PROCEEDINGS

of the Tomsk Technological Institute

of Emperor Nicholas II

Vol. 14. 1909. No. 2.

III.

S. V. Lebedev,

ON THE RECTIFICATION OF ALCOHOL IN BATCH-OPERATED APPARATUSES.

1–33.

On the Rectification of Alcohol in Batch-Operated Apparatuses

At the beginning of 1904, one of the large plants of European Russia applied to the Director of the Tomsk Technological Institute, Prof. Zubashev, with a request to provide some guidance regarding methods of rectification with chemical purification of alcohol which, replacing charcoal filtration, would at the same time yield a known increase in the output of first-grade product. A sample of raw alcohol was also sent.

At the suggestion of Prof. Zubashev, in order to clarify this question, a series of experiments was carried out in the laboratory. These experiments led to results having a certain practical significance, and in the data thus obtained one may find some material for elucidating the chemical aspect of rectification—a question that until now has been almost completely undeveloped. For this reason the indicated experiments are of some interest and are therefore being made known, all the more so because they involved painstaking laboratory labor and such work at the plant as could be carried out only thanks to the kindness of the owners and managements of the plants where the experimental distillations were performed.

Since the question had been posed from a purely practical standpoint and, moreover, required as rapid an answer as possible, the task could not, of course, include a comprehensive consideration of the extremely complex question of rectification with preliminary chemical purification of raw spirit.

In accordance with the conditions of the task set, all the work of finding the method sought had to be concentrated on improving the quality of the raw material being processed.

Although this question has so far been little developed, nevertheless very many attempts have been made to resolve it. All meth—

The methods proposed for this purpose, despite their outward diversity, may be divided into two groups: one of them includes methods that purify the raw spirit by extracting, in one way or another, part of its impurities; the other includes methods that seek not to remove them, but merely to change their form, converting them from compounds that are difficult to separate by rectification into compounds easily separated by the latter.

The methods of the first category, which in most cases require special devices and apparatuses, are based, in most cases, on the ability of certain bodies either to retain part of the impurities of the raw material (charcoal filtration, Bang’s method), or to cause these impurities to separate from the spirit in one form or another (Traube’s method).

The basis of the methods of the second category is the action of various chemical compounds on the impurities of the spirit being purified, which, in accordance with the conditions set, must serve as the basis for solving the question under consideration.

From the point of view of the methods of the second category, for the chemical purification of raw spirit during rectification, a very large number of different chemical compounds have been applied at various times; but in most cases these are only practical trials with one substance or another, and accordingly the results of all this are reduced, for the most part, to an enumeration of the chemical compounds used for this purpose, the number of which, it must be said, is very great. There are very few systematic investigations in this field, owing to the difficulty of carrying them out—and there are absolutely no such investigations in which analyses were made along the way both of the initial material and of all the resulting products of rectification, in the total change of which one must seek the key to resolving the question of the action of the given chemical compound. To a certain extent an exception to this is a series of experiments carried out by the St. Petersburg Chemical Laboratory of the Ministry of Finance, aimed at replacing charcoal filtration by the action on purified spirit of potassium permanganate with alkalis.

But even the indicated experiments, judging by Professor Kucherov’s report,¹ are far from exhausting this complex question, especially since they are all concentrated on a single sample of spirit and therefore do not make it possible to generalize the results obtained.

The absence of any indications for determining the required quantity of reagents for each individual case leaves open the question of the further application of the indicated means.

Likewise, the question of the useful effect of these reagents in rectification remains rather unclear; this, of course, should be sought not only in diagrams constructed for only the 1st grade on the basis of measurements of oxidizability and tasting, but also in a comparative chemical analysis of the second and third grades of rectification, and in distillations both with reagents and without reagents.

Taking as the basis for resolving the question posed the action of chemical agents on the impurities of raw spirit, let us first consider the qualities of the latter.

In order to give a certain idea in this regard, I cite, for raw spirits, the data of the Central Chemical Laboratories of the Ministry of Finance in the city of Odessa (Table I) and St. Petersburg¹ (Table II), as well as analyses carried out in the laboratory of nutrients of the Tomsk Technological Institute for 1905 (Table III).

From the indicated material one may conclude what great variety, both quantitatively and qualitatively, may be encountered in the impurities of raw spirits, which depends on the diversity of the material being processed and the totality of the conditions of factory operation.

During rectification, some impurities remain in the still of the apparatus; others pass unchanged into the distillate; still others, during distillation, undergo certain changes, interacting with one another and with the spirit being distilled. On the other hand, the spirit itself, in the course of distillation, is partly oxidized by the oxygen of the air, which is always present at the beginning of the distillation both in the apparatus and in the spirit being rectified, where it is in the dissolved state. The oxidation products act on the unoxidized spirit, resulting in new compounds which pass, as impurities, into the distillate. In the processes of formation of new compounds that pass into the impurities of the rectified spirit, it is very probable that a certain role is also played by the overheating of the spirit near the tubes of the steam coil of the still of the rectifying apparatus; in connection with the high temperature of the heating steam and the layer of porous scale covering these tubes, this creates very favorable conditions—

Table 1.

Results of analyses made in the laboratories of the Excise District and the Central Chemical Laboratory of the Ministry of Finance in the city of Odessa of raw spirits received in Kazan in 1900–1903.

Table II.

Results of investigations of raw spirits carried out by the Central Chemical Laboratory of the Ministry of Finance in Petersburg:

favorable for various kinds of reactions connected with the transformation both of the alcohol itself and of the impurities present in it into new, diverse chemical compounds.

These processes are promoted to a certain extent by the long duration of the run and by the enormous surface of the plates of the rectification column, which, of course, cannot be unconditionally clean. Apart from the action of any chemical reagents, rectification itself partly alters and partly separates the impurities of the raw spirit. In this, some of them are easily separable, while others are not. Accordingly, the action of chemical reagents must be directed toward the impurities that are difficult to separate.

The definition of “difficult to separate” may be interpreted in two different directions, according to the tasks of rectification, which seeks to obtain: first, as pure a product as possible, and second, as much first-grade product as possible. From the first point of view, the impurities of the raw spirit passing through must be included under the indicated definition.

Table III. Analyses of raw spirits from Siberian distilleries, performed by the nutrition laboratory of the Tomsk Technological Institute for 1905–1906.

all the fractions of rectification and therefore always fall, in greater or lesser quantities, into the 1st grade—such as acids and esters.

In addition to impurities that pass through the entire rectification, there are also, in crude spirit, those from which the greater middle part of the run is usually free. Nevertheless, some of these impurities, from the point of view of the possibility of increasing the yield of first grade, must likewise be regarded as difficult to separate. The reason for this is, on the one hand, that they are wholly inadmissible in first grade, even in minimal quantities; and, on the other hand, their inability to concentrate within the narrow limits of the initial or final fractions. Among such impurities, the most difficult to separate, owing to their comparatively very great extension over the fractions of rectification, are aldehydes, which pass through the first 13–22% of the rectification fractions. In order to give an idea of how the various impurities of the raw material are distributed over the fractions of rectification, I give the results of analyses of two experimental distillations carried out in 1905 at the plants of Mr. Vytnov and Mrs. Koroleva (the Porosensky plant), located near the city of Tomsk.

Both distillations were carried out with apparatuses of the Möller–Fugelsang system, of approximately identical dimensions. The charges, both in these distillations and in the remaining ones to be discussed below, fluctuated around 40,000°–45,000°, a normal value relative to the working capacity of the given apparatuses. It should be noted here that the apparatus at Mr. Vytnov’s plant was in complete order and operated very distinctly, which cannot be said of the Koroleva apparatus, already sufficiently old, requiring repair, and moreover with a column installed not quite correctly. Thus one of the distillations was carried out under favorable conditions, the other under unfavorable conditions, which, from the standpoint of the experiments, is also of some interest, since it makes it possible to judge, to some extent, the influence of unfavorable distillation conditions on its results under the conditions of chemical purification, which will be discussed below.

In the tables IV–V presented, as in the other analyses, the numbers show the content of aldehydes, acids, and esters in milligrams per 1 liter of absolute alcohol.

Fusel oil was determined by Rose’s chloroform method and

Eng.-Techn. S. V. Lebedev.

given in volume % as absolute alcohol; aldehydes were determined with the aid of Gayon’s fuchsin–sulfite–acid reagents*) and a Duboscq colorimeter.

The content of aldehydes, acids, and esters in the spirit is expressed in the form of ethyl aldehyde, acetic acid, and acetic-ethyl ester.

From the tables IV–V presented, it is clear that the most difficult impurities to separate are esters, acids, and aldehydes. Accordingly, the choice of those or other chemical compounds that can serve the purposes of preliminary purification of raw spirit must be such that the action of the substances employed is directed chiefly toward aldehydes, acids, and esters.

Having outlined the direction in which the question posed must be resolved, it remains to consider the choice of the corresponding chemical compounds.—This choice, of course, cannot present any great difficulty, since it is indicated by the chemical nature of the impurities of the spirit themselves, against which the action of the sought agents must be directed.—Moreover, numerous attempts to resolve this question have long since established the undoubtedly useful action of alkalis and weak oxidizing agents, for example, \(KMnO_4\), potassium permanganate, to which Balling pointed as early as 1845.

Of course, potassium permanganate acts not only on aldehydes, but also on the spirit itself; but since, with weak oxidizers, the rate of their action on ethyl alcohol and aldehyde is very different, it is therefore natural that their action is directed chiefly toward the aldehyde, affecting the spirit itself comparatively weakly and, in part, the fusel oil.

The success of the course of the reaction in the indicated direction is influenced to a considerable degree by the conditions under which the reaction proceeds, such as the concentration of the spirit being purified, the quantity and concentration of the reagent solutions taken, the temperature at which the reaction occurs, etc.

) Note.* Gayon’s reagent is prepared as follows: to a solution of 1 g of fuchsin in a liter of water, 20 cc of acidic sodium sulfite of 30° Baumé are added; after an hour, 4.8 cc of sulfuric acid of specific gravity 1.84 are added. In testing, for each cubic centimeter of this reagent, two cubic centimeters of the spirit under examination, diluted to 40 degrees, are taken.—Comparison with standard solutions in quantitative determination is made after 20 minutes.

Table IV

Run No. 1 at the plant of Messrs. Vintnovy.

Table V

Run No. 5 at the plant of Koroleva.

Accordingly, one of the first tasks is to work out these conditions.—Of the oxidizing agents suitable for this purpose, potassium permanganate remains, as the most appropriate for the present case and as one that has to some extent been tested.

The action of this reagent proceeds smoothly and easily either in an acidic or in an alkaline medium. It is clear that in the case under consideration only alkali can be used. The latter is necessary, apart from the consideration just mentioned, as a means against the acids and esters always present in raw spirit.

Since, with the present state of fermentation technology, every raw spirit always contains a certain quantity of acids and esters, alkali is therefore, from this point of view, a means applicable to all raw spirits—of course in the proper quantity and under the proper conditions, varying depending on the quality of the spirit being purified.—Thus, Professor Verigo’s remark¹ that caustic alkalis, like other chemical reagents², do not constitute a universal means applicable to all raw spirits, since for some this may be useful and for others harmful, must be understood in the sense that alkali is not a universal means only if there exists some “universal recipe,” according to which alkali is used for every spirit without regard to its properties. Consequently, one may reject “universal recipes,” but not the universality, in the present case, of alkali—at least with the qualities of raw materials that are possible at the present time.

The action of caustic and carbonated alkalis consists chiefly in the following: free acids are neutralized, esters are saponified, with free alcohol being liberated, and aldehydes partly resinify and partly polymerize.

Under the influence of caustic alkalis, a certain quantity of alcoholate is formed, which, very probably, under the influence of the circumstances already mentioned earlier, decomposes toward the end of the distillation.

The decomposition products, entering into interaction with ethyl alcohol and with various compounds of fusel oil that are present during this period in especially large relative quantity, give rise to new chemical compounds. As a result

Note: ¹ Proceedings of the Technical Committee, vol. XV, p. 56.

of these decomposition and interaction processes appear in the final fractions of rectification in a significantly increased amount, in comparison with the preceding fractions, characteristic impurities of the initial low wines, namely: aldehydes, esters, and acids. Owing to this circumstance, caustic alkali, which is desirable as being more energetic in comparison with carbonated alkali, at the same time proves dangerous when present in considerable excess. But since, on the one hand, the reaction with potassium permanganate requires an alkaline medium, i.e. a certain excess of free alkali, and on the other hand the same requirement is imposed by the reaction of saponification of esters and by the process of resinification of aldehydes, the necessary quantity of alkali must therefore be used in the form of carbonate, giving approximately only \(1/5\) of it in the form of caustic soda.

Laboratory experiments showed that even a very considerable excess of carbonated alkali does not exert any noticeable negative influence on the results of distillation, whereas a significantly smaller excess of caustic soda causes certain undesirable phenomena, of which more will be said below.

The content of acids and esters in raw spirits rarely exceeds such an amount as is equivalent, calculated per 1 liter of absolute alcohol, to 0.4 g of chemically pure caustic soda. Accordingly, the total quantity of alkali is taken as twice the indicated figure, which for a charge of 1000 will amount to approximately 100 g of chemically pure caustic soda. In this, \(4/5\) is given in the form of sodium carbonate and only \(1/5\) in the form of caustic soda. Of course, the indicated figure may vary in individual cases. The conditions under which the alkali is to be charged are determined by the work it is to perform: neutralization of free acids (soda), saponification of esters, and resinification of the aldehydes present in the spirit (NaOH). Resinification of the aldehydes present in the spirit after the action of chameleon upon it (NaOH). Accordingly, all the alkali is charged in three portions. In the first charge, to the spirit diluted with the amount of water necessary to bring it to \(45\text{–}50^\circ\), from which, judging by the charge, 60–100 buckets of water have been discharged, as needed for dissolving the potassium permanganate and other reagents, there is added, with vigorous stirring in a thin stream, a solution of the entire quantity of soda required for the given charge.

After 20–30 min., a solution of \(1/3\) of the caustic soda is added in the same manner; the remaining \(2/3\) of it is charged 2–3 hours after the addition to the spirit of the potassium permanganate solution.

The action of chameleon ($\mathrm{KMnO_4}$) has already been described in its main features above; accordingly, the conditions for adding it and for its action on the alcohol being purified should be as follows: the weakest possible concentration of the reagent solution, a low temperature, an alkaline medium, and the fullest possible contact of the reagent with the alcohol being purified. Since under such conditions the reaction proceeds not especially rapidly, a sufficient interval of time must therefore be allowed for its completion; this, of course, presents no difficulty, since with the usual availability of factory vessels and the normal course of work one can always set aside for this purpose a period of 16–20 hours.

In accordance with these requirements, the treatment of the raw spirit with chameleon was carried out as follows: a solution of potassium permanganate with a concentration of 0.1–0.05% was gradually poured from a special small tank with a tap, over the course of 15–20 minutes, into the dilution, which had already received, 2 hours earlier, the entire required quantity of soda and $\frac{1}{3}$ of the caustic soda. During the pouring-in of the chameleon solution, and likewise thereafter for 20–30 minutes, the alcohol in the dilution is vigorously stirred either with a special agitator or by hand with metal stirrers.

As a result of the action of chameleon on the alcohol, there is obtained a considerable decrease, but not a complete removal, of the aldehyde, some quantity of which must always be present in the alcohol purified in this manner, as the result of the action of potassium permanganate on the ethyl alcohol itself.

In order to reduce this residue as well, after 2–3 hours, or better shortly before pumping the dilution into the still of the apparatus, the remaining $\frac{2}{3}$ of the caustic soda allotted for the given charge is added; this addition is carried out as in the preceding case.

For the complete interaction and completion of the reactions between the alcohol being purified and the added reagents, after the dilution has been pumped into the still of the apparatus and it has been heated to $50$–$60^\circ\mathrm{R}$, the alcohol is kept at this temperature for 30 minutes, after which a sample is taken and filtered through an ordinary filter. If, in doing so, the filtrate proves to be entirely colorless, distillation is begun in the usual manner; if even the slightest signs of a yellow tint are present in the filtrate, the temperature should be raised by another 5 degrees and the alcohol left in the still at that temperature until the yellow coloration in the filtrate has completely disappeared.

Having thus outlined the necessary reagents and clarified, to a certain extent, on the basis of general considerations, the conditions under which

...which these reagents must act upon, let us now turn to establishing the necessary quantity of potassium permanganate.

As is already clear from what has been said, its action appears to be rather complex, directed in various ways; owing to this, a simple arithmetical calculation on the basis of one reaction or another is extremely difficult and hardly possible, especially if one takes into account the parallel action of alkalis and the variety of impurities in raw spirits. It is impossible, for example, to place the aldehyde content in the spirit being purified in direct dependence on the necessary quantity of potassium permanganate, as is evident from Table VI.

The data furnished to us by the ordinary analysis of spirit are so general, embracing at once whole groups of diverse compounds, consisting, in essence, of a whole series of different chemical representatives, very possibly related quite differently to certain chemical compounds, that it is not possible to put these data into direct arithmetical dependence on the quantity of reagents acting—as Table VI shows quite clearly, with the analyses of raw spirits for which the most advantageous required quantity of chameleon was established.

From these data one can only conclude that, with respect to potassium permanganate, aldehydes and, in part, fusel oils apparently have the chief significance, both quantitatively and qualitatively. In fact, spirits Nos. 1 and 7, and also Nos. 8 and 10, almost identical (in pairs) according to the analytical data, prove to be very different with respect to \(KMnO_4\).

On the other hand, spirits Nos. 1, 3, 5, 6, 9 show that, with the same relation to chameleon, the content of fusel oil and aldehydes may in each separate case fluctuate very strongly. The content of acids and esters in raw spirits apparently has very little significance with respect to the reagent indicated. Thus, the question of the quantity of \(KMnO_4\) required for each individual case has to be solved by some method based on the equivalent action of all factors connected with the oxidation process.

Naturally, in this connection, the principle of the Barbet reaction suggests itself first of all—the action of chameleon in the cold; but the matter is that this reaction, having in view chiefly the minimum quantities of substances comparatively easily and rapidly oxi-

Table VI.

On the Rectification of Alcohol in Batch-Operated Apparatuses

...arising and encountered in rectified spirits. When, however, one has to deal with crude spirits, and moreover under quite special conditions in comparison with those recommended for Barbet’s reaction, the matter is not limited to these compounds alone: apparently, a certain portion of the fusel oil is also included, and, to a certain extent, the alcohol itself.

In addition, the accompanying action of alkalis, the high temperature, and the duration of heating—all this compels one to turn to such a method for determining the quantity of potassium permanganate in which, as far as possible, at least the especially characteristic features of factory distillation would be observed, namely high temperature and a considerable duration of heating. Since, in the question under consideration, it is not only the absolute quantity of impurities in the entire rectified product obtained that is of importance, but the conditions of their distribution among the fractions also play a considerable role, it therefore seemed that the desired method should include distillation with division of the run into fractions and known control over the distribution of at least the most harmful impurities of the alcohol—for example, aldehyde, which is particularly capable of being concentrated within narrow limits of the run.

Of course, it would have been simplest of all to solve the question by practical work at a factory with a given spirit, but in the present case it was not possible to have such conditions at one’s disposal. There was only a sample of crude spirit in hand; on the basis of its properties and peculiarities it was required to give the desired answer, and therefore it was necessary to seek other ways.

Proceeding from the considerations indicated, after several modifications, the following method was finally adopted for determining the necessary quantity of potassium permanganate: crude spirit, diluted accurately to 45°, is taken in an amount of 300 cc and distilled on a water bath in a one-liter flask with a Glinsky dephlegmator equipped with screens. The distillation is conducted extremely slowly, so that in 4–5 hours only \(1/3\) of the total alcohol taken is distilled over. The distillate is collected in graduated cylinders with stoppers, each of 10 cc capacity. Each sample is tested for aldehyde by means of metaphenylenediamine hydrochloride\(^1\), and in this way the length of the aldehyde fraction obtained is approximately determined, which is esta—

Note: \(^1\) The reagent must be perfectly white, and its solution must be prepared separately each time.

Eng.-Techn. S. V. Lebedev

is poured off with an accuracy of up to 1 cc in the subsequent distillations. After this one proceeds to determine the amount of potassium permanganate required for the given alcohol, used for this purpose in a solution with a concentration of 0.02%.

The determination is based on the consideration that, with the aid of a corresponding, strictly definite quantity of chameleon, the aldehyde content in the alcohol, and consequently also the aldehyde fraction obtained during distillation, can be reduced to a minimum, which immediately increases when the quantity of \(\mathrm{KMnO_4}\) taken deviates from the norm in one direction or the other. Of course, aldehyde is adopted here as the criterion not because it exhausts the whole essence of the process taking place, but only as a rather convenient indicator of a sort. The determination itself is carried out as follows: to the raw spirit, taken in an amount corresponding to 300 cc of alcohol of 45° strength, the necessary amount of alkali is added, then a known amount of potassium-permanganate solution, after which the volume is brought with distilled water to 300 cc. Next, everything is shaken vigorously and transferred to a distillation flask, where it is left at room temperature, tightly stoppered with a rubber stopper, for 12 hours. After this period has elapsed, it is heated with the condenser set for reflux to \(50^\circ R\) and is left at this temperature for 45 min. During this time the initially turbid liquid becomes completely clarified. Then distillation is carried out under exactly the same conditions as in testing the raw spirit. Next, the length of the aldehyde fraction is determined; with the correctly chosen amount of \(\mathrm{KMnO_4}\), for average raw spirits it fluctuates within the limits of 3 to 6 cc, and for poor ones—from 6 to 9 cc.

With some experience, being guided to a certain extent by the external signs of the reaction (the rate at which the pink coloration disappears, the clarification time, the length of the initial aldehyde fraction, etc.), it proves quite sufficient to make 3 distillations in order to establish the required norm accurately.

The results for the raw spirits tested in this way are given in Table VI, where the magnitude of the aldehyde fractions obtained in laboratory distillations of the raw spirits without reagents, with alkali, and with alkali together with chameleon is shown. From the named table it is evident that the action of alkali alone, at least with respect to aldehydes, amounts to \(1/3\)–\(1/2\) of the total action of alkali with chameleon.

From these data, attention is also drawn to the circumstance that the absolute content of aldehyde in raw spirit still...

does not always settle the question of the length of the aldehyde fraction; at least under laboratory distillation conditions, for example, spirits Nos. 6, 8, and 10, with almost the same aldehyde content, prove to be very different with respect to the magnitude of the aldehyde fraction, as well as with respect to potassium permanganate. On the other hand, comparison of spirits No. 1 and No. 3 with spirits No. 2 and No. 4 makes it possible to observe a certain proportionality in the indicated respect. All the remaining spirits, however, are, in the sense under consideration, approximately the same as No. 1, so that No. 8 and No. 10 are something of an exception, the cause of which apparently must be sought both in the quality and in the quantity of their other impurities.

The amounts of potassium permanganate determined by the indicated method for the various spirits prove to be very precisely established; thus even very small deviations (0.1 zolotnik per \(1000^\circ\) of spirit) to one side or the other from the norm found already entail, as the tests at Mr. Vytnov’s plant showed, a noticeable decrease in the yield of first grade.

With regard to alkalis, on the basis of the data from the indicated tests one may add the following: their action proves best approximately at a double quantity relative to the theoretical amount, while a further increase gives no improvement.

Compared with carbonates, caustic alkalis give somewhat better results with respect to shortening the aldehyde fraction; but, on the other hand, when present in excess they cause certain undesirable phenomena during the subsequent course of the distillation, consisting in the fact that, after the complete disappearance of aldehyde in the fractions, traces of it again begin to appear after a short interval. The phenomenon disappears when caustic sodium is replaced by sodium carbonate.

This was especially clearly noticeable in spirits rich in fusel oil, namely Nos. 3, 8, and 10 (Table VI).

This circumstance compelled caution in the use of NaOH; and since it nevertheless has its advantages, only \(1/5\) of the alkali used is therefore assigned to it, while the remaining \(4/5\) of the required alkali is given in the form of sodium carbonate.—Having limited myself to what has been said concerning the laboratory work, I turn to the factory trial distillations.

In all, 8 were carried out, but 2 of them, owing to a misunderstanding that occurred in the taking of samples, cannot be taken into consideration.

Of the remaining ones, 4 were carried out at the Vytnovs’ plant and 2 at the plant of Mrs. Koroleva, located near the city of Tomsk.

The quantity of reagents taken (calculated as chemically pure), the sampling carried out during rectification, and the quality both of the spirit being processed and of the products obtained are shown in Table VII.

Analyses of the spirit being rectified, samples of which were taken from the receiver in each charge, indicate large fluctuations in its quality, especially in the runs at the Vypinov plant. These fluctuations arose because of an unexpected stoppage, just before the experiments, of the work of the rectification department, and because of the accumulation during this period of raw spirit continuously arriving from the distillery, which continued to operate without interruption throughout this time.

In addition, the taking of each fraction separately, unusual for the plant, caused—especially when two rectification apparatuses were operating—some difficulty with the vessels, which, not having been adapted to these requirements, lacked the proper pipelines leading directly to the collecting vat.

The consequence of all this was an uneven addition of second-grade spirits to the raw material being distilled, which caused the noted considerable fluctuation in the quality of the spirit being processed in each individual case. Thus, in run No. 1 only raw spirit was introduced; in No. 2, in addition, there was a small quantity of second grade, consisting of a mixture of the initial and final corresponding fractions taken from another operating apparatus; into run No. 3 there entered a very large quantity of second initial spirit; in No. 4 an addition was made of second initial grade from run No. 3 and part of the mixed second grade from the neighboring apparatus. In this way there arose a great nonuniformity in the quality of the material being processed, with which, willy-nilly, it was necessary to reconcile oneself, since, when making use of the courtesy of a private rectification plant that was operating at full capacity and, moreover, connected with a distillery, it was of course difficult to count on complete convenience and special conditions desirable for carrying out the experiments.

Runs No. 5 and No. 6 were carried out at another plant, of course with a different spirit, as is evident from the analysis.

Although fluctuations in the quality of the material being processed significantly hinder the possibility of comparison and of drawing conclusions, nevertheless the material obtained still elucidates, to a certain degree, the question under consideration.

In order to make the results obtained more clear and comparable, Table VIII is proposed, where the content of impurities in each separate fraction is expressed as a percentage of their initial quantity in the spirit being processed.

Since, moreover, for evaluating the results of rectification, in addition to the quality and quantity of the first grade obtained and of the remaining fractions, the total amount of impurities carried by all the fractions together is also of significance, Table VIII therefore also gives the content of impurities in the sum of all the rectification fractions, expressed, as in the preceding case, as a percentage of the original amount of these impurities in the spirit being processed.

From Table VIII, judging by the figures in columns No. 1 and No. 5, which give the content of the remaining impurities in the sum of all the rectification fractions, it is evident that, under ordinary rectification conditions, the acids possess the greatest ease of separation; next comes fusel oil; the last place belongs to the ethers and aldehydes.

If, however, the ease of separating impurities is assessed by their ability to concentrate in a more or less short fraction, then the order obtained is different: under this condition, first place belongs to the fusel oils; next come the aldehydes; then follow the acids and ethers.

The addition of reagents under the adopted conditions apparently gives no improvement with respect to fusel oil, and in some cases, probably depending on the quality of the raw material being processed, may even cause a noticeable deterioration; thus, for example, judging from Table VIII, the total content of fusel oil in the sum of all the rectification fractions in runs without reagents proves to be less than with reagents.

It is quite understandable that, under the given conditions of addition and action of the chemical reagents, one can count very little on improvement with respect to fusel oil. Usually the basis of this process is the oxidation of some part of the latter, which proceeds considerably more difficultly than the oxidation of aldehydes and therefore, of course, requires quite different corresponding conditions as compared with those adopted.

In the case under consideration, however, the entire purpose was directed toward ensuring that the action of the chemical reagents, and especially of potassium permanganate, would be directed chiefly toward the aldehydes, these most readily oxidized compounds; it is therefore entirely natural that the fusel oil remains under these conditions almost untouched or is even changed in an undesirable direction. The latter circumstance must be connected with the action of an excess of alkalis and, of course, with the known chemical nature of the fusel oil of the given spirit.

TABLE

Table VII.

2.

—
119,0 gr.
—

3.

8,54 gr.
23,1 gr.
122,7 gr.

5.

—
—
—

6.

11,52 gr.
23,1 gr.
122,7 gr.

TABLE

Evidently the cause of this phenomenon lies in the action of alkalis on the ether compounds contained in raw spirit, which, having a very high boiling temperature, usually remain in the still of the rectifying apparatus; under the action of alkalis, however, they decompose into compounds with a lower boiling temperature, which, passing into the distillate, lower the quality of the latter in comparison with ordinary conditions.

The action of alkalis with respect to acids consists in neutralizing the latter and, thereby, reducing the amount of ethers obtained in the distillate.

Note. The sample of the 1st grade of distillation no. 6 was not delivered to the laboratory, and therefore only its strength was established (which was done at the plant); the remaining data have provisionally been taken from the data of the corresponding sample of distillation no. 5.

VIII

The decrease in the quantity of acids themselves in the rectification distillates must be observed to a very slight degree.

It should be noted that, in general, very little acid passes into the rectification fractions in comparison with the other impurities of the raw spirit, since a large part of them passes, during the distillation process, into ethers; therefore the fluctuation of their content in the rectification distillates may be very small.

From Tables VII and VIII it is very evident that, in all distillations, the action of the reagents with respect to ethers is manifested: their content in the sum of all fractions falls to \(1/3\)—\(1/2\) in comparison with the data for ordinary rectification.

A direct comparison of the figures in Table VIII, showing the content of esters in the sum of all rectification fractions, leads to the conclusion that the combination of all three reagents (runs No. 3 and No. 5) gives the greatest effect; next comes soda, while caustic soda occupies the last place.

If, however, one takes into account that the alcohol from run No. 4, judging by the analytical data, is a more favorable medium for the action of alkalis than the alcohol from run No. 2, then it must be considered that the advantage remains with caustic soda, which is also evident from a comparison of the ester content in the final fractions of the runs.

The considerable strengthening of the action of alkalis on esters in the presence of potassium permanganate must apparently be connected with the decrease, under these conditions, in the quantity of aldehydes, which in the present case have indirect significance as compounds capable, during rectification—especially at its beginning—of passing into acids, the consequence of which, of course, is the formation of esters.

The undesirable phenomena observed in laboratory runs in the presence of an excess of caustic soda were entirely unnoticeable under factory operating conditions.

For further clarification of this question, a factory run was carried out with a very considerable excess of caustic soda; but, unfortunately, for the reasons already indicated earlier, its results could not be used.

The chemical reagents under consideration have an especially favorable effect with respect to aldehydes.

Quantitatively, the results are close to those obtained with respect to esters; qualitatively, however, they are considerably higher, since, besides the large decrease in the quantity of aldehydes in the sum of all rectification fractions, there is also a considerable reduction of the large initial aldehyde fraction, which is not observed with respect to esters and acids. This conclusion follows from Tables VII and VIII, in addition to which, for greater clarity, Table IX is also introduced, where the content of aldehydes in the two initial rectification fractions is given, within whose limits the principal work of the chemical reagents with respect to aldehydes takes place.

The figures given in Table IX give the quantity of aldehydes carried off by an individual fraction, expressed as a percentage of the quantity of them that was contained in the raw spirit being processed.

In Table IX, for run No. 6, the fraction designated in Tables VII and VIII as “II a” is combined with fraction “II b.”

Table IX.

From the data presented it is evident that the greatest effect is obtained with the combined action of all the reagents. The action of caustic soda is approximately equal to one half of the combined action with potassium permanganate; the action of sodium carbonate is somewhat weaker than that of caustic soda, although it comes close to it. From the data presented it is also evident that the amount of aldehydes removed—this impurity being the most difficult to separate by means of ordinary rectification—increases greatly with the aid of the indicated chemical purification, although in factory operation to a considerably lesser degree than under laboratory conditions.

In addition to a reduction in the amount of aldehydes under the action of the reagents, judging from Tables VII and VIII, there is also observed a considerable change (toward an increase) in the concentration of this impurity in the initial fraction of rectification. This may be a consequence either of intensified separation of aldehyde under the new conditions during this period of distillation, or it may be the result of a reduction of this fraction at the expense of a part of it poor in aldehydes, so that, with the same or even a smaller absolute content of aldehyde in the entire fraction, its percentage content increases.

The latter explanation is apparently the one that applies here. In part, the data of Table VII, and chiefly the laboratory distillations of raw alcohol, which establish the length of the initial fraction,

show that the aldehyde content, enormous at the beginning of this fraction, rapidly falls to a comparatively small percentage, which then continues through a rather long interval. Judging from laboratory distillations, the entire aldehyde fraction may be divided into 3 parts: in the first, very short one, the principal mass of aldehyde is concentrated; then follows a long fraction with a comparatively small aldehyde content; after which, in the last part, whose magnitude occupies a middle position between the first two, the aldehyde content falls to zero.

Of course, the same character of distribution remains also in plant distillation.—Since the transition from the IIIrd initial grade to the IInd grade is made on the basis of Saval’s reaction, in which, besides aldehyde, other impurities of the raw material also have significance, the amount of which, very possibly, is reduced by chemical purification, it is therefore natural that the transition to the IInd initial grade in this connection occurs somewhat earlier than under ordinary conditions; this, probably, also explains the noted increase in the percentage content in the initial III grade.

As regards the qualities of the first grade obtained with the use of chemical reagents, then, judging by the data of the ordinary analysis of the alcohol (Table VII), its quality is approximately the same as with ordinary rectification, only the yield is greater.

In making a comparative assessment of the 1st grade obtained with chemical purification and without it by means of Prof. Verigo’s reaction (the salicylic aldehyde test) and Barbet’s reaction (measurement of the oxidizability of alcohol by a weak solution of chameleon), one must even note a certain advantage of the rectified alcohol with preliminary chemical purification over ordinary rectified alcohol.

The reason for this lies, as Prof. Kucherov believes, in the oxidation of unsaturated compounds,¹ with which rectification copes poorly, and which, without the assistance of potassium permanganate, pass into the 1st grade, thereby lowering its quality.

In order to ensure, as far as possible, the uniformity of analogous rectification fractions in various distillations, in addition to such signs as the appearance, taste, and odor of the distillate at the transitions from one fraction to another, there was also applied Saval’s qualitative test with strong sulfuric acid (10:10) and Prof. Verigo’s reaction with a one-percent solution in pure ethyl alcohol (96°) of freshly distilled (b.p. 196° C.) salicylic aldehyde.

Note. ¹) Trudy Tekhn. Komiteta—vol. XIV; p. 120.

The reaction is carried out as follows: to 10 cc of the alcohol being tested, 25–30 drops of the indicated solution of salicylic aldehyde are added, the mixture is shaken, and then, carefully along the wall of the flask, 20 cc of strong, chemically pure sulfuric acid is poured in, after which everything is again thoroughly shaken and left for 15–20 min.

If, after this time has elapsed, the mixture assumes only a canary-yellow color, then the given alcohol is free from fusel oil and aldehyde; but in the case of even the most insignificant quantities of these impurities, a reddish tint appears in reflected light, and such a content of these impurities as in raw spirits or in the second grades of rectification is sufficient to produce at once a very dark coloration.—The transition from the third preliminary fraction to the second preliminary fraction is made when, in the Savalle test (10:10), on pouring the alcohol together with the acid, no coloration of the mixture is obtained without boiling.

The transition from the second preliminary fraction to first grade is made when the alcohol fully withstands the Savalle test and, in addition, when in the reaction of Prof. Verigo the noticeable pink coloration appears no earlier than 4–4½ min. after mixing with the acid.—A fraction satisfying the latter condition appears 3–5% (counting by the charge of the apparatus) later than the fraction that withstands the Savalle test.

Under these conditions and, of course, with a correct transition to the final fractions of rectification, the whole of the 1st grade freely withstands the strictest excise tests.

The relative quality of the transition fractions to first grade can be judged from the data of Table X. These data refer to distillations carried out at various plants in order to determine the possible maximum withdrawal of 1st grade.

Table X¹)

Note¹): Borrowed from the reports of the excise inspector of the Tobolsk and Akmolinsk provinces.

From the foregoing it is evident that, with the proper withdrawal of the first grade, the quality of its general sample and of the fractions adjacent to it differs very little from one another.

Having considered the action of the reagents separately with respect to the various impurities of raw spirit, we shall attempt to make a comparative evaluation of the results of rectification in the runs examined.

In practice it is usually customary to evaluate these results by taking, as the sole criterion, the yield of the first grade.

It is clear that such an evaluation cannot be correct, since the yield of the first grade is affected by the quality of the raw material being processed; and, for evaluating the results of rectification, the quality of the first grade obtained— which is not always the same—also has significance, as does the quality of the remaining rectification fractions. Accordingly, in a comparative evaluation of rectification results, it must be assumed that they are directly proportional to the withdrawal of the first grade, directly proportional to the amount of impurities separated from it relative to their content in the raw material, and inversely proportional to the amount of impurities remaining in the sum of all rectification fractions.

The quality of the raw material being processed is introduced here indirectly, by expressing the impurities as a percentage of their initial content in the raw material.

Thus, the rectification effect, for example, for aldehyde in run No. 1 will be expressed as follows:

\[ \frac{77.9 \times 100}{93.5} = 83.3 \]
(see Tables VII and VIII).

For acids in the same run:

\[ \frac{77.9 \times (100 - 10.5)}{21.5} = 324.3 \]

Table XI gives the numbers obtained in this way for all the runs performed. The greater the rectification effect, the larger the corresponding number.

Table XI.

Assuming that rectification without any auxiliary means gives the normal degree of purification, taken as 100, we shall express, on the basis of the data of Table XI, the results of distillations No. 2, No. 3, No. 4, and No. 6 as percentages relative to the results obtained in distillations No. 1 and No. 5.

Taking, moreover, as the normal conditions of distillation No. 1—both with respect to the quality of the raw material processed and with respect to the merits of the rectifying apparatus and the working conditions—we shall likewise express, as percentages of it, the results of distillations No. 5 and No. 6.

The following Table XII shows clearly that distillations No. 1, No. 2, and No. 3, though equivalent in yield of first grade, prove to be far from identical in the effect achieved in these cases by rectification. Distillation No. 6, with a yield of 74.5%, proves, in the effect it achieves, to be somewhat inferior to distillation No. 3, with a yield of 84.0% of first grade.

This same table gives a clear representation of the comparative action of the reagents, as was already mentioned earlier.

The numbers of Table XII, of course, cannot be regarded as applicable in each individual case.

Table XII.

It is self-evident that their magnitude is the resultant of the chemical nature of the impurities in the raw material being processed, the quality of the rectifying apparatus, and the work of the rectifier; nevertheless, the figures given still provide a certain idea of the relative action of the reagents and of their significance as factors that increase the effect of rectification.

At every plant the quality of the alcohol being processed is always subject to certain fluctuations; accordingly, the quantity of potassium permanganate required must, of course, also change—this especially important reagent, an excess of which, just as...

and a deficiency may reduce the yield of the first grade. Therefore, for factory practice it is very important always to know whether the amount taken is correct or not.

The laboratory test mentioned above is inapplicable under the conditions of factory operation, and therefore other methods have to be sought. A good criterion of the correctness of the amount of chameleon taken may be the totality of phenomena accompanying the action of this reagent on the alcohol being purified. The characteristic signs are: the rate at which the pink coloration disappears after the addition of chameleon; the color of the filtrate of a sample taken from the still of the apparatus after preliminary heating of the latter; and the characteristic sharp odor in the initial fractions when the reagent is in excess.

The slightest excess of the chameleon used is immediately revealed by a yellowish tint of the filtrate of a sample taken from the still of the apparatus after preliminary heating of the alcohol in it. In essence, the matter here is the reaction proposed by Liebig and Bang for detecting aldehydes by the action on them of an excess of alkali.

A deficiency of potassium permanganate is indicated by the comparatively rapid disappearance of the pink coloration after its solution has been poured into the chameleon alcohol.

As for the alkali, on the basis of the data available so far, observations in this respect should be reduced to checking, by means of litmus paper, the reaction of the contents of the still of the rectifying apparatus, which from the beginning to the end of the run must be distinctly alkaline.

Of course, it would be desirable to dwell in more detail on determining the optimum in this respect, since it may be thought that too great an excess of alkali also has its inconveniences; but for the present this is a question for further work in the same direction.

When comparing the results of laboratory runs with the addition of chemical reagents with factory runs under the same conditions, it turns out that the effect obtained with respect to aldehydes is extremely different in the one case and in the other.

Under factory conditions, judging from the data of Table VIII, the decrease in the amount of aldehyde in the sum of all the rectification fractions, with the use of alkalis alone, rises by 8–15% compared with the results of ordinary rectification and by 25–40% with the combined action of alkalis and chameleon.

In the laboratory distillations, no quantitative determination of aldehydes was carried out; nevertheless, one can form a clear idea on this point, judging from the change, upon the addition of reagents, in the magnitude of the initial aldehyde fraction (Table VI), and from the detailed qualitative tests of this latter fraction which accompanied each distillation. For alcohol of average quality, by the action of alkali alone, the initial aldehyde fraction is reduced by \(1/3\)—\(1/2\). It is by this action of the alkali that one may explain why the quantity of chameleon required proves in reality to be considerably smaller in comparison with that which would be required judging from the content of aldehydes alone in the crude alcohol, without even taking into account that the oxidizing action of this reagent, besides aldehydes, extends also to certain other impurities of the crude spirit and also, in part, to the ethyl alcohol itself.

By the combined action of alkalis with chameleon, the initial aldehyde fraction for alcohols of average quality is reduced, instead of the original length of approximately 25 c.c., to 3–6 c.c. At the same time, in the first 2–3 cubic centimeters the aldehyde content is equal to the content in the middle of the aldehyde fraction obtained in distillation without reagents, while the remaining 3–4 cubic centimeters contain the very minimum quantity of this impurity.

Thus, on the basis of what has been stated, it may be said with confidence that under laboratory conditions, by the action of alkali alone, the quantity of aldehydes in the distillate is in any case reduced by no more than 30–40% of their initial content, and by 10–15% under the combined action of alkalis and chameleon. Consequently, there is a sharp discrepancy between the results of laboratory and factory distillations. The latter cannot be explained by an incorrect quantity of chameleon taken in factory operation, since the slightest change of it in one direction or the other from the norm established for the given alcohol entails a deterioration in the rectification results obtained. Therefore the cause must be sought not in an incorrectly taken quantity of potassium permanganate and other reagents, but in the difference between the conditions of factory and laboratory distillations. This difference consists: first, in the long duration of the factory distillation, which in the present case, with respect to the aldehydes contained chiefly in the initial fractions of rectification, of course cannot be of great significance; second, in the very large volume of the rectifying apparatus in relation to the mass of alcohol being distilled, which

is not present to such a degree in laboratory distillations and, finally, in the peculiarities of factory distillation connected with the operation of the many-plate high column of the rectifying apparatus, which creates entirely distinctive conditions inaccessible to laboratory distillations.

With the great gradualness in the filling of the whole apparatus with alcohol vapors, and the considerable interval from the beginning of heating the still until the appearance of distillate in the sight glass, there occurs, of course, at the expense of the air present in the apparatus and dissolved in the liquid of the still, partial oxidation of the alcohol and of the aldehyde remaining in the alcohol after the action upon it of chemical reagents.

As a result, a new quantity of aldehyde is obtained, the consequence of which, in comparison with laboratory distillations, is an increase in its content in the initial fractions of factory rectification.

Besides aldehydes, acids are obtained in a similar way; but no considerable increase in their quantity can be observed in the distillates of rectification, since, owing to favorable conditions, they for the most part pass into esters.

Thus, during the very process of rectification, new quantities of various impurities of alcohol are obtained, which for the most part pass into the distillate, in part are carried along by the alcohol condensing in the still of the apparatus, and in part remain for a very long time in the column, thus escaping the action upon them of chemical reagents and at the same time serving as a source for the formation of new compounds.

The great difficulty of isolating the impurities of the initial fractions by means of periodic rectification is indicated by the experiments of Sorel and Barbet. The latter circumstance prompted Guillaume to arrange, in his continuously operating apparatus, a special device for the concentration and removal of these impurities, which in periodically operating rectifying apparatuses can be achieved to a certain degree by the use of preliminary chemical purification of the raw spirit being processed.

On the basis of all that has been said, one may arrive at the following conclusions concerning chemical purification during rectification in periodically operating apparatuses.

a) The use of chemical reagents for the purposes of rectification may, under certain conditions, be an important factor with respect to increasing the yield of first grade and improving its quality.

c) The development of any universal “recipe” for the chemical purification of alcohol, both with respect to the quality and quantity of the reagents used and to the manner of their employment, cannot take place, owing to the great diversity in the content and chemical nature of the impurities in raw spirits.

d) Owing to the constant presence, among the impurities of raw spirits, of aldehydes, acids, and esters, alkalis and mild oxidizing agents are universal means; and therefore, for each individual case, the question is only one of their quantities and conditions of action.

In addition to applying preliminary chemical purification to raw spirits, investigations were begun in the laboratory regarding its extension also to the reprocessing of the third products of rectification; moreover, besides the reagents mentioned, certain other chemical compounds applicable to this case were tested as well. However, the results of these works, being unfinished, cannot be presented here, since for certain reasons the work on the question indicated had to be discontinued for a fairly long period before its completion.

In conclusion, I consider it my pleasant duty to express my gratitude to the owners and administrations of the factories of Messrs. Vytnov and Mrs. Koroleva, who kindly permitted the distillation experiments described to be carried out at their premises.

S. V. Lebedev.

Tomsk.

Laboratory of Nutrients
Tomsk Technological Institute.

1. Journal Vinokurenie, 1906, No. 5, p. 3; No. 7, p. 2. ↩↩↩↩

2. Also known are sulfuric acid, chameleon, chloride of lime, etc. ↩