0.6–1.0 gram of liquid were weighed into stoppered flasks of about 60 cc capacity, and the content of acetic acid was determined by titration with a Ba(OH)₂ solution, the concentration of which was about 0.3 N. Phenolphthalein served as the indicator. In the case of insufficient agreement of the results, a third sample was also taken.

In exactly the same way the concentration of acetic acid in the residue from the distillation was determined. Since its weight was about 600 grams, the change in composition during the collection of the distillate under investigation, in view of the small quantity of the latter, could be considered negligible.

After completion of an experiment, a certain quantity of water or of a weaker acetic-acid solution was added to the flask, with the calculation that the acetic-acid content in the distillation flask should be, at high concentrations, 3–5%, and at medium concentrations, 10–12%, lower than in the preceding experiment, and the distillation was again carried out.

For checking the data obtained, a second series of experiments was carried out under somewhat modified conditions.

The heating of the vapor was regulated so that neither condensation of vapors before their entry into the condenser nor superheating was observed.

The sign that this had been achieved was the presence, on the bulb of the thermometer, of a drop of liquid which remained unchanged for several minutes, and the absence of streams of liquid in the flask.

At the moments of the beginning and end of collecting the sample, the temperature of the vapor was noted. Thus here, although rather roughly (the thermometer had divisions of 0.2°), the boiling temperature was also determined.

In all other respects the work did not differ from the preceding series of experiments.

Then an additional series of experiments was carried out for high concentrations; the work was conducted as in the first series.

The results of all three series of observations, presented in the appended tables and diagram, agree excellently with one another and lie on a smooth curve.

The theoretical derivation of a formula expressing the dependence between the composition of the boiling solution and the composition of the distillate is greatly complicated by the fact that, owing to the difference in boiling temperature, there changes both

the number of associated molecules of acetic acid, as well as the degree of its decomposition into ions, and also by the existence of its various hydrates. But, on the other hand, it proved comparatively easy to find an empirical formula that agrees excellently with the experimental data, namely:

\[ P d = 0.725 p + 0.275 p^4 \]

where:

\[ P d = \text{concentration of } \mathrm{CH_3COOH} \text{ in the distillate} \]

\[ p = \text{concentration of } \mathrm{CH_3COOH} \text{ in the boiling liquid.} \]

The values of \(Pd\) for various concentrations are given below in the tables, and on the diagram with the experimental results the corresponding curve has been drawn.

Table of observations of the distillation of acetic-acid solutions

I. With superheating of the vapor.

II. Without overheating.

Composition of the distillate of acetic-acid solutions, calculated by the formula

\[ P_d = 0{,}725\,p + 0{,}275\,p^4 \]

To the article by V. P. Markov.

DIAGRAM

of the results of experiments on the distillation of acetic acid solutions.

Vertical axis: Distillate
Horizontal axis: Boiling liquid