On the Mechanism of Oxidation of Bituminous Coal by Atmospheric Oxygen

Article 6.

B. V. Tronov.

I. Literature Review.

On the question of where the process begins in the oxidation of fossil coals by air—a process that sometimes leads to the spontaneous combustion of coal—there are several theories.

In 1861, Gundmann[^1] put forward the supposition that the principal cause of the phenomena of self-oxidation and spontaneous combustion is pyrite admixed with the coal. This opinion was supported in 1866 by Liebig[^1][^2].

Some investigators, especially in England and America, adhere to this view to this day. Parr found that coals with a high content of pyritic sulfur do indeed have a special tendency toward spontaneous combustion[^2]. Graham[^3] likewise attributes the decisive role in the process of spontaneous combustion to the sulfurous iron compounds contained in coal—pyrite and marcasite. Walter, Bielenberg, and Gausswald[^4] think they find confirmation of the pyrite theory in the easy ignition of semicoke from coal containing much iron sulfate. However, many scholars point out that by no means all coals with a high content of sulfurous iron are capable of spontaneous combustion[^2]. In the Moscow basin there occur, for example, bogheads with large inclusions of pyrite that are not prone to spontaneous combustion in air. On the surface of pieces of such boghead, after several months, iron sulfate appears where the pyrite had been, but the coal itself remains without visible changes.

The same was observed on other coals by Macpherson, Simpkin, and Wild[^5].

The fact that semicoke from coal with a high content of iron sulfate ignites more readily may be due not to the formation of sulfurous iron compounds, but to its suboxides, which possess pyrophoric properties[^2].

It is also known that frequently self-igniting coals contain little pyrite and little sulfur in general. In B. S. Krym’s[^6] experiments, certain Donets coals with a low pyrite content ignited more readily than coals richer in it.

The coals of Kuzbass generally contain little sulfur, in particular little pyritic sulfur, yet they ignite comparatively very easily. M. D. Bonh and S. G. Makovskii^79) assembled a large body of literature on the sulfur content of these coals and analyzed several coals. As a result, comparing all available data, they likewise came to the conclusion that there is no dependence of any kind between sulfur content and spontaneous ignition.

Graham^7) advanced another pyrite–fusain theory, according to which the component of coal that is especially dangerous with respect to spontaneous ignition is fusain with finely comminuted pyrite contained in it. A. M. Gladshtein and L. M. Berman^8) agree with this view. A. Putilin^9) also expresses a compromise view. He recognizes the importance of pyrite, but believes that “the role of pyrite in the spontaneous ignition of coal apparently consists in pulverization and the initial warming of the coal.”

F. Mullert^10) writes: “Summarizing, it may be said that the cause of the spontaneous ignition of coals lies in the ability of the organic coal substance formed without access of air to absorb oxygen from the air; the presence of pyrite and marcasite plays a secondary role in this.”

In general, apparently, it must be admitted that if sometimes sulfurous compounds of iron can facilitate spontaneous ignition, then in any case, for many coals, including the Kuzbass coals of particular interest to us, these compounds cannot be regarded as the decisive factor in the outbreak of coal fires. In clarifying the causes of spontaneous ignition, attention must be directed mainly to the substances constituting the organic mass of coal. Here it is necessary to study the relation to oxidation of:

a) the petrographic components of coal;
b) the chemical compounds constituting coal;
c) the individual atomic groups of which the molecules of these compounds consist.

Stott and Wheeler^11) determined the absorption of oxygen at different temperatures by fusain, vitrain, clarain, and durain separately. It turned out that at 15° and 50° oxygen is absorbed considerably more strongly by fusain, to which vitrain, clarain, and durain are far inferior. At 100° the most active prove to be vitrain and clarain, behind which durain and fusain lag. The authors explained this by the fact that at lower temperatures adsorption predominates, for which fusain, by virtue of its physical structure, is more capable, whereas with stronger heating oxidation proper takes place, and the substance of vitrain and clarain is evidently more readily subject to the actual chemical action of oxygen. Francis and Wheeler^12) oxidized vitrain, durain, and fusain from Nottinghamshire coal with air at 100° for 5 months. They found that under these conditions the composition of vitrain changed most of all: the content of C decreased from 79.7% to 72.1%, i.e., by 7.6%; H—from 5.2% to 3.9%, i.e., by 1.3%; the content of O increased from 12.1% to 21.1%, i.e., by 9%. For fusain the smallest ...

changes: C—85.9%—81.1% = 4.8%; H—3.9%—3.3% = 0.6%; gain in O = 13.8%—8.4% = 5.4%. The calorific value of vitrite decreased by 7949—6266 = 1683 cal., and that of fusite by only 6466—5760 = 706 cal.

Kreulen¹³) likewise came to the conclusion that vitrite is more readily oxidized; next in his sequence is durite and, in last place, fusite. Oxidizability was assessed from the initial temperature.

However, for Saxon bituminous coal, Kюndle¹⁴), after oxidation with air at 225°, found identical changes for the whole mass of coal and for the fusite isolated from it. It is also pointed out¹⁵) that in laboratory experiments coals and their petrographic constituents are taken in a pulverized state, whereas under natural conditions fusite has the greatest surface of contact with air, and vitrite the smallest. Thus, the substance of vitrite, by its chemical character, may be more accessible to oxidation, but this difference may be smoothed out and even outweighed by a physical factor—the magnitude of the surface—and the possibility is not excluded that fusite may prove to be the center of the initial rise in temperature.

As regards the individual classes of organic compounds entering into the composition of bituminous coal, Francis and Wheeler¹⁶) concluded that humins are the most readily oxidized; resins occupy second place, followed by structural plant residues and, finally, hydrocarbons. Of the bitumens obtained by treating coal with pyridine and chloroform, according to the same authors¹⁷) the component β, soluble in pyridine but not soluble in chloroform, proved to be the most prone to oxidation in oxygen. This component differs from the others by its high oxygen content; it contains humic substances (in the colloidal state). Bitumen γ, consisting mainly of resins (and compounds related to them) and hydrocarbons, is considerably more resistant to oxidizing agents.

All this is in full agreement with the fact that humic coals ignite more readily than sapropelites (for example, for Barzas sapropelite no cases of spontaneous combustion have so far been recorded¹⁸)).

Erdmann¹⁹) considers the residual coal to be the most readily oxidized part, since the residue obtained from brown coal after extraction of bitumens and humic acids ignited at a lower temperature than the original coal. The brown coals of Germany studied by him had a humic character.

Of great importance for elucidating the mechanism of oxidation of bituminous coal are works in which the presence of particular functional groups in the molecules of the compounds composing coal was determined, as well as the changes in functional groups occurring during coal oxidation.

In this respect we are especially interested in humic substances, whose ready oxidizability is evident from the foregoing.

Fuchs devoted much attention to elucidating the structure of humic acids^20,21,22,23). He found that the molecules of humic acids contain the following active atoms and groups of atoms:

Other authors have found somewhat different ratios among the indicated active groups. For example, for the ratio of the number of phenolic hydroxyls to the number of carboxyls, the following values are found in the literature^24,25,26): 4, 1; 5, 4; 3, 2; 4, 6.

The very skeleton of the molecules of humic acids, in Fuchs’s opinion^27), is a condensed system of six-membered rings, partly benzene, partly hydroaromatic, and partly heterocyclic with oxygen.

The more complex humic substances found in coals proper—up to humic residual coal inclusive—are formed from humic acids, apparently by decarboxylation and, perhaps, dehydration at the expense of certain hydroxyls (not of phenolic character). It is thought^28) that the residual coal of humic formations consists of polycyclic neutral compounds, among which aromatic rings predominate.

Some authors object to such an interpretation of the structure of the nucleus of humic substances. For example, a scheme is proposed^29) according to which, in coal and in humic acids, benzene rings predominate, not condensed but linked through open chains and, in part, through oxygen heterocycles. However, this objection concerns only the skeleton of the molecules, and not the functional groups, which are much more readily determined.

There is rather little information on changes in functional groups during the oxidation of coal by air. Judging from all the literature data, freshly mined coal first begins physically to adsorb oxygen, which then enters into chemical interaction with the coal substance, initially forming unstable peroxide compounds^30). Some investigators have in fact observed the liberation of iodine in the reaction of coals with acidified potassium iodide, which they explain by the presence of peroxides in the coal^31). However, V. S. Krym and S. I. Se-

Semenikhin^32) believe that iodine can (in any case, partly) be liberated through the action of trivalent iron:

\[ 2\mathrm{Fe}^{+++}+2\mathrm{I}^{-}\rightarrow 2\mathrm{Fe}^{++}+\mathrm{I}_{2}. \]

In their opinion, during the oxidation of coals the formation of peroxide compounds occurs to a very slight extent, and these are rapidly converted into substances containing oxygen in a more stable form.

It must also be pointed out that, under the same conditions, quinones can also liberate iodine:

\[ \mathrm{O{=}C} \begin{matrix} & \mathrm{CH} & \mathrm{CH}\\[-2mm] / & & \backslash\\[-1mm] \mathrm{CH} & & \mathrm{CH}\\[-1mm] \backslash & & / \end{matrix} \mathrm{C{=}O} +2\mathrm{HI} \rightarrow \mathrm{HOC} \begin{matrix} & \mathrm{CH} & \mathrm{CH}\\[-2mm] / & & \backslash\\[-1mm] \mathrm{CH} & & \mathrm{CH}\\[-1mm] \backslash & & / \end{matrix} \mathrm{COH} +\mathrm{I}_{2}. \]

Concerning other changes in functional groups, the following is known. During the oxidation of coals the number of hydroxyl groups increases. For coal from the Amur mine (Donbass), it was found^32) that the OH content rose from 0.17 g per 100 g of fresh coal to 4.92 g after 15 days of oxidation at 150°. For coal from the N. Chaikino mine, 0.59 g and 5.10 g were found. However, the determination of hydroxyls was carried out with caustic potash in the presence of phenolphthalein (the excess KOH was titrated with hydrochloric acid), i.e., by a method which is suitable for acids, is wholly unsuitable for alcohols, and does not give clear results for phenols (all the more so since the acidity of phenols depends to a high degree on other parts of the molecule). There are also direct indications^31) that, during the oxidation of coal, carboxyl groups appear in it.

V. S. Krym and S. I. Semenikhin^32) also note an increase in the content of carbonyl groups. In this case it reaches a maximum after some time and then decreases somewhat. Thus, the same coal from the Amur mine showed, in the fresh state, a carbonyl content of 0.41 g per 100 g of the original dry coal; after 12 days—0.79 g; and after 15 days of oxidation only 0.77 g. For coal from the N. Chaikino mine, respectively, 0.32 g, 0.68 g, and 0.58 g (after 16 days) were found. The determination was carried out with hydrochloric-acid phenylhydrazine, the excess of which was titrated with iodine:

\[ \begin{matrix} \mathrm{R}\\ \mathrm{R}_{1} \end{matrix} >\mathrm{C{=}O} +\mathrm{H}_{2}\mathrm{N}\!\cdot\!\mathrm{NH}\!\cdot\!\mathrm{C}_{6}\mathrm{H}_{5}\!\cdot\!\mathrm{HCl} \rightarrow \begin{matrix} \mathrm{R}\\ \mathrm{R}_{1} \end{matrix} >\mathrm{C{=}N}\!\cdot\!\mathrm{NH}\!\cdot\!\mathrm{C}_{6}\mathrm{H}_{5} +\mathrm{HCl}+\mathrm{H}_{2}\mathrm{O}; \quad \mathrm{C}_{6}\mathrm{H}_{5}\!\cdot\!\mathrm{NH}\!\cdot\!\mathrm{NH}_{2} +2\mathrm{I}_{2} \rightarrow \mathrm{C}_{6}\mathrm{H}_{5}\mathrm{I} +\mathrm{N}_{2} +3\mathrm{HI}. \]

In the laboratory of organic chemistry of the Tomsk Industrial Institute, M. Yu. Grigor’ev and K. K. Semenova^33) studied the changes occurring in the bitumens of certain Kuzbass coals during the oxidation of these coals by air at 150–200° and 250°.

The following table gives the yields of bitumen (with alcohol–benzene extraction), the acid numbers, ester numbers, and saponification numbers in mg KOH per 1 g.

As can be seen from the table, in all the coals examined the acid number at 150–200° either decreased or increased only slightly. Oxidation at 250° led, for the first three coals, to an increase in this number, sometimes a very considerable one; one coal (No. 4) gave a lower acidity compared with the initial state, but a higher one than at 150°. The ester number in two coals (Nos. 1 and 2) rose sharply at 150–200° and was somewhat lower (though still higher than the initial value) after oxidation at 250°; in coal No. 3 it decreased continuously, then again increased somewhat. It has not yet been possible to derive any definite general regularity, which, however, is understandable, since only the bitumens were investigated, and not all the constituent parts of the coal. At the same time, the yield of bitumen also changed during oxidation.

We now turn to the principal question considered in the present work: the question of which atomic groups in coal are the first to undergo the action of oxygen. Having resolved this question, one may hope, by determining the corresponding groups, to evaluate coals (at least approximately) from the standpoint of their danger with respect to spontaneous ignition.

After the pyrite theory proved insufficiently satisfactory and the importance of the organic part of coal for its oxidation began to be clarified, the idea was put forward that the oxygen absorbed by coal acts first of all on carbon atoms joined by a double bond. Ferdinand Fischer^34) in 1899 studied the action on coals of bromine water, as well as of a solution of KBr and KBrO₃ in the presence of hydrochloric acid. The coals proved to be very active with respect to these solutions, the absorp-

bromine uptake reaches 30–36% of the weight of the coal. The same coals, after some oxidation, absorbed much less bromine.

From these facts Fischer concluded that oxygen and bromine act on the same constituent parts of coals, namely on unsaturated compounds. He therefore recommended determining the propensity of coals to spontaneous ignition by their ability to react with a solution of $\mathrm{KBr}+\mathrm{KBrO_3}+\mathrm{HCl}$. Fischer also tested iodine (in an alcoholic solution), which, according to him, acted on coals only very slowly, and chlorine, which reacted very vigorously. Habermann[^35] and Dering[^36] also indicate that, from the ability of coals to add bromine, one may judge their tendency toward spontaneous ignition[^37].

In Fischer’s work there is also some discussion of the possible significance of sulfur-containing organic compounds of coal, but the author ascribes the leading role in the process of spontaneous ignition to unsaturated compounds.

The opinion expressed by Fischer has gained such recognition that the authors of some recent works apparently accept it as something self-evident. Thus, E. V. Voitova[^38] writes directly: “The phenomena of spontaneous ignition and self-oxidation of coals during storage depend to a considerable degree (at least, in the opinion of the majority of investigators) on the presence in them of unsaturated compounds, which absorb oxygen rather avidly and determine the course of exothermic reactions. By the quantity of such compounds one may judge the tendency of such coals toward oxidation.”

To determine the ability of coals to oxidize, the iodine number is sometimes used. For example, in the work of E. V. Voitova we find a table in which the spontaneous-ignition points of several coals according to Erdmann and the iodine numbers (according to Gubl) are compared:

A sufficient parallelism is observed, which indeed gives grounds to think that the atomic groups in coals which are active also with respect to iodine are more readily oxidized.

E. Erdmann[^19] in 1922 drew attention to the fact that polyphenols also readily react with oxygen, and explained the easy oxidizability of residual coal obtained from German brown coals by the presence of aromatic nuclei with several hydrox-

strong groups. However, this explanation is almost nowhere given in the more recent literature. G. L. Stadnikov^39) (1933) considers it acceptable for the brown coals of Germany, but not of general significance. Later,^40) this author once again returns to the question of the significance, for the oxidation of coal, of benzene nuclei with several hydroxyls, but in a rather indefinite form, quite rightly noting: “We have no data at our disposal that would make it possible to compare the rate of oxidation of unsaturated acids with the rate of oxidation of humic acids.”

In other recent monographs and articles, Erdmann’s opinion, it seems, is not cited at all.

Finally, Mullert^41) points to the possibility that hydroaromatic compounds may also participate in oxidative processes.

Thus, the most generally accepted opinion at present is that the leading role in the process of spontaneous ignition of coals belongs to unsaturated compounds. This opinion is based chiefly on the high activity of coals with respect to haloids, i.e., to substances that are regarded as typical reagents for double bonds between carbon atoms, and also on a certain parallelism between the oxidizability of coals and their ability to react with haloids. Among other atomic groups that oxidize more or less readily, the literature contains indications of benzene rings with several hydroxyls, and hints of sulfur-containing groups and hydroaromatic rings.

II. Work of the Departments of Organic Chemistry of TII and TSU on determining the oxidizability of various organic compounds.

To determine from which functional groups of the complex compounds composing the organic mass of bituminous coal the oxidation process begins, it would be very useful to have information on the oxidizability of the most diverse organic substances.

In the laboratories of organic chemistry of the Tomsk Industrial Institute and Tomsk State University, work has been under way for a number of years (since 1923) to determine the oxidation rate of organic compounds of various classes. So far the following have been studied: hydrocarbons^42) — saturated, aromatic, and cyclanes. Halogenated hydrocarbons^42). Alcohols with one^43, 44, 45, 46) or several^47) hydroxyls, as well as phenols^48). Simple and compound ethers^49). Aldehydes and ketones^50). Carboxylic acids of the saturated series^51), unsaturated^52), halogen-substituted^53), acids of the benzene series^54), hydroxy acids^55), keto acids^55, 56). Carbohydrates^47, 57, 58). Sulfo acids^54). Nitroso compounds^59). Amines: saturated^60), cyclane^60), aromatic^61). Nitrogenous heterocycles^62). Amino acids^54, 63). Quaternary ammonium salts^64).

The following were used as oxidizing agents for the time being: $\mathrm{KMnO_4}$ in neutral, acidic, and alkaline media; $\mathrm{CrO_3}$ in water (sometimes in the form of a chromic mixture) or in acetic anhydride; and diluted $\mathrm{HNO_3}$. The experiments were carried out at temperatures from 0° to 25°, most often at 18–20°. The reaction rate was determined by titrating, after known intervals of time, the active oxygen remaining.

The calculation of constants did not always give results of any satisfactory value. Therefore, instead of constants, relative values of the reaction rate were sometimes determined, by comparing the intervals of time during which the same percentage had reacted. This gave results similar to comparison by constants in those cases where constants could be calculated.

The most convenient oxidizing agent for the work proved to be permanganate, which is inapplicable only to hydrocarbons and carbohydrates insoluble in water. Nitric acid can also nitrate and esterify. $\mathrm{CrO_3}$ in small concentrations acts too weakly; in addition, it sometimes gives colored products that hinder titration, as was observed, for example, with phenols.

Therefore, in the table given below, the values of the average relative oxidizability of a number of characteristic representatives of various classes of organic compounds are derived only from the results of experiments with potassium permanganate. As the unit of comparison, the rate of oxidation of permanganate by methyl alcohol was adopted, with the concentration of each component of the mixture ($\mathrm{CH_3OH}$, $\mathrm{KMnO_4}$, and $\mathrm{KOH}$) equal to 0.03 mole per liter and at a temperature of 20°. This experiment gave the following results:

Since the oxidation rate often varied very greatly depending on the reaction of the medium, the table gives, for each compound, figures referring to neutral, alkaline, and acidic solutions. For acids and amines, the solution of the salt of the substance being oxidized with $\mathrm{NaOH}$, $\mathrm{KOH}$, or $\mathrm{H_2SO_4}$ was taken as the neutral medium; in addition, the reaction rate is indicated simply in aqueous solution, i.e., in the acidic or alkaline medium created by the substance itself. The table includes only the results of experiments made at room temperature, generally between 17° and 21°.

The values of the average relative rates are taken for the most part for the interval from 0% to 40% or 50%. Where they go beyond these limits (for example, in the case of very rapidly reacting—

...compounds), or do not reach them at all (a very slow reaction); the corresponding numbers are enclosed in parentheses. The same has been done in those cases where the results have not yet been reduced to unity and the comparison has been made only from a few experiments. With very large changes in the rate during the course of the process, averages were nevertheless calculated, but a question mark (?) was placed.

Determination of the reaction rate for very readily oxidized compounds could not be very exact, since the mixing of the solutions itself took no less than 1–1.5 sec. In addition, as specially arranged experiments showed, the substances most active toward permanganate also react with iodine and, consequently, the process did not cease after the addition of a solution of potassium iodide and hydrochloric or sulfuric acid.

Nevertheless, our work still makes it possible to judge the order of magnitude of the rates. The mixing of the reacting components took approximately the same time. The same was also true for nitration. In addition, in terms of the rate of reaction with permanganate and with iodine, phenols, for example, were of the same order[^65].

The numbers in the tables presented show that, of all the organic substances we studied, the first place in activity toward oxidation by permanganate is occupied not by unsaturated alcohols and acids, but by phenols. Moreover, this applies not only to polyphenols, which reacted tens, and sometimes hundreds and thousands (pyrogallol) of times faster than allyl alcohol and unsaturated acids. It turned out that even simple phenol and cresols are, in oxidizability, somewhat superior to unsaturated compounds.

After the phenols, in rate of reaction with permanganate, come unsaturated acids and allyl alcohol, as well as aromatic amines. Next come aldehydes and diketones, fatty amines (in alkaline medium), certain polyhydric alcohols, and carbohydrates (also in the presence of alkali). Hydroaromatic alcohols are oxidized much more difficultly. It has not yet been possible to determine the oxidizability of sulfur-containing organic compounds.

Our experiments were carried out not with the oxidizing agent that acts on coals under natural conditions, but all known facts show that free oxygen also acts readily on phenols and, one must suppose, reacts with them faster than with unsaturated compounds. Almost all phenols color fairly rapidly in air. Pyrogallol is even used in gas analysis for absorbing oxygen, since it is, apparently, the most readily oxidized of all well-studied organic substances.

As for double bonds, on the contrary, there is reason to think that in their activity toward atmospheric oxygen they yield to phenols even more than in the rate of reaction with permanganate. It is known that terpenes are readily oxidized in air, but it is not the double bonds that react in them, but other parts of the molecule. Thus, in limonene[^66] one of the CH₂ groups is oxidized;

[chemical structural scheme]

In pinene^67), rupture of the four-membered ring occurs, again with preservation of the double bond:

[chemical structural scheme]

According to other data^68), pinene with oxygen forms a peroxide of the following structure:

[chemical structural formula]

Here, too, the double bond does not take a direct part in the reaction.

Thus, everything that is known about the rate and direction of the action of oxidizing agents on organic compounds compels us to regard phenolic groups as the most readily oxidized parts of the molecules constituting the organic matter of fossil coals. In this process, not only benzene rings with several hydroxyls—which may be few in coals—are oxidized, but also those with a single hydroxyl.

Of secondary importance may be undefined groups, aminated benzene rings (but not pyridine rings, since pyridine, its homologues, and even quinoline were oxidized considerably more slowly), and, perhaps, in part organic sulfur.

III. Scheme of the oxidation of coal by atmospheric oxygen and conclusions from it.

In the preceding chapters we have seen that:

1) Of the constituent parts of fossil coals, humic substances are more readily oxidized, from humic acids to the residual coals of humus origin, which are probably constructed similarly to humic acids. Some bitumens are also readily oxidized, but precisely those which in their properties resemble humic substances; for example, they are extracted by the basic reagent pyridine and do not dissolve in neutral chloroform, i.e., they possess an acidic character.

2) Humic acids have in their molecules a fairly large number of phenolic hydroxyls, which, apparently, are retained mainly also upon conversion into more complex products. In any case there are noticeably more phenolic groups than undefined ones.

3) Of all sufficiently studied organic compounds, phenols are the most readily oxidized both by potassium permanganate and, so far as can be seen at present, by free oxygen.

All this leads to the idea that atmospheric oxygen acts in the first instance precisely on the phenolic groups of coals, and that those benzene rings which have only one hydroxyl are also oxidized.

As is known, in the oxidation of phenols with one hydroxyl, hydroxylation of the hydrocarbon units occurs first of all. For example, \( \mathrm{C_6H_5OH} \), under the action of air, gives quinone, and the intermediate product is probably hydroquinone:

\[ \begin{array}{ccc} \begin{array}{c} \mathrm{COH}\\[-2pt] \mathrm{HC}\ \hexagon\ \mathrm{CH}\\[-2pt] \mathrm{HC}\ \ \ \ \mathrm{CH}\\[-2pt] \mathrm{CH} \end{array} & \longrightarrow & \begin{array}{c} \mathrm{COH}\\[-2pt] \mathrm{HC}\ \hexagon\ \mathrm{CH}\\[-2pt] \mathrm{HC}\ \ \ \ \mathrm{CH}\\[-2pt] \mathrm{COH} \end{array} \longrightarrow \begin{array}{c} \mathrm{C{=}O}\\[-2pt] \mathrm{HC}\ \hexagon\ \mathrm{CH}\\[-2pt] \mathrm{HC}\ \ \ \ \mathrm{CH}\\[-2pt] \mathrm{C{=}O} \end{array} \end{array} \]

In the next stage of the reaction the hydroxyls lose hydrogen atoms. Carbonyl groups are formed.

The oxidation of phenolic groups in coals probably proceeds analogously. If a benzene ring with several hydroxyls is oxidized, then hydrogens must leave at once and hydroxyls will be replaced by carbonyl oxygens. In the oxidation of groups with one hydroxyl, new OH groups will first appear, and only later will the accumulation of carbonyls take place.

Since oxygen atoms weaken the bonds between the carbon atoms, with further action of oxygen the ring may be ruptured, one of the carbonyls departing in the form of carbonic anhydride, while the second O atom from the molecule \(O_2\) may take its place. An anhydride will be obtained, which under the action of water will give an acid.

A rupture of the bond between the carbonyl groups without the action of oxygen is also possible. In that case carbon monoxide should be evolved.

The indicated transformations may be expressed approximately by the following scheme:

\[ \begin{array}{ccc} \begin{array}{c} {-}C{-}COH\\[-2pt] {-}C{\!}\diagup \quad \diagdown CH\\[-2pt] \quad I\\[-2pt] {-}C{=}CH \end{array} & \longrightarrow & \begin{array}{c} {-}C{-}COH\\[-2pt] {-}C{\!}\diagup \quad \diagdown COH\\[-2pt] \quad II\\[-2pt] {-}C{=}COH \end{array} \longrightarrow \begin{array}{c} {-}C{-}C{=}O\\[-2pt] {-}C{\!}\diagup \quad \diagdown C{=}O;\\[-2pt] \quad III\\[-2pt] {=}C{-}C{=}O \end{array} \end{array} \]

\[ \begin{array}{ccc} \begin{array}{c} {-}C{-}C{=}O\\[-2pt] {-}C{\!}\diagup \quad \diagdown C{=}O + O{=}O\\[-2pt] \quad III\\[-2pt] {-}C{-}C{=}O \end{array} & \longrightarrow & \begin{array}{c} {-}C{-}C{=}O\\[-2pt] {-}C{\!}\diagup \quad \diagdown O + CO_2;\\[-2pt] \quad IV\\[-2pt] {=}C{-}C{=}O \end{array} \end{array} \]

\[ \begin{array}{ccc} \begin{array}{c} {-}C{-}C{=}O\\[-2pt] {-}C{\!}\diagup \quad \diagdown O + H_2O\\[-2pt] \quad IV\\[-2pt] {=}C{-}C{=}O \end{array} & \longrightarrow & \begin{array}{c} {-}C{-}C{=}O\\[-2pt] {-}C{\!}\diagup \quad \diagdown OH\\[-2pt] \quad V \quad OH;\\[-2pt] {=}C{-}C{=}O \end{array} \end{array} \]

\[ \begin{array}{ccc} \begin{array}{c} {-}C{-}C{=}O\\[-2pt] {-}C{\!}\diagup \quad \diagdown C{=}O\\[-2pt] \quad III\\[-2pt] {=}C{-}C{=}O \end{array} & \longrightarrow & \begin{array}{c} {-}C{-}C{=}O\\[-2pt] {-}C{\!}\diagup \quad | \quad + CO.\\[-2pt] \quad VI\\[-2pt] O{=}C{-}C{=}O \end{array} \end{array} \]

All these transformations are equally possible both in a ring belonging to a condensed polynuclear system and in a ring separated from the others by side chains.

The process described here of a profound chemical alteration of the coal substance is, of course, preceded by the physical absorption of oxygen—adsorption. The possibility of the intermediate formation of unstable peroxides is also not excluded, although it apparently cannot be considered proven. The observed liberation of iodine upon the addition to coal of an acidified solution of potassium iodide may be explained by the reaction of quinone groups with hydroiodic acid (see above).

From the proposed scheme of coal oxidation the following follows:

1) If mainly phenolic groups with one hydroxyl are oxidized (simply because there must be more of them than polyphenolic groups), then under the action of air on coal

initially an increase in weight must occur. Let us compare the molecular weights of the benzene ring shown in the scheme and of the products of its transformation without the gases evolved:

\[ \begin{gathered} \mathrm{I}—91;\quad \mathrm{II}—123;\quad \mathrm{III}—120;\\ \mathrm{IV}—108;\quad \mathrm{V}—126;\quad \mathrm{VI}—92. \end{gathered} \]

The weight increases considerably during the initial absorption of oxygen; subsequently, over several stages of transformation, it remains almost unchanged. The only exception is the process of evolution of carbon monoxide, but at high temperatures it is of secondary importance.

For example, Mahler\({}^{69}\) found that, when air was passed for 30 hours through powdered coal, for 150 g of coal at \(20—30^\circ\), \(1\) ml of \(\mathrm{CO_2}\) and \(2.88\) ml of CO were evolved; at \(65^\circ\) the amounts of \(\mathrm{CO_2}\) and CO were \(4.25\) ml and \(5.59\) ml, and at \(105^\circ\), 30 and \(6.67\) ml. L. N. Berman and I. E. Kagan\({}^{70}\) arrived at the same conclusion; they oxidized coals from the Moscow Basin at temperatures from \(100\) to \(175^\circ\). The yields of \(\mathrm{CO_2}\) and CO, in percent, changed with increasing temperature for various coals as follows:

No. 1: \(\mathrm{CO_2}\) from 0.15 to 0.42; CO from 0.18 to 0.22
No. 2: \(\mathrm{CO_2}\) from 0.20 to 0.75; CO from 0.16 to 0.19
No. 3: \(\mathrm{CO_2}\) from 0.21 to 0.67; CO from 0.15 to 0.19
No. 4: \(\mathrm{CO_2}\) from 0.11 to 0.52; CO from 0.12 to 0.15.

From this it is clear that investigators who dealt with high temperatures note, as a general regularity, first an increase in the weight of the coal, then the preservation of the weight almost constant for a prolonged time. Thus in Kreulen\({}^{71}\), who subjected coals to very prolonged oxidation at \(105^\circ\), for 958 h an increase in weight occurred; then, for 1275 h, it remained unchanged, and afterward began to fall.

The decrease in the weight of coal upon further heating with air must be explained chiefly by decarboxylation of the carbonic acids being formed:

\[ \begin{array}{c} \begin{array}{ccccc} & {-}\mathrm{C}{-}\mathrm{C}{=}\mathrm{O} & & & {-}\mathrm{C}{-}\mathrm{H}\\ & \diagup & \diagdown\mathrm{OH} & & \diagup\\ {-}\mathrm{C} & & \mathrm{V} & \longrightarrow & {-}\mathrm{C}\ \mathrm{VII} + 2\mathrm{CO_2}.\\ & \diagdown & \diagup\mathrm{OH} & & \diagdown\\ & {=}\mathrm{C}{-}\mathrm{C}{=}\mathrm{O} & & & {=}\mathrm{C}{-}\mathrm{H} \end{array} \end{array} \]

The molecular weight of the ring, which has begun to break down, falls to 38 in VII.

To this are added the processes of transition from III to IV \((120 \to 108)\) and from III to VI \((120 \to 92)\).

2) In the functional groups of the molecules constituting coal, essentially the following changes should take place:

a) The content of phenolic hydroxyls should at first increase, and then decrease.

b) The content of carbonyl groups will likewise increase at the beginning of oxidation, and later will begin to decrease, but somewhat later.

c) The content of carboxyl groups will do the same, but both the increase and the subsequent decrease should proceed still more slowly.

The few data available in the literature on this question^31, 32) generally confirm these conclusions qualitatively. Precise quantitative regularities are difficult to expect here, since different parts of the molecules—in particular, different hydroxylated benzene rings—may at the same moment be at different stages of oxidation. In one part of a molecule the process of accumulation of phenolic hydroxyls may still be proceeding; in another part, or in another molecule, at the same time the phenolic groups may perhaps have begun to be converted into quinone groups; a third part (or molecule) may already have reached cleavage with the formation of carboxyl groups, etc. Moreover, not only phenolic groups can be oxidized.

3) According to the proposed scheme, the gaseous products of coal oxidation should include both carbon dioxide and carbon monoxide.

Workers in the coal industry are very interested in the question of whether, from the results of analysis of the air in a mine, it is possible to determine the onset of spontaneous ignition of coal—in particular, whether the presence of carbon monoxide is not an indication of fire. The old theory, which explained the spontaneous ignition of coals by the rapid oxidation of unsaturated compounds, made it likely that carbon monoxide is evolved only at rather high temperatures, already characteristic of a fire. The evolution of CO during the oxidation of unsaturated compounds without a strong rise in temperature is theoretically incomprehensible and has not been observed experimentally. Formation of CO was observed under the action of air, for example, on hexylene \(\mathrm{CH_2{=}CH—CH_2—CH_2—CH_2—CH_3}\), but the temperature was \(125\text{–}145^\circ\).

The phenolic theory of coal oxidation, as is evident from the scheme given above, proposes a very simple mechanism for the evolution of carbon monoxide. Experience also shows that, for example, the most readily oxidized phenol, pyrogallol, forms CO when reacting with oxygen at ordinary temperature.^72, 73)

As regards bituminous coal, practical experience indicates that carbon monoxide can be detected even in places where there are no signs whatever of fire. Graham and Storr^74) in South Yorkshire (England) found in the air of normal old workings from 0.006% to 0.022% CO; in workings with slight heating there was from 0.006% to 0.014%; in the vicinity of places with strong old heating, up to 0.19% CO was found.

In the first large laboratory investigation of the process of oxidation of coal by air, carried out by Richter^75), it was proved

the formation of CO₂ and H₂O, but nothing is said about CO, although the experiments were carried out at 180°–200°. However, Mahler⁶⁹), already at 20°–30°, detected both CO₂ and CO.

A. P. Shakhno and D. A. Podzharskaya⁷⁶) analyzed the air from jars in which, for several months, crushed coals from the Central Asian deposits Kizil-Kiya and Kok-Yangak had been kept. In all cases they found contents of both CO₂ (from 0.36% to 0.88%) and CO (from 0.05% to 0.21%). The same results were obtained in work carried out in 1937 at the Kuzbassugol Research Institute by N. F. Denisova⁷⁷). Large pieces of coal (cubes with an edge of 12 cm) stood at ordinary pressure and at temperatures of 25° and 45° in an atmosphere of air, oxygen, a mixture of air with oxygen, and gas containing 11–15% O₂ and 6–10% CO₂. The duration of the experiments was from 58 to 73 days. For greater certainty, the determination of carbon monoxide in the gas above the coal was carried out in two institutes (also at the Institute of Sanitation and Hygiene) and by two methods: with palladium chloride and with iodic anhydride. Everywhere carbon monoxide was likewise detected, the content of which, for samples kept at 25°, sometimes reached 1% and more.

In the laboratory of organic chemistry of Tomsk State University, analyses were made⁷⁸) of the air in jars containing crushed samples of Kuzbass coals that had stood for a little more than a year. In all jars both carbonic anhydride (0.1%–0.24%) and carbon monoxide (0.05%–0.28%) were found.

Already after the report on the results of the present work (at the meeting of the Scientific and Technical Council of KNIUI on December 24, 1938), carbon monoxide was found in the air at the experimental section of mine No. 3 of the Kiselevsk mine field, but upon a detailed examination of the entire section no fire was detected.

Thus, it must be regarded as undoubtedly established that bituminous coals, when oxidized in air at ordinary temperature, yield not only dioxide but also carbon monoxide. Consequently, in this respect coals behave in the same way as certain phenols. This once again argues for the correctness of the proposed oxidation scheme.

In general, the phenolic theory explains without contradiction a whole series of facts and explains them, as it seems to us, better than other theories of coal oxidation. Of course, for a more complete elucidation of the mechanism of oxidation and spontaneous ignition of coal, further investigations are still necessary. It would be very interesting, for example, to study the oxidation of various phenols by air under the conditions under which coal is oxidized, and to carry out a complete analysis of the products, including the gaseous ones. It would also be interesting to clarify in detail the relationship between the content of phenolic groups in coals and the oxidizability of the latter, to try experimentally to trace the change of these groups in the process of oxidation, etc. However, all this will probably lead only to a refinement of the theory, and will not compel its rejection.

The work was carried out with funds from the Kuzbass Scientific Research Coal Institute (KNIUI).

REFERENCES

1. E. Erdmann.—“Brennstoff-chemie” III, 258 (1922).
   V. S. Krym.—“Chemistry of Solid Fuel,” 2nd ed., p. 272 (1936).
2. G. L. Stadnikov.—“Chemistry of Coal” (1933), p. 280.
3. Graham. Br.—Ch. 8, 62 (Ref.).
4. Walter, Bilenberg, Hauswald. Br.—Ch. 7, 333 (Ref.)
5. Macpherson, Simpkin, Wild. Br.—Ch. 8, 62 (Ref.)
6. V. S. Krym. Kh.T.T. 1, No. 4, p. 3 (1930).
7. Graham. Transact. Inst. Min. Eng. 1925, p. 41.
8. A. M. Gladshtein and L. M. Berman.—Materials of the 1st All-Union Conference on the Changes in Coal during Storage (1932), p. 71.
9. A. Putilin.—“Recent Data on the Spontaneous Ignition of Coal” (1933), p. 20.
10. F. Mullert. “Sulfur in Coal” (1932), p. 20.
11. M. Stopes, R. W. Wheeler. Fuel. II, 122 (1923).
12. W. Francis, R. W. Wheeler. Journ. Chem. Soc. 127, 112 (1925).
13. D. J. W. Kreulen—Chem. Weekbl. 29, 282; Chem. Zbl. 1932, II, 2402.
14. Br.—Ch. 9, 297 (1928).
15. G. L. Stadnikov. “Chemistry of Coal,” p. 285.
16. W. Francis, R. W. Wheeler. Fuel. XI, 356—359 (1932).
17. V. S. Krym. “Chemistry of Solid Fuel,” 2nd ed., pp. 273—4.
18. V. G. Palvelev. Kh.T.T. VIII, No. 9, p. 793 (1937).
19. E. Erdmann. Br.—Ch. 3, 296 (1922).
20. W. Fuchs, W. Stengel, F. Bangert. Br.—Ch. 9, 181 (1928).
21. W. Fuchs, W. Stengel. Br.—Ch. 10, 303 (1929).
22. W. Fuchs. Br.—Ch. 11, 332 (1930).
23. W. Fuchs. Zeit. angew. Chem. 44, 116 (1931).
24. N. M. Karavaev, I. B. Rapoport, O. A. Holler. Kh.T.T. V, 510 (1934).
25. H. Stach. Braunkohle. 1933, 617.
26. G. L. Stadnikov.—“The Origin of Coals and Petroleum,” 3rd ed., p. 423 (1937).
27. V. S. Krym.—“Chemistry of Solid Fuel,” 2nd ed., p. 140.
28. G. L. Stadnikov.—“Chemistry of Coal,” p. 156.
29. R. V. Randall, M. Benger, C. M. Groocock. Proc. Roy. Soc. (London) A 165, 432—52 (1938); Chem. Abstr. 1938, No. 14, p. 5373.
30. V. S. Krym.—“Chemistry of Solid Fuel,” 2nd ed., p. 272.
31. Shevchenko.—“Chemical Abstract Journal,” 1935, No. 12.
32. V. S. Krym and S. I. Semenikhin. Kh.T.T. VIII, No. 9, pp. 782—786 (1937).
33. Work being prepared for publication.
34. Ferd. Fischer. Zeit. angew. Chem. 1899, pp. 764—7, 787—790.
35. Journ. f. Gasbel. 1906, 422.
36. Fischers Jahresber. 39, 13 (1908).
37. Strache—Lant.—“Chemistry of Coal,” p. 263 (1931).
38. E. V. Voitova.—Kh.T.T. VII, No. 3, pp. 249—262 (1936).
39. G. L. Stadnikov. “Chemistry of Coal,” p. 282.
40. Ibid., p. 290.
41. F. Mullert. “Sulfur in Coal” (1932), pp. 10—11.
42. B. V. Tronov, P. P. Popov, V. F. Kolotovkina and P. Kh. Stadash. “Proceedings of S.Kh.T.I.” No. 2 (1931).
43. B. V. Tronov, A. A. Lukanin and I. I. Pavlinov.—Zh.R.Kh.O. 59, 1157—1172 (1927).
44. B. V. Tronov, V. F. Udodova and M. I. Chizhova.—Zh.R.Kh.O. 59, 1149—1156 (1927).
45. B. V. Tronov and A. I. Kravchenko.—“Proceedings of T.G.U.” 87, No. 2, pp. 100—104 (1935).

1. B. V. Tronov and V. I. Lutkova—Work submitted for publication.

2. B. V. Tronov and M. P. Sinyavin.—“Izv. S.Kh.T.I.,” No. 2 (1931).

3. B. V. Tronov, A. I. Kravchenko, and E. V. Yudenicheva.—Work submitted for publication.

4. B. V. Tronov and F. Z. Gorfinkel.—Work submitted for publication.

5. A. I. Kravchenko. “Trudy T.G.U.” 94, 55—74 (1938).

6. B. V. Tronov, M. T. Yalovaya, and K. Ya. Budavova.—“Izv. T.I.I.” 54, No. 11, pp. 3—17 (1935).

7. B. V. Tronov and M. T. Yalovaya.—“Trudy T.G.U.” 94, 36—47 (1938).

8. B. V. Tronov and E. G. Kondrakhina.—“Trudy T.G.U.” 87, No. 2, pp. 105—114 (1935).

9. B. V. Tronov and A. A. Grigorieva.—Zh.R.Kh.O. 61, pp. 653—665 (1929).

10. B. V. Tronov and I. I. Zharikov.—“Izv. S.Kh.T.I.,” No. 2 (1931).

11. A. I. Kravchenko—“Trudy T.G.U.” 94, 44—54.

12. M. D. Bon and I. M. Karpеnko.—Work submitted for publication.

13. E. V. Bakulina.—work not yet completed.

14. B. V. Tronov and A. A. Grigorieva. Zh.R.Kh.O. 60, 1014 (1928).

15. L. S. Nikonova.—“Izv. S.Kh.T.I.” No. 2 (1931).

16. B. V. Tronov, L. N. Dyakovova, and M. T. Khodalevich.—Work submitted for publication.

17. B. V. Tronov and L. S. Nikonova.—Zh.R.Kh.O. 61, 541—9 (1929).

18. B. V. Tronov and L. S. Nikonova. Zh.R.Kh.O. 61, 2390 (1929).

19. V. I. Lutkova.—Work is being prepared for publication.

20. B. V. Tronov and V. Z. Grigorieva.—Work submitted for publication.

21. A. Beutmann, O. Zeitschel.—Chem. Zbl. 1914. II, 1235.

22. See, for example, P. Karrer—“Course of Organic Chemistry,” p. 741.

23. K. Suzuki Chem. Zbl.—1937, 1, 2612.

24. A. P. Shakhno and D. A. Podzharskaya.—Kh.T.T. VI, No. 5, pp. 379—403 (1935).

25. L. N. Berman and I. E. Kagan.—Kh.T.T. V, No. 2, pp. 115—125 (1934).

26. D. J. W. Kreulen. Fuel.—5, 345 (1926).

27. Calvert, Cloëz. Lieb. Ann—130, 248 (1864).

28. Boussingault. Lieb. Ann. 130, 243 (1864).

29. A. Putilin.—“Recent Data on the Self-Heating of Coal” (1933), pp. 106—107.

30. Richters. Dingl. polyt Journ. 195, 258;
    V. S. Krym.—“Chemistry of Solid Fuels,” 2nd ed., pp. 271—2.

31. A. P. Shakhno and D. A. Podzharskaya. Kh.T.T. VIII, No. 9, pp. 813—820 (1937).

32. N. F. Denisova.—“Establishing the Conditions for Bringing Coal into a Stable State with Respect to Atmospheric Oxygen” (unpublished).

33. B. V. Tronov and S. G. Makovsky.—Work submitted for publication.

34. M. D. Bon and S. G. Makovsky—Work submitted for publication.

Appendix

Average values of the relative reaction rate of organic compounds with potassium permanganate

Table 1.

Average values of the relative reaction rate

…of organic compounds with potassium permanganate

Table 2

Mean values of the relative reaction rate

organic compounds with potassium permanganate

Table 3

Average values of the relative reaction rate of organic compounds with potassium permanganate

Table 4

Average values of the relative reaction rate of organic compounds with potassium permanganate

Table 5.

Average values of the relative reaction rate of organic compounds with potassium permanganate

Table 6.