B. V. Tronov and L. S. Nikonova.

Study of the Oxidation Products of Pyridine by Potassium Permanganate under Various Conditions

In a previous work¹) we studied the rate of oxidation of pyridine and of certain other cyclic amines by potassium permanganate in the presence of alkalis, acids, and simply in aqueous solution. In the oxidation of pyridine it is especially interesting that the amine, while reacting rapidly with permanganate in an acidic medium (and in general in the form of a salt), is oxidized very slowly by itself and still more slowly in the presence of alkali. Obviously the mechanism of the reaction must differ depending on the conditions. It has not yet been clarified whether the acceleration of oxidation in the presence of acids is explained by a greater saturation of the nitrogen atom (its transition to the pentavalent state), or whether the hydrogen appearing at nitrogen is of significance here. The latter explanation seems probable if one takes into account the regularities observed in the oxidation of alcohols²).

Thus, CH₃ CH₂ OH is readily oxidized by both KMnO₄ under a variety of conditions and by CrO₃ in aqueous solution, whereas CH₃ CH₂ O CH₂ CH₃, although rather readily subject to the action of alkaline permanganate, is oxidized considerably more difficultly by KMnO₄ in an acidic medium and almost does not react with CrO₃ at the concentration used in the experiments. Such a sharp difference can be explained only by the fact that, in the reaction with CrO₃ and partly with acidic KMnO₄, the hydroxyl hydrogen of the alcohol takes part first of all, whereas in the ether this hydrogen is replaced by a hydrocarbon radical.

To clarify the mechanism of the reaction, it was also of interest to study it from the qualitative side, to investigate the oxidation products formed under various conditions.

The material available in the literature concerns chiefly pyridine homologues. The simpler homologues, such as picolines, with KMnO₄ in alkaline solution give predominantly pyridinecarboxylic acids³); in acidic solution with KMnO₄, d-picoline is destroyed with formation of formic and acetic acids.

This, of course, is fully explained by the different stability of the pyridine ring in an acidic or alkaline medium.

There are two works on the oxidation of pyridine itself. M. Delépine⁴) treated pyridine with an excess of KMnO₄ in the presence of an equivalent amount of H₂SO₄ at 70°. He obtained CO₂, NH₃, and HNO₃; toward the end of the reaction free oxygen was evolved. J. Dorronsoro⁵) carried out oxidation with sodium persulfate in acidic solution, also on heating to 70°–80°, and isolated the same products, CO₂, NH₃, and HNO₃.

We repeated the experiment on the oxidation of pyridine by permanganate in the presence of sulfuric acid, but only with an equimolecular amount of oxidizing agent and at ordinary temperature.

In addition, experiments were carried out in alkaline solution and without the addition of acids or alkalis, both at room temperature and at

¹) B. V. Tronov and L. S. Nikonova, “Proceedings of the V Mendeleev Congress,” p. 87.
²) B. V. Tronov and A. A. Lukashkin, Journal of the Russian Physico-Chemical Society, chemical part, 59, 1157 (1927); 60, 181 (1928).
³) A. Ladenburg hiel, Ann. 247, 14 (1888).
⁴) M. Delépine, C. r. 184, 206–8 (1926).
⁵) J. Dorronsoro, Ann. Soc. Erp. Fis-Quim 25, 409–10 (1927), Chem. Zbl. 1928 I, 1192.

At \(70—80^\circ\), \(KMnO_4\) was taken everywhere (as were also \(KOH\) and \(H_2SO_4\) in the corresponding experiments) on the basis of one molecule per molecule of \(C_5H_5N\). Depending on the conditions, the reaction proceeded until the solution became colorless in from 3 days to 4 months.

The products of oxidation were investigated as follows: the precipitated manganese peroxide (sometimes containing potassium) was filtered off. The unreacted pyridine was distilled off, if necessary (in experiments in acid solution, after preliminary addition of alkali); its amount (together with the \(NH_3\) formed) was determined by titration with \(FeCl_3\) and \(H_2SO_4\) (by Schulze’s method¹). To determine the ammonia formed, a definite portion of the solution was taken, neutralized with \(HCl\), and concentrated on a water bath; then, by adding an excess of \(KOH\), \(NH_3\) was liberated and distilled with steam into water. The ammonia passed over in the first portions of the distillate, while the greater part of the pyridine still remained in solution. From the distillate the \(NH_3\) was precipitated with ethyl oxalate in the form of the almost insoluble oxamide

\[ \begin{array}{c} COOC_2H_5\\ |\\ COOC_2H_5 \end{array} + 2NH_3 \longrightarrow \begin{array}{c} CONH_2\\ |\\ CONH_2 \end{array} + 2C_2H_5OH \]

Preliminary tests showed that the precipitation of ammonia under such conditions proceeds sufficiently completely even from dilute solutions. Pyridine does not interfere with the determination.

In the aqueous solution, after the distillation of pyridine and \(NH_3\), \(CO_2\) was determined; it was liberated by the action of \(HCl\) and collected in a gasometer. In the remaining solution, oxalic acid was determined by the usual method, in the form of the calcium salt.

The nitric acid formed during oxidation was determined by Devarda’s method, reducing it (in a separate sample of the solution) to \(NH_3\).

The acid solution, after removal of \(CO_2\), was extracted repeatedly with ether, but in the ethereal extracts nothing could be detected except oxalic acid. The quantitative results of the experiments are given in the following table:

Table No. 1.
Products of the oxidation of pyridine by \(KMnO_4\) under various conditions.

| No. of experiment | Substances taken | Temperature | \multicolumn{6}{c}{Amounts of products in grams} |
|---:|---|---|---:|---:|---:|---:|---:|---:|
| | | | \(MnO_2\) (with impurities) | \(C_5H_5N\) (unreacted) | \(NH_3\) | \(HNO_3\) | \(CO_2\) | \(C_2H_2O_4\) |
| 1 | 79 g \(C_5H_5N\) + 158 g \(KMnO_4\) + 3 liters water | \(18^\circ—22^\circ\) | 125 | 49.33 | 0.0853 | 1.787 | 15.97 | none |
| 2 | Same | ” | — | 49.51 | 0.0859 | — | 14.38 | ” |
| 3 | 25 g \(C_5H_5N\) + 50 g \(KMnO_4\) + 1.5 liters water | \(70^\circ—80^\circ\) | — | 15.6 | none | 0.827 | 2.71 | 0.843 |
| 4 | 79 g \(C_5H_5N\) + 158 g \(KMnO_4\) + 56 g \(KOH\) + 3 liters water | \(18^\circ—22^\circ\) | — | 49.94 | ” | 1.760 | 21.8 | 20.28 |
| 5 | Same | ” | — | 54.67 | ” | — | 21.68 | 21.48 |
| 6 | 25 g \(C_5H_5N\) + 50 g \(KMnO_4\) + 17.72 g \(H_2SO_4\) + 1.5 liters water | \(70^\circ—80^\circ\) | 42 | 18.9 | ” | 1.557 | 5.778 | 6.661 |

In the following table the yield of the various oxidation products is calculated in percent relative to the amount of pyridine that reacted. Isolation of pyridine after the experiments under our conditions could not be quantitative; part of it was undoubtedly lost both during the oxidation itself and during distillation. Therefore, in experiments in alkaline medium, where all the evolved \(\mathrm{CO_2}\) was absorbed by the alkali, it is best to determine the amount of amine actually consumed in the reaction by calculating it from the yield of the carbon-containing oxidation products, \(\mathrm{CO_2}\) and oxalic acid.

Experiments in neutral medium do not make it possible to do this, since \(\mathrm{CO_2}\) could not be retained completely. Although, during reduction of \(\mathrm{KMnO_4}\), \(\mathrm{KOH}\) is formed, which apparently might trap \(\mathrm{CO_2}\), it was partly neutralized by manganese acids. Weighing the resulting precipitate always gave figures considerably exceeding the possible weight of \(\mathrm{MnO_2}\) and even of \(\mathrm{MnO(OH)_2}\). Thus 158 grams of \(\mathrm{KMnO_4}\) could give only 87 g of \(\mathrm{MnO_2}\), or 105 g of \(\mathrm{H_2MnO_3}\), whereas 125 g of precipitate dried at 100 degrees were found. From 50 g of \(\mathrm{KMnO_4}\), 42 g of precipitate were obtained (instead of 27.5 g of \(\mathrm{MnO_2}\), or 33.2 g of \(\mathrm{H_2MnO_3}\)). Consequently, the precipitate must have contained something else—evidently potassium. Therefore, the content of \(\mathrm{CO_2}\) found after the experiment is diminished.

On the other hand, the good agreement of the figures obtained for the residual pyridine in neutral and in alkaline solution makes it possible to think that approximately the same amount of it was consumed in both cases.

On this basis, for the experiments in neutral solution we also place in the table approximate figures for the yields of \(\mathrm{HNO_3}\) and \(\mathrm{NH_3}\).

Table No. 2.

Yield of oxidation products, in percent relative to the pyridine that reacted.

From the tables given, the following conclusions may be drawn:

1) Under the action of permanganate in neutral and alkaline medium, every molecule of pyridine in which oxidation has begun is rapidly destroyed further to \(\mathrm{CO_2}\) or \(\mathrm{C_2H_2O_4}\); all more complex intermediate products are evidently very unstable.

2) In a neutral medium, the carbon atoms of pyridine are immediately oxidized to \( \mathrm{CO}_2 \). Only in the experiment with heating was the formation of a small amount of oxalic acid observed; oxidation with an alkaline permanganate solution gives almost identical yields of \( \mathrm{CO}_2 \) and \( \mathrm{C}_2\mathrm{H}_2\mathrm{O}_4 \).

3) During oxidation, the nitrogen of pyridine is liberated for the most part, probably in the form of \( \mathrm{N}_2 \) molecules. A portion of it was always liberated in the form of \( \mathrm{HNO}_3 \). In experiments at ordinary temperature, only a little nitric acid was obtained (about 15%); on heating, its yield increases (to almost 20%) in a neutral medium and to more than 40% in an alkaline solution.

Oxidation of pyridine with neutral permanganate at ordinary temperature also gave a small amount of ammonia. Nitrogen oxides and \( \mathrm{HNO}_3 \) were not detected anywhere. We also repeated the oxidation of pyridine with \( \mathrm{KMnO}_4 \) in the presence of sulfuric acid, but likewise (in contrast to Delepine’s experiments) with an equimolecular amount of oxidizing agent. Most of the pyridine was recovered unchanged; among the oxidation products only acetic acid was found.

1. Bcr 20, 3391 (1887). ↩