P. F. Volodin.

Degree of Softening of Caking Coals and Their Thermal Stability as Indicators of Coking Properties

I. Coking capacity, caking capacity, and melting of coal.

By coking capacity is understood the ability of a coal, or of a coal blend, under conditions of dry distillation without access of air at relatively high temperatures, on the order of 800° (in old fireclay ovens)—1100° (in modern high-speed silica-brick ovens), to yield coarse-lump coke that withstands the established test for mechanical strength.

It is quite obvious that, in order to obtain large pieces of coke from finely crushed coal, the coal must, in the process of coking, pass either through a stage of melting of the individual grains, which as a result merge into a homogeneous continuous mass, or through a stage of liberation from the grains of liquid, resinous substances that impregnate the entire coal mass and cement it upon subsequent solidification as a result of the driving off of vapors and gases. In accordance with this, there exist in science two points of view, two theories of coking: the theory of coal melting, whose originator is one of the greatest authorities in the field of coal research, the Soviet scientist G. L. Stadnikov, and the theory of cementation, whose most prominent representative is the major English specialist in the field of coking, R. A. Mott. In the further exposition facts will be presented on the basis of which it will be possible to judge which of these theories corresponds to the actual state of affairs to the greatest extent. But, one way or another, one thing may be considered beyond doubt: the first and principal indication of the coking capacity of coals is their ability, upon heating, to form a more or less homogeneous fused mass, which, during the shrinkage of the coal-coke mass, breaks up into separate large pieces.

It is, of course, quite natural that in the laboratory study of the question of the indicators of coking capacity, primary attention was focused on this property of coal, on the development of methods for its quantitative characterization. Different investigators approached the solution of the question in different ways; however, two main directions may be noted which, perhaps to a significant degree, reflect the two theoretical premises mentioned above. The first of these directions denotes the indicated property of coals by the term “caking capacity.” By this term, depending on the proposed method of determination, is understood either the fused character and homogeneity of a crucible coke button, as well as its swelling, determined qualitatively by eye; or the ability of coal, in a crucible test, to give a caked residue when variable amounts of some inert additives (sand, coke, anthracite, etc.) are admixed with it; or, finally, testing the button obtained in a crucible test from a mixture of coal with a constant amount of inert additive by crushing under a variable load. Both the terminology itself and the nature of the testing methods indicate that the premise for this direction is a theory which assumes the presence, or subsequent formation, in coal of

of binding (cementing) material. The caking capacity of a coal is the higher, the greater the amount of inert additive that one and the same quantity of it is capable of binding; the caking capacity is the higher, the more firmly a certain quantity of inert additive is bound to the coal. From this it followed that caking is the same as coking capacity. Practice, however, has shown that caking cannot be identified with coking capacity. Quite apart from the fact that the so-called “fat” coals, which show very high caking capacity, are as a rule unsuitable for coking without the addition of other “lean” coals, coals possessing one and the same caking capacity, as may be judged from the results of experiments by the well-known German investigator P. Damm (1), may differ very sharply in their coking properties.

The second approach defines the same property of coals as the result of their softening, of their transition into the plastic state. Of the methods proposed for the quantitative characterization of the plastic state of coals in laboratory research practice, two have become more or less widely used: the penetrometric method of the German investigators G. Agde and L. Linker, and the method of the English investigator Foxwell. According to the method of Agde and Linker (2), a thin steel rod, loaded from above with a weight of definite mass, is placed on a coal charge placed in a steel tube and lightly tamped. As the coal softens during heating (at a rate of 2–3° per minute), the rod begins to sink into the coal charge. The depth of penetration, on a scale enlarged 10 times, is marked by a pointer on a circular disk. Curves can be plotted from the points: temperature and depth of penetration. According to Foxwell’s method (3), the study of the plastic state of coals is based on measuring the magnitude of the resistance of a current of inert gas (usually nitrogen), passed through the coal charge, to the coal softening during heating. The curve is plotted from the points: temperature and magnitude of the resistance in millimeters of water (or, at high resistances, mercury) column.

When carrying out an experiment by the Agde and Linker method, in the case of a small load on the needle, coals possessing the ability to swell give such rises and descents of the needle throughout the entire experiment that it is difficult to determine the most important point—the maximum of melting. In the case of a large load, the needle gives a jump at the beginning of coal softening, and again it is impossible to determine the point of maximum melting. A modification of this method using needles of different cross-section for coals of different fusibility is not a successful solution of the problem, since in this case it is difficult to compare the data obtained for different coals.

As for Foxwell’s method, the following should first of all be noted. When testing coals of a high degree of fusibility, at a certain temperature there comes a moment of complete cessation of gas movement through the coal; the resistance is then so great that rubber tubes cannot withstand the pressure necessary to overcome it. If the flow rate of the constant quantity of inert gas passing through the coal is sharply reduced, then such a high resistance at the point of maximum melting is not observed, but in this case it is very difficult to determine the point of the beginning of melting, since the change in resistance will be barely perceptible. The absolute values of the resistances in millimeters of water or mercury column in two parallel experiments almost do not coincide, but the method makes it possible to record fairly accurately the temperature points of the beginning and maximum of melting, by the abrupt change in resistance, and they agree closely in parallel experiments.

By either method, about 3 hours must be spent on a single determination. Quite apart from the fact that the considerable complexity of carrying out

the experiments and their duration make these methods of little suitability for use in a plant laboratory; the data obtained themselves do not provide the necessary guidance in determining the coking properties of coals. For this reason, neither method has gained any currency at our coke-chemical plants, despite their fairly wide use in the practice of research laboratories.

The second line also includes the method of L. M. Sapozhnikov (4). The idea of the method is to reproduce, on a small laboratory coal sample (100 grams), the conditions of coal coking in a coke oven. A coal charge placed in a steel cylinder is heated from one side, namely from below, at a rate of 3° per minute. Throughout the experiment the coal in the steel cylindrical cup is under a constant load pressure, calculated as 1 kilogram per square centimeter. The perforated bottom of the cup, covered with a thin layer of asbestos board, and an identical perforated die on top, transmitting the pressure of the load to the coal (also resting on a thin layer of asbestos), allow the gases and vapors of decomposition to escape toward the hot (lower) and cold (upper) sides of the coal charge. When the coal is charged into the cup, a thin paper sleeve is placed in it over the full height of the charge. This sleeve disappears even before the coal reaches the temperature at which melting begins, and thus an opening is formed in the coal charge in the vertical direction. As in a coke oven, when the required temperature is reached, a layer of softened coal (the plastic layer) appears in the lower part of the charge; as the temperature rises, it will move together with it into the upper layers of the charge, while semicoke forms below and, with subsequent degassing, coke. At the same time, the opening in the coal charge at the place where the plastic layer forms is drawn together by the molten plastic mass, and the thickness of the latter can be measured with the aid of a thin steel rod (needle). A rod freely lowered from above into the opening will stop at the upper boundary of the layer and, passing through it under slight pressure, will rest below against the semicoke. The determined thickness of this plastic layer, in millimeters, is the first indication of the coking capacity of coals. It has been established experimentally that good coking coals or coal blends should give a layer 20–22 mm thick. The second indication in determining coking capacity, according to L. M. Sapozhnikov, is the amount of shrinkage of the coal charge by the end of the experiment.

At the basis of the original method proposed by Professor I. V. Gebler (5) lies the concept that caking coals (in common industrial terminology), when rapidly heated, are capable of softening to such an extent that the plastic mass formed penetrates into the small pores formed by grains of sand. A distinguishing feature of this method is the small amount of time required to carry out the test. Into a thin-walled steel tube (wall thickness 0.5 mm, clear cross-section 1 square centimeter, and length 50 millimeters) with a perforated insert bottom (which, during the experiment, is covered with a thin layer of asbestos board), a charge of coal (about 1 gram) is poured. With the aid of a small hand press the coal in the tube is pressed, and onto its smooth surface there is poured, from a previously weighed boat or crucible with contents, sand in an amount of 3 grams (in the experiments of Professor I. V. Gebler and of mine, a fraction of washed and calcined quartz sand was used with grain sizes retained when sieving on a sieve with 144 openings per 1 sq. cm). The tube thus “charged” is fastened in an apparatus consisting of a round disk, at the top of which a coupling is fixed, and below a steel tube with a flared lower part for fastening the tube containing coal and sand. Through the coupling there is passed a steel

rod, having at the bottom a “plunger” that freely enters the tube, and at the top a device for setting a load of 1 kilogram. When the “charged” tube is introduced into an electric furnace with a constant temperature of 900°, the coal rapidly absorbs heat, melts, and penetrates into the pores between the sand grains. Upon subsequent solidification of the plastic mass that has penetrated into the sand, it binds and retains a certain quantity of it. After 5 minutes the experiment is stopped. The tube is removed from the apparatus. The remaining unbound sand is poured into a boat, and it is weighed again. From the difference in the weights of the boat with sand before and after the experiment, it is possible to determine what quantity of sand was bound by the softened coal. The quantity of sand bound by the coal, expressed in grams and multiplied by 100, Professor I. V. Gebler calls the “softness number” (M). The numbers (M) obtained by Professor I. V. Gebler’s method, no less distinctly than the long-duration tests by L. M. Sapozhnikov’s method, establish the differences between coals of different degrees of fusibility. The agreement of the numbers from two parallel experiments is very good. For illustration, Table 1 gives the data of proximate analysis,

Table 1.

plastometric indices and “softness numbers” for 8 coals of the Osinovsky and Prokopyevsky deposits of the Kuzbass. The proximate analysis and plastometric testing of these coals were carried out by the laboratory of the coke-chemical shop of the Kuznetsk Metallurgical Plant named after I. V. Stalin, while the softness numbers were determined by me.

The method of I. V. Gebler is taken as the basis of the present investigation. However, under the conditions of carrying out experiments to study the degree of softening of coals at various temperatures, it proved necessary to introduce certain changes into it. Thus, for example, in testing certain “fatty” coals at temperatures below 900° (600, 700°), they swell so strongly that the sand rises above the edges of the tube and spills out of it; furthermore, at these temperatures, in experiments with coals rich in tar, another abnormality is observed: part of the sand proves to be bound not as a result of the penetration of molten coal into it, but as a result of condensation and subsequent decomposition of tar in its upper layers.

To avoid the ejection of sand from the tube and to eliminate the possible influence on changes in the softening numbers of tar vapors condensing in the sand, the experimental conditions were changed as follows: 1) the coal and sand were interchanged, i.e., the sand was poured directly onto the bottom of the tube, and on top of it was placed a briquette prepared beforehand from a measured quantity of coal; 2) the insert bottom of the tube was made solid and was fastened from below, as tightly as possible, so that the gases and vapors of decomposition would not pass through it, but would be directed exclusively upward and would leave the tube through the gaps between the walls and the loosely fitting piston rod entering the tube.

Photographs 1, 2, and 3 give an idea of the apparatus required for carrying out the experiment. Photograph 1 shows: a) an electric crucible furnace; b) the apparatus for testing coal in assembled form, with the tube lowered into the furnace; c) a thermoregulator, making it possible to maintain a constant temperature in the furnace; d) a galvanometer for observing the temperature, measured by means of a thermocouple inserted into the furnace.

Fig. 1.

Photograph 2 shows a press for making coal briquettes. Photograph 3 shows the details of the apparatus: a) a disk with a hollow tube fastened in it, in the flared end of which the reaction tube “charged” with sand and coal is fixed; b) a rod, onto the lower end of which the “piston” is screwed (lying alongside), which freely enters the opening of the reaction tube and at the top of which there is a small disk serving as a support for the load; c) the reaction tube, next to which lies

insertable bottom and a pin for fastening the latter; d) a weight of 1 kg, with a hole drilled in the middle for placing it on the rod.

The experiment is carried out as follows. From a measured quantity of coal, a coal briquette is made by means of a press. The measured quantity of sand (approximately 3 grams) is poured into a porcelain crucible, and weighing is carried out. The insertable bottom is firmly fixed in the reaction tube, and then the sand from the crucible is poured into the tube. The coal briquette is lowered into the tube and placed on the sand. The “charged” tube is fastened in the lower part of the hollow tube of the apparatus, and the apparatus is placed on an electric furnace, which by this moment has the required temperature; the tube with the coal and sand is lowered into the furnace. Immediately after this the weight is placed on the rod. After the established time has elapsed, the apparatus is removed from the furnace and placed on a tripod. The tube is easily removed from the apparatus, and therefore there is no need to wait for it to cool. With the aid of forceps it can be detached from the apparatus, the bottom removed, and the sand poured into the crucible from which the sand for the given experiment had been taken, provided it has not been cemented by the molten coal. The difference between the weights of the crucible with sand before the experiment and after the experiment will show what quantity of sand was bound by the coal. For example, the weight of the crucible with sand before the experiment is 15.482 grams, after the experiment 13.865 g; the difference is \(15.482 - 13.865 = 1.617\); the softening number is 161.7, or, rounding off, 162. Hereafter all numerical characteristics of the degree of softening by the amount of sand bound by the coal that are given in the text and tables were obtained under these modified experimental conditions.

The “softening numbers” obtained by the method described above must be assumed to be connected, first of all, with the quantity of fusible soluble substances formed as a result of the decomposition of the coal. An idea of this connection is given by experiments with mixtures of different ratios of completely non-caking coal from the Aralichevo deposit and an extract obtained with naphthalene from coking coal of the Prokop’evsk deposit (Table No. 2). This extract is a mixture of substances (“bitumens”) capable, upon heating, of forming a liquid mass that readily flows out through a very narrow opening of a vessel. For the experiments the Aralichevo coal and the “bitumens” were taken in a state of grain fineness passing through a sieve with 900 openings per 1 cm².

Table 2

Two coals containing equal quantities of fusible substances may, of course, also have different “softening numbers.” In this case the difference in the “softening numbers” will be connected with the viscosity of the melt formed by the fusible substances. As an illustration, one may cite the results of experiments with a mixture of constant composition of the same Aralichevo coal and coal-tar pitch. Table No. 3 gives the “softening numbers”

for the mixture in a ratio of 1:1 at different experimental temperatures. As the temperature is raised, the fluidity of the coke increases, and the “softening numbers” increase.

Table 3

The thickness of the plastic layer in L. M. Sapozhnikov’s experiments has the property of additivity, which constitutes a great advantage of this method over others, since it makes it possible, from the magnitude of the layers of the components, to find easily the percentage composition of a mixture with a specified thickness of the plastic layer. The “softening numbers” do not possess this property, and this, naturally, creates enormous difficulties for using the method for practical production purposes.

The first task of the present investigation was therefore to find a way of expressing the degree of softening in such quantities as would make it possible to calculate a mixture of several components with a specified value. Experiments with mixtures of “bitumens” and Aralichev non-caking coal suggested expressing the degree of softening of coals by the quantity of “bitumens,” by comparing the “softening number” obtained for the coal with the percentage content of “bitumens” in the mixture. From the data of Table No. 3 a curve was constructed, for whose points the ordinates were the “softening numbers,” and the abscissae were the percentage amounts of “bitumens” in the mixture. From this curve, for any coal, according to its “softening number,” it is possible to determine the quantity, corresponding to this softening number, of fusible substances of a definite composition, or a certain “bitumen equivalent.” It goes without saying that this found quantity of “bitumens” may by no means correspond to the actual quantity of fusible substances formed in the coal during its decomposition, and is of a purely conventional character. In the present case the actual (unknown) percentage content in the coal of fusible substances “z,” in terms of the softening effect produced by it on the coal, expressed by the softening number “y,” is equivalent to the percentage content found from the curve in the given mixture of fusible substances “x.” Investigation of the curve determining the dependence of the softening numbers on the percentage content of bitumens in the mixture showed that it can be expressed by an equation of the form:

\[ \lg y = \frac{a}{x} + b \cdot \lg x + c, \]

where \(Y\) is the “softening number,” expressed, as before, in grams of sand bound by the coal when multiplied by 100; \(X\) is the percentage content of “bitumens” in the mixture; \(a\), \(b\), and \(c\) are constants.

Selecting from the experimental data three points of the curve and thus having the values \(X\) and \(Y\), by solving a system of three equations with three unknowns we find the values of the constants \(a\), \(b\), and \(c\). Then, substituting these found values of the constants into the equation, we find the value of \(Y\) for pro-

intermediate points. For the curve shown in Figure 4, the constants were determined from the points:

\[ x=100,\ y=300 \quad (298\ \text{from Table 2 rounded to }300); \]

\[ x=60,\ y=75 \quad (\text{from Table 2 rounded to }75); \]

\[ x=20,\ y=6. \]

The values of \(Y\) then obtained by calculation for 80 and 40% content of bitumens in the mixture are relatively close to those found experimentally. Table 4 gives the values of \(Y\) from experiment and those found by calculation with the constants \(a=10\), \(b=3\), and \(c=-3.625\). The values of the constants were found from the following conditions. The value of the “softening number” in the experiment with 100% bitumens was rounded to 300; the equation for this point is

\[ \lg 300=\frac{a}{100}+b\lg 100+c. \]

The value of the “softening number” in the experiment with a mixture of 60% bitumens and 40% non-caking coal was rounded to 75; the equation for this point is

\[ \lg 75=\frac{a}{60}+b\lg 60+c. \]

The value of the “softening number” in the experiment with a mixture of 20% bitumens and 80% non-caking coal was 6; the equation for this point is

\[ \lg 6=\frac{a}{20}+b\lg 20+c. \]

\[ (1).\ \lg 300=\frac{a}{100}+b\lg 100+c;\qquad (1).\ 2.4771=\frac{a}{100}+2b+c; \]

\[ (2).\ \lg 75=\frac{a}{60}+b\lg 60+c;\qquad (2).\ 1.8751=\frac{a}{60}+1.7782b+c; \]

\[ (3).\ \lg 6=\frac{a}{20}+b\lg 20+c;\qquad (3).\ 0.7782=\frac{a}{20}+1.3110b+c; \]

\[ (1).\ 247.710=a+200b+100c;\qquad (1).\ 247.710=a+200b+100c; \]

\[ (2).\ 112.506=a+106.682b+60c;\qquad (2).\ 112.506=a+106.682b+60c; \]

\[ (3).\ 15.564=a+26.020b+20c;\qquad (4).\ 135.204=93.318b+40c. \]

\[ (2).\ 112.506=a+106.682b+60c \]

\[ (3).\ 15.564=a+26.020b+20c \]

\[ (5).\ 96.932=80.662b+40c. \]

\[ (4).\ 135.204=93.318b+40c \]

\[ (5).\ 96.932=80.662b+40c \]

\[ (6)\quad 38.272=12.656b \]

\[ b=\frac{38.272}{12.656}\simeq 3. \]

From (5)

\[ 40c=96.932-80.662\times 3=96.942-241.986=-145.044 \]

\[ c=\frac{145.044}{40}\simeq -3.625. \]

From (1)

\[ a=247.710-600+362.500\simeq 10. \]

Substituting into the equation the found values of the constants \(a\), \(b\), and \(c\), we find the value of \(y\) for \(x = 90\), \(x = 80\), \(x = 70\), \(x = 50\), \(x = 40\), and \(x = 30\).

For example:

\[ \lg y=\frac{a}{x}+b\lg x+c=\frac{10}{50}+3\lg 50-3.625=0.2+5.097-3.625=1.672. \]

Hence \(y = 47\).

For a mixture of 80% bituminous coals and 20% non-caking coal, the calculated value is \(y = 162\), while directly from experiment the softening number is 158; correspondingly, for a mixture of 30% bituminous coals and 70% non-caking coal, the calculation gives \(y = 27\), and experiment gives 24.

Table 4

Table No. 5 gives the results of experiments with coal mixtures. In this case, from the “softening numbers,” the “equivalents” were found with the aid of the curve; from the equivalents of the components, the calculation was made according to the mixing rule for the “equivalent” of the mixture, and from the “softening number” the same “equivalent” was found with the aid of the curve. Comparing one with the other shows that the equivalents possess the property of additivity. For example, experiment gives for coal PZh—1 from Table No. 5 a softening number of 181, to which, according to the curve in Figure 4, there corresponds the equivalent 84; correspondingly, for coal PS—1 the softening number is 19, and the equivalent 34. A 1:1 mixture should give an equivalent equal to \(84 \cdot 0.5 + 34 \cdot 0.5 = 59\); in the experiment with this mixture a softening number of 67 was obtained, to which, according to the curve, the equivalent 57 corresponds, i.e., a value close to that found by calculation.

Table 5

II. Caking Properties and the Nature of the Decomposition of Caking Coals

The most widespread point of view on the question of the cause of the caking of coals is the theory according to which caking coal contains, in ready form, fusible and soluble substances, and sometimes may even consist almost entirely of them. Under the influence of heating, part of the organic mass of the coal melts; another part dissolves in it; and in this mixture of molten and dissolved substances the remaining infusible and insoluble part of the organic mass is dispersed. Such a conception—that caking coals possess fusible substances and non-caking coals do not—of course explains most simply the difference in their behavior upon heating.

As regards differences in the coking results of two identical, well-caking coals, these, in accordance with this theory, could be attributed to differences in the composition of the fusible substances, to, let us say, different viscosity of the melt, and so on. However, having made such assumptions about the presence of fusible substances in caking coals, the investigator feels the need for experimental proof and inevitably arrives at the idea of trying to isolate them from the coal and study them in detail. The application to the study of coals of methods of separation analysis, alongside the broader aims of elucidating their chemical nature in general, also pursued this narrower task—the elucidation of the essence of the process of caking and coking. Since the most diverse solvents and different conditions were used in this work, investigators arrived at conclusions sharply differing from one another. The greatest interest in work on the extraction of coals is presented, one must suppose, not by the conclusions—which subsequently, under the pressure of newly discovered facts, may change—but by the factual side of the matter. Comparison of the facts obtained and observations made should be the best path for judging the correctness of one conclusion or another.

Let us consider first of all the results obtained in the extraction of hard coals without heating, using such an excellent solvent in all respects as benzene. It is a sufficiently strong solvent, first; neutral, second; and is easily removed from the extract by distillation, third. It turns out that extraction of coal with cold benzene, at its usual standard laboratory fineness, gives a negligible yield of extract. Thus, for example, S. A. Gusinskaya and E. I. Beskina (Dnepropetrovsk Metallurgical Institute) (6), using this method, extracted: 0.16% of the weight of the combustible mass from fat coal (PZh), 0.06% in one case and 0.09% in another from coking coal (K). The same coals, when extracted with benzene in a Soxhlet apparatus, i.e., with hot benzene, showed considerably higher yields, namely: 0.33%—PZh coal and 0.23%—K coal. Some investigators consider one of the reasons for such meager extract yields to be the relatively small surface of the coal washed by the solvent. This explanation is supported, among other things, by the fact that F. Fischer, by grinding coal to a grain size of 1 micron, was able, upon extraction with benzene in a Soxhlet apparatus, to obtain yields twenty times greater than those obtained under the same conditions with standard laboratory grinding. And nevertheless, if we assume that cold benzene extracts from coal subjected to such grinding 20 times more than in the experiments of Gusinskaya and Beskina, the yield for coal K would still not exceed 2% of the weight of the combustible mass.

Quite different results are given by benzene as a solvent under conditions of coal extraction in an autoclave, for example at a pressure of 55 atmospheres, allowing the temperature to be brought to 285°. Applying these conditions, F. Fi—

1. Izvestiya TII, vol. 57, issue 2.

Table 6

Fischer (7) charged into the autoclave coal in grains the size of peas and nevertheless obtained yields of extractable substances many times higher than in the extraction of finely crushed coal in a Soxhlet apparatus. Table No. 6 gives data characterizing the coals extracted by F. Fischer, the yields and composition of the extracts according to solubility in petroleum ether (oily bitumen, soluble in petroleum ether, and solid bitumen, insoluble in petroleum ether), the melting temperatures of the extracts, and the decomposition temperatures of the solid bitumens. From these data one can establish a certain dependence of caking capacity on the total yield and composition of the extract, and on the decomposition temperature of the solid bitumen. Thus, for example, the coal with the poorest caking capacity gave the smallest yield of extract; in weakly caking coals the extracts contain a higher percentage of solid bitumen, the melting temperatures of the extracts are correspondingly higher, and the decomposition temperatures of the solid bitumens are lower than those of extracts from well-caking coals.

The explanation for the increased yields during extraction in an autoclave may be either the effect of high pressure, owing to which the solvent penetrates through the pores deep into the coal grains, or the effect of high temperature, owing to which decomposition of the coal and formation of readily soluble substances occur, or the combined effect of both factors. It should not be forgotten that those investigators who tried to select as low-boiling a solvent as possible, or carried out extraction with a cold solvent, did so primarily out of fear of possible changes in the composition of the organic mass of the coal under the influence of heating. Before extraction, F. Fischer tested the coals to determine the decomposition temperature and found that, in this case, the lowest point of the beginning of decomposition was a temperature of 310°. Since in the experiments the temperature did not exceed 285 (25° lower), decomposition even of the most thermally unstable of the coals subjected to extraction was excluded. It would thus seem that fears of possible changes in the composition of the organic mass of the coal were unfounded, that the extracts represent substances present in the coal in finished form, and that the increased yields should be attributed exclusively to favorable conditions for the action of the solvent. But the question arises: must every change in the composition of the organic mass of coal necessarily be accompanied by the liberation of colored vapors and gases, i.e., by visible decomposition of the coal, which served F. Fischer as the criterion for establishing the point at which decomposition begins?

If coal is subjected to more or less prolonged heating under conditions excluding its oxidation and visible decomposition, then it is possible—indeed, it turns out—to obtain during extraction significantly more soluble substances than from coal not subjected to such treatment. W. Fuchs (8) cites the results of Harger’s experiments carried out in this way, when the amount of extract increased sevenfold. Coal containing 36.7% volatile matter was subjected by the said investigator several times to thermal treatment in a stream of nitrogen within the temperature range 230–340°, and after each heating experiment it was extracted; the yield of extract increased from 1.2% in the original coal to 8.5% in the coal subjected to thermal treatment.

“Fat” coal from the Osinovsk deposit of Kuzbass was subjected by me to sequential extraction with alcohol-benzene: first without any treatment, then the residual coal was extracted again after thermal treatment in a stream of nitrogen for 4 hours at 300°; the second residual coal was again extracted with alcohol-benzene after treatment for 4 hours at 350°; the third residual coal, without new treatment, was extracted with pyridine; the fourth residual coal in

in turn, without thermal treatment, was subjected to extraction with phenol. The results of the experiments are given in Table No. 7.

Table 7.

The proximate analysis of this coal showed: laboratory moisture 1.66%, ash on the dry substance 9.8%, volatiles on the hot mass 28.71%.

One curious detail should be noted in passing: the pyridine extract dissolved in alcohol-benzene to the extent of 75%, the phenol extract dissolved in alcohol-benzene to the extent of 22%, and in pyridine to the extent of 100%. It must therefore be assumed that, when the coal was treated with pyridine and phenol, new quantities of substances soluble in alcohol-benzene were formed.

The same coal, subjected to thermal treatment for 1 hour 10 minutes with a gradual rise in temperature during this time from 200 to 375°, when extracted with alcohol-benzene, gave the solvent 4.47%. Extraction with pyridine after a repeated identical treatment gave a further 6.37%. In the first case the sum of the alcohol-benzene and pyridine extracts amounted to 12.56%; in the present case, to 10.84%. There is every reason to attribute the higher yield in the first case to the longer thermal treatment.

The alcohol-benzene extract of the second series of experiments dissolved to the extent of 57% in petroleum ether and to the extent of 89% in benzene. The pyridine extract of this same series of experiments dissolved to the extent of 25% in petroleum ether and to the extent of 35% in benzene. In the total extracts, the substances soluble in benzene amounted to 6.2% out of 10.84%, while the substances soluble in petroleum ether, in relation to the substances soluble in benzene, amounted to 66%, i.e., approximately the same ratios as those obtained by F. Fischer for well-caking coals when extracted with benzene under pressure. To this it must be added that the conditions of thermal treatment excluded the possibility of decomposition of the coal with the evolution of colored vapors and gases. Thus, for example, as a result of treating the coal for 4 hours at 300° and then (without extraction) for another 4 hours at 350°, the loss in weight was 2.7%, while in the absorption apparatus the gain from the carbon dioxide and moisture evolved during heating amounted, relative to the weight of the coal, to 2.58%. When treated for one hour at 375°, the loss in weight of the dry coal was 1.03%, and the gain in the apparatus was 1%.

The facts cited above are, of course, difficult to reconcile with the proposition that the composition of the organic mass of coal remains unchanged when it is extracted under conditions of high pressure and temperature. These facts are convincing proof of the contrary. Therefore, in explaining the cause of the increased yields of extractable substances with preliminary thermal treatment, investigators inevitably come to the conclusion that this is the result of the breakdown of complex high-molecular compounds contained in the coal into simpler, readily soluble ones. Within certain temperature limits this breakdown can occur without the splitting off of volatile liquid products and gases; subsequently, as the decomposition process deepens, vapors and gases appear.

According to the results obtained by A. Pott and his coworkers

studies, the moment of the beginning of visible decomposition of coal corresponds to the greatest yield of soluble substances.

A. Pott, H. Broche et al. (9) carried out experiments on the extraction of various coals in an autoclave with a mixture of tetralin with primary phenols and naphthalene at temperatures: a) below the coal-decomposition temperature, b) corresponding to the coal-decomposition temperature, and c) lying above the coal-decomposition temperature. These investigators regarded as the temperature point of coal decomposition the moment of a sharp increase in the amount of vapors and gases liberated from the coal. Table No. 8 presents the results of extraction of five coals at the indicated temperatures.

Table 8.

From the results of Foxwell’s investigations (3), in determining the temperature interval of the plastic state of coals, it can be seen that the temperature at the beginning of coal melting lies considerably above the temperature at the beginning of visible decomposition. Thus, for example, of 10 coals investigated, six had a temperature at the beginning of melting of 400° and higher. The temperature of the melting maximum in his experiments ranged from 409° to 515°, and the temperature of the end of the plastic state ranged, for different coals, from 450 to 620°.

As is known, the point of the beginning of melting by Foxwell’s method is determined by the sudden sharp increase in the resistance to the inert gas flowing through the coal; the maximum resistance corresponds to the maximum of melting, and the end of the plastic state is determined by the decrease of the resistance to its initial value in the unmelted coal. Table 9 gives the results of experiments I performed on the plastic state of three coals from different Kuzbass mines by Foxwell’s method.

Table 9.

The decomposition of coal, in physicochemical terms, is the breakdown of a solid body—but one which even externally is not always homogeneous, and in fact apparently consists of a very large number of diverse compounds—into gaseous, liquid, solid meltable, solid soluble, and solid infusible and insoluble substances.

substances. Since it is not possible to take exact account of the changes occurring in the composition of the organic mass of coal within a given temperature interval during such decomposition, it is also impossible to give an exact quantitative assessment of the degree and character of decomposition. One can judge only relatively the degree of decomposition of a coal by the amount of extract of soluble substances or by the amount of gaseous and liquid products it releases at a given temperature, and the character of decomposition by the composition of the extracts or distillation products. Nevertheless, even such approximate accounting makes it possible to establish a definite dependence between the degree and character of decomposition, on the one hand, and the caking capacity of coals, on the other.

Anthracites and dry long-flame coals represent two extreme types in the series of hard coals. What they have in common is the absence of any appreciable ability to melt and cake, but this common property is due to different causes. Anthracites yield almost no extract, or at any rate yield very small amounts of extract. Thus, according to F. Fischer (7), anthracite coal extracted by him with benzene under pressure showed an extract yield of only 0.5%. This extract consisted mainly of oily bitumen. Noticeable decomposition of coals of this type begins at a very high temperature, and the amount of volatile decomposition products is then very small. In short, anthracites are characterized by a low degree of decomposition, which does not provide the necessary amount of products responsible for melting and caking. In contrast to anthracites, in long-flame coals decomposition with a large release of volatile products begins at a relatively low temperature. As may be judged from the decomposition temperatures of solid bitumens from the extracts of the first two coals in Table No. 6, the formation of stable high-melting and soluble substances is not ensured here. Decomposition of the coal proceeds deeply from the very beginning, with predominant formation of volatile products. Between these two extreme types are located the most diverse types of coals with different degrees and characters of decomposition, and with different degrees of melting and caking.

Extremely interesting data were obtained by Foxwell (3) in studying the decomposition of one sample of caking coals. The coal was heated gradually to a certain final temperature, and the yields of decomposition products were determined. Table No. 10 presents the experimental results for five final temperatures, indicating the yields of the principal decomposition products.

Table 10.

Comparison of the results presented in the table with the results of experiments studying the plastic state makes it possible to establish differences in the character of decomposition of caking coals during the melting period and in the subsequent period of completion of the coking process. At final

at temperatures of 400, 500, and 600°, i.e., during the period of the plastic state, the captured decomposition products are dominated by liquids over gaseous products (if the weight ratios of the former and the latter are compared). With the end of the period of the plastic state, the release of new quantities of tar ceases, and thereafter only gaseous substances are split off from the solid residue.

Thus, three mutually distinct periods of coal decomposition may be noted. The first period is the period, if one may put it so, of hidden decomposition; its beginning is difficult to determine, and its end may be considered the beginning of visible decomposition. Some idea of the degree and character of coal decomposition in this period can be formed from the quantity and composition of the extracts. The second period is the period of the plastic state, up to and including the formation of semicoke. The third period is that of residual gas evolution. The difference in the character of decomposition between the last two has been noted above. All the principal processes decisive for coke quality undoubtedly occur in the second period. However considerable the degree of coal melting may be, strong cementation in the coke residue cannot be ensured if, during distillation and decomposition from the maximum of melting to the formation of semicoke, the losses of the substances responsible for coal melting are very large; but even in those cases where the indicated losses do not reach threatening proportions, the strength of the bonds in coke made from coal of a relatively high degree of melting is often severely disturbed by changes in the volume of the coal charge during the melting period.

Experiments to determine the degree of softening of coals at various furnace-heating temperatures, carried out by me with coals from different deposits of the Kuzbass, give some indication of the common source of those negative phenomena which create a threat to the mechanical strength of coke during the period of the plastic state of coal. Experiments to determine the “softening numbers” and the corresponding “equivalents” were conducted by the method described in the first part of the present work, at furnace temperatures of 500, 600, 700, and 800°.

First of all, the following curious fact may be noted. According to the data of Foxwell’s experiments, the maximum melting of coals lies at approximately 500°; in my experiments all coals give higher softening numbers at 600° than at 500°. The explanation is not difficult to find. As was noted above in considering the results of the investigation of caking coal, after 500° the formation of considerable quantities of liquid substances is still possible, contributing to an increase in the degree of coal softening. In contrast to the gradual and relatively slow heating of coal in experiments by Foxwell’s method, in the present case the coal is introduced into a medium having the final temperature; heating proceeds in a concentrated manner; a small charge of coal in a very short interval of time is heated almost simultaneously to the same temperature at all points; melting also proceeds in a concentrated manner, and the liquid products formed do not have time to volatilize from the coal before the process is completed.

Next, attention is drawn to the difference in the behavior of coals of different industrial grades in experiments at temperatures of 700 and 800°. Coking coals show, at these temperatures, only very slight changes toward a decrease or increase in the numbers compared with the temperature of 600°. The fat Osinovskie and Leninskie gas coals invariably show a considerable decrease in the numbers at 700 and at 800°.

The test results are given in Table 11.

In the successive course of the decomposition process, in the second period, i.e., in the period of the plastic state, there are two stages. One is the stage

Table III

decomposition of some part, more or less significant for well-caking coals, of the compounds contained in the coal, with the formation chiefly of compounds that are liquid at these temperatures and soluble; it continues up to the maximum of melting. The other is the stage of formation of the solid residue. It is accompanied by the distillation of part of the liquid compounds and by a new decomposition of the compounds responsible for melting into gaseous, liquid, and solid residue.

One may suppose that in some coals there is a considerable quantity of such thermally unstable compounds, which under high-temperature conditions decompose at once and very rapidly—to volatiles, i.e., liquid low-boiling and gaseous substances, on the one hand, and a solid residue, on the other, bypassing the first stage. In this way one can explain the lowering of the “softening numbers” of the Osinovskie and Leninskie coals at experimental temperatures above 600°. Prolonged heating of coal at temperatures below the temperature of visible decomposition can also, evidently, lead to decomposition of these unstable compounds, with formation of liquid, readily boiling “bitumens” (as is shown, for example, by the increase in the yield of alcohol-benzene extracts during preliminary thermal treatment, with oily bitumen predominating in them) and of a non-melting, insoluble residue. In any case, the experiment of heating for 8 hours at 300° under completely identical conditions, and the subsequent testing for “softening numbers,” of the “fat” coal from Osinovskoe and the “coking” coal from Prokopevsk showed that they underwent entirely different changes as a result of the treatment. After treatment, the “fat” coal lowered its softening number at an experimental temperature of 600°, while its softening number at 800° remained unchanged. The “coking” coal, on the contrary, raised its softening number at 600°, while its softening number at 800° decreased somewhat, but only slightly.

These same thermally unstable compounds are apparently the least stable also during oxidation of coals. New samples of the same two coals were kept for a long time in a drying oven at a temperature of 75–80°, and then their softening numbers were again determined at 600 and 800°. It turned out that, during oxidation, the lowering of the softening numbers of the fat coal again occurs most strongly at 600°.

Table No. 12 presents the results of experiments on thermal treatment (without extraction) and oxidation of typical representatives of the Osinovskie fat and Prokopevsk coking coals. The softening numbers given in the table are the arithmetic means of two parallel experiments in all cases without exception.

Thus, these unstable compounds, present in considerable quantity in some well-caking coals, are apparently the common source of all those negative phenomena with which the deterioration of coke strength is usually associated. They can impede shrinkage as a consequence, on the one hand, of an increase in viscosity during the sudden decomposition of the initially melting and soluble substances formed from them, and, on the other hand, because of the rapid accumulation of large amounts of vapors and gases during this decomposition. And since shrinkage will not occur when there is a considerable loss of substance by the melting coal, there will also be no necessary compaction of it, which will undoubtedly be reflected in the strength of the bonds in the coke. They, on the contrary, can cause sudden, sharp shrinkage as a result of large losses of volatiles during excessively rapid decomposition. And since here, at the same time, formation of the solid residue will be proceeding just as rapidly, it is not difficult to imagine what breakage this sharp shrinkage will produce in it. Finally, they may be the cause of weakening of the bonds in the coke as a result of large losses of binding material. It is understandable—

Table 12.

therefore, how important it is to take their presence in coal into account, even if by such a perhaps imperfect method as recording the difference in equivalents at 600° and 800°. This difference, in any case, may serve as a certain characteristic of the greater or lesser stability of the coal during the period of its melting.

In the first part of this work it was shown that the values of the “equivalents” in blends possess the property of additivity at a test temperature of 800°. The same property is exhibited by them, as the experiments showed, also at a temperature of 600°, as can be seen from Table No. 13. But in this case the difference of the equivalents between the two indicated temperatures also possesses the property of additivity, and thus its consideration in preparing blends is facilitated.

Table 13.

In fact, if there are two components \(A\) and \(B\), whose “equivalents” respectively for \(800^\circ\) and \(600^\circ\) are \(X_A^{800}\) and \(X_A^{600}\) for one, and \(X_B^{800}\) and \(X_B^{600}\) for the other, and if from them a mixture is prepared consisting of \(a\%\) of component \(A\) and \(b\%\) of component \(B\), which will give the “equivalents” \(X_C^{800}\) and \(X_C^{600}\), then, assuming additivity of the “equivalents” in mixtures, we have

\[ X_C^{800}=\frac{a}{100}\cdot X_A^{800}+\frac{b}{100}\cdot X_B^{800} \]

and

\[ X_C^{600}=\frac{a}{100}\cdot X_A^{600}+\frac{b}{100}\cdot X_B^{600}; \]

then:

\[ X_C^{800}-X_C^{600} = \left( \frac{a}{100}X_A^{800} + \frac{b}{100}X_B^{800} \right) - \left( \frac{a}{100}X_A^{600} + \frac{b}{100}X_B^{600} \right) = \frac{a}{100}\left(X_A^{800}-X_A^{600}\right) + \frac{b}{100}\left(X_B^{800}-X_B^{600}\right), \]

that is, the difference of the “equivalents” of the mixture is equal to the difference of the “equivalents” of the first component, multiplied by its share in the mixture, plus the difference of the equivalents of the second component, multiplied by its share in the mixture.

With more than two components in the mixture, the difference of its “equivalents” will be the sum of the difference of the equivalents of the first component multiplied by its share, of the second by its share, of the third by its share, and so on.

III. Industrial conditions of coking coals of different degrees of softening and thermal stability.

The process of transformation of coal into coke in the chamber of a coke oven presents an extremely interesting picture. What first of all attracts attention here is the character of heat propagation in the coal charge throughout the entire coking period. When temperature measurements were made at various points of the coal charge across the width of the chamber, it turned out that in the layer of coal immediately adjacent to the oven wall the temperature rises very rapidly; in layers situated closer to the axis of the chamber, for many hours no noticeable rise in temperature is observed; when, however, the temperature begins to rise at these points, it proceeds at a rate little different for layers farther from the wall and for layers situated closer to it. In a certain temperature interval the rate of temperature rise at the center of the chamber even surpasses that in the coal layer located at \(1/4\) of the chamber width from the wall. The diagram of Hielgenstock (10) given in Figure 5 clearly shows these features of heat propagation in the coal charge of a coke oven.

In contact with the incandescent walls of the oven chamber, the outer layers of the charged coal rapidly receive heat from them, and decomposition processes begin, leading to the formation, along the entire length and height of the oven, of a layer of plastic mass from molten, dissolved, and solid substances dispersed in the solution. Subsequently, with further rise in temperature, in one zone of this layer, facing the wall, the decomposition process leads to the formation of a solid residue, while from the oppo-

from the opposite side, as the temperature rises, it, acting on the adjacent layer of coal, also causes processes of decomposition and melting in this latter layer. Contracting on one side and continuously increasing on the other, the layer of plastic mass gradually moves from the wall toward the middle of the oven. This plastic layer, constituting, as it were, a thermal barrier, prevents the uniform spread of heat through the entire thickness of the coal charge and accounts for the above-noted peculiarities in the propagation of heat within it.

Such a pattern of heat propagation in the coal charge is a phenomenon of a positive character. If the spread of heat into the depth of the coal charge were not arrested by the plastic seam forming at the very beginning, and if in the layers remote from the wall there occurred a continuous and, it goes without saying, very slow rise in temperature, it would be impossible to obtain in the oven a more or less homogeneous coke; indeed, in general it would be difficult to find coals suitable under such conditions for coking purposes. The point is that, as experience shows, certain—

Fig. 5.

—coals, under very prolonged heating, begin already at 300° to release liquid oily substances; at 350° this release becomes quite noticeable, and at 400° considerable quantities of oil resembling oily bitumen are released. The distillation of liquid products occurring under conditions of very slow heating and decomposition will cause a retardation of the processes of melting and dissolution in the coal; at the same time the process of decomposition will deepen, and part of the soluble substances will decompose into volatile products and a coke residue. In the final analysis, an excessively drawn-out heating process may lead to a complete loss of the ability to melt and sinter even in coal known to be well caking.

The facts and considerations set forth in the preceding chapter of the present work show that the temperature interval 300–600° is the most responsible and important one for the coking process. In this temperature interval there occurs the breakdown of complex organic compounds with the formation of substances responsible for melting; in its middle the very process of formation of the plastic mass from coal takes place; and at its end—the formation of semicoke. It is natural, therefore, that the greatest interest should be attached precisely to the rate of temperature rise in this temperature interval.

Under production conditions, as is known, the duration of coking is connected chiefly with three factors: the moisture content of the coal being coked, the heating temperature of the walls, and the width of the oven chamber. With little difference in the moisture content of the charged coals and in the heating temperature of the wa—

of the furnace, the coking period is related only to the width of the oven chamber. R. A. Mott and R. V. Wheeler (1), on the basis of data from various authors, derive average time indices for the temperature interval 300–600° at the center of the charge and at \(1/4\) of the distance from the wall across the chamber width for coke ovens with different chamber widths and different lengths of the coking period, according to Table 15.

Table 15

This corresponds to a heating rate of: \(1^\circ/\mathrm{min.}\), \(1.7^\circ/\mathrm{min.}\), and \(2.5^\circ/\mathrm{min.}\) in the coal layer located at \(1/4\) of the chamber width from the wall; \(1.4^\circ/\mathrm{min.}\), \(2^\circ/\mathrm{min.}\), and \(3.3^\circ/\mathrm{min.}\)—at the center of the coal charge.

When comparing the data on the rate of heat propagation in the coal charge of ovens of different chamber widths with the considerations set forth above concerning the influence of slow heating on the decrease in the degree of coal fusion, the following fact becomes understandable: in wide chambers, as a rule, coals and charges that are “fatter,” i.e. of a high degree of fusibility, yield better coke, whereas in narrow chambers coals and charges that are relatively “lean,” i.e. with a reduced degree of fusion, are successfully coked.

The aforementioned authors, who carried out extensive coking experiments under production conditions on a great variety of coals and mixtures of them, arrive in this respect at the following conclusions: “Considering a series of samples obtained from well-caking coals with a high content of volatile matter and their mixtures, we see that stronger coke was obtained in wide ovens, whereas with poorly caking coals with a high content of volatile matter, stronger coke was obtained in narrow ovens” 11.

The broadest and most complete investigation of the coking capacity of coals of the Kuznetsk Basin under production conditions was carried out in 1931 by an operational group of Soyuzkoks at the Kemerovo coke plant, in the ovens of the 3rd battery of the obsolete O. Piette system without regeneration (12). The average chamber width of these ovens was 375 mm, and the coking period 16–18 hours. Preliminary testing of coals and charges was performed by coking small coal samples in iron boxes (500 × 350 × 400 mm in size), placed in the coke oven before it was charged with coal. The strength test of the box-coking coke was carried out in a drum of reduced dimensions, compared with the standard, with 25 kilograms of coke pieces larger than 40 mm loaded into it after screening. In this way, in particular, coals from various seams of the Prokopyevsk, Osinovsk, and Leninsk deposits were tested without any additions of other coals.

The coal of seam III Internal of the Prokopyevsk deposit is a typical coking coal. Experimental coking of it without the participation of other coals, carried out at another time at the same Kemerovo installation, showed that it gives coke of unusually high strength, with a residue in the drum of 350 kilograms. The external feature distinguishing coke from the coal of the Gorely seam from coke from the coal of seam III Internal is its lower fusedness, greater

heterogeneity. The granular structure of cokes, as the result of a relatively low degree of fusion, is in general a distinctive feature of cokes from coals of the PS type, which are relatively poorly caking. The appearance of cokes from the coals of the Osinovsk and Leninsk deposits has much in common. In the report of the operational group, coke from the coals of the Osinovsk deposit is characterized as follows: “The coke obtained in coking Osinovsk coals alone in iron boxes was silvery-light, fine-porous, but with deep longitudinal cracks, being somewhat similar to the coke of Leninsk coals.” In the report preserved at the Kemerovo plant on experimental oven coking carried out in January 1927 by Prof. Rubin, there are data from drum tests of cokes from seams 15 and Elbansky of the Osinovsk deposit when coked separately; in the first case—the drum charge was 257, in the second—230 kilograms. The coke is characterized as follows: “coke with great longitudinal fissuring, easily breaks into nut.”

Great fissuring in pieces of coke is inherent in both Osinovsk and Leninsk coals and is the reason for their easy disintegration during testing in the drum into small pieces below 40 mm in size.

The common character of the phenomenon leading to reduced strength of coke from the well-caking coals of the Osinovsk and Leninsk deposits gives grounds for assuming the presence of a common source of this phenomenon as well. Confirmation of this point of view may be the fact that elimination of the cause producing the reduction in coke strength in these coals is again achieved by a common method. The coals of the Leninsk deposit, investigated by coking in boxes, gave, when tested in the small drum, a very low residue—16.5% and less; the coals of the Osinovsk deposit showed strength, by residue in the drum, from 50 to 62%; the coal of the Gorely seam of the Prokopyevsk deposit gave a residue in the drum of about 48%. When 45% of coal from the Gorely seam was added to a mixture of different seams of the Osinovsk deposit, taken in an amount of 55%, the residue in the small drum of coke from such a charge was 68%. A charge of 50% mixture of different seams of the caking coals of the Leninsk deposit and 50% of the Gorely seam gave coke which showed a residue in the drum of 66%. A mixture of 30% Osinovsk, 25% Leninsk, and 45% of the Gorely seam gave coke with a residue in the drum of 51%.

This character of behavior of the indicated coals was also confirmed when testing cokes from their mixtures with coal of the Gorely seam in the large drum. A mixture of 55% Osinovsk and 45% of the Gorely seam, in oven coking, gave coke with a residue in the drum in one case of 293 kg, in another of 312 kg. A mixture of 55% Leninsk coals and 45% of the Gorely seam gave coke with a residue in the drum—in one case of 292 kg, in another case of 312 kg; while a mixture of 50% Osinovsk and 40% of the Gorely seam, with the addition of only 10% Leninsk coals, already gave coke with a strength of only 283 kg of residue in the drum.

Analyzing the plastometric indices of coal caking capacity proposed by L. M. Sapozhnikov, G. L. Stadnikov comes to the conclusion that the ordinates of his classification diagram, along which the thickness of the plastic layer is plotted, “express the content of fusible substances in coals” (3). Comparison of coals with different thicknesses of the plastic layer in terms of the fusion and homogeneity of the coke residue confirms the correctness of this conclusion.

The same may also be said with respect to the comparison of coals by thickness of the plastic layer and softening numbers (Table 1).

But from this follows the conclusion that the plastic layer does not in all cases represent the same state of softening of the coal or an identical

degree of melting. The plastic state of coal must be determined by the presence in it, at a given temperature, of a minimum of substances that are in the molten and dissolved state. If, under the given temperature conditions, such a minimum of molten and dissolved substances is not formed, then no plastic layer is formed either. And, indeed, one of the coals I studied, which yielded a very poorly caking button, when tested in L. M. Sapozhnikov’s apparatus in the laboratory of the Coke-Chemical Shop of the Kuznetsk Metallurgical Plant (KMZ), showed a plastic layer equal to 0.

The results of Foxwell’s experiments on the gas permeability of coals during their transition into the plastic state convince one that the plastic layer of one and the same coal cannot, at all points across its width, represent the same degree of melting. That part or zone of the plastic layer of the charge in the coke oven which is adjacent to the coal not yet deeply affected by the decomposition process will evidently be at that degree of softening which is characterized by a minimum of substances in the molten and dissolved state. The pores between the individual coal grains in this zone are not closed; the vapors and gases of decomposition are removed relatively easily in the direction of least resistance to their motion, i.e., toward the center of the charge. That part or zone of the plastic layer which is adjacent to the semicoke is in the state of a hardening process that has begun. Here open pores and cracks are being formed; the path of least resistance for the vapors and gases of decomposition runs from here toward the wall through the semicoke and coke. The third zone of the plastic layer, situated between the two outer zones, is the zone of maximum softening of the coal, the zone of the maximum amount of fusible and soluble substances formed during decomposition. For coals of a high degree of melting, the escape of vapors and gases in this zone is impeded, and they can develop considerable pressure in order to escape toward the cold or hot side of the coal charge. The pressure developed by vapors and gases in this zone of maximum melting may play a positive role, compacting the plastic mass before the formation of semicoke and thereby promoting the creation of stronger bonds in it, when it is moderate. However, if it is very great, it may lead to sharply negative phenomena in the sense that, by pressing the coke tightly against the walls, it will make its pushing from the oven difficult, in the first place, and, as a result of compression, will cause numerous cracks in it, in the second. Especially dangerous, both in this zone and in the zone of incipient hardening, is a sudden sharp increase in the vapors and gases of decomposition, which is connected with the formation in coals of fusible or soluble thermally unstable substances. A sudden sharp increase in vapors and gases in the zone of maximum melting may create pressure of enormous force, with all the consequences resulting from this. A sudden sharp increase in the evolution of vapors and gases in the zone of incipient hardening will cause an equally sharp acceleration of this process, as a result of which the formation of open pores will cease, and the vapors and gases of decomposition will seek an outlet by forming new cracks in the forming semicoke.

Since the transition of coal into the plastic state is connected with the decomposition of coal and since, in turn, the character of decomposition may change with changes in heating conditions, indicators cannot be established independently of the operating regime of the coke oven. In the present case as well, the degree of softening of coal, expressed in “bitumen equivalents” at 800°, and the degree of thermal stability during melting, expressed as the difference of “equivalents” at 800 and 600°, could receive a numerical expression of those pre—

properties, which characterize coals and blends capable of yielding coke with a high drum index, only in relation to definite production conditions. The best way to verify the suitability of these indices in general, and to establish numerical characteristics for a given regime in particular, would be a method of prolonged and extensive experimental coking under production conditions of the most varied coals, so that it would be possible to compose both very poor and very good blends according to the indices, alongside normal blends determined by the quantity and quality of the coals supplied.

For understandable reasons, the possibility of such verification under production conditions was excluded for me. The only thing that could be done was to conduct tests, over a certain period, of the coals arriving at the coke plant of KMZ and to test the current charges of this plant in the summer of 1936.

From July 20 to August 5, 1936, in the laboratory of the coke shop of the Kuznetsk Metallurgical Plant named after I. V. Stalin, with the help of student trainees from Irkutsk State University, I tested 32 coal samples arriving at the shop from the Osinovskii, Prokop'evsk, and Sudzhensk mines, and 22 samples of plant charges. The participation of coals of different grades in these charges, on average for the indicated period, was as follows: PZh — 30%, K — 45%, and PS — 25%.

Tables 16 (coals) and 17 (charges) give the numerical characteristics of the caking indices proposed by the author of the present work. It was not possible to determine more or less accurately the quality of the coke by the drum test from each charge separately; the fluctuations of the drum test of coke over these days were from 290 to 311 kg.

The coals and charges listed in the tables are plotted on the diagram of Figure 6 in the system: \(E\) — “equivalents of softening numbers” at an experimental temperature of 800°, \(D\) — the difference (positive or negative) of the “equivalents” at experimental temperatures of 800° and 600°.

In Table 16 the coals, beginning with No. 9 through No. 25 inclusive and, in addition, Nos. 29 and 32, are marked by the mines as “K,” i.e., coking coals. If one is guided strictly by the data of the plastometric indices of the KMZ coke-shop laboratory, then not one of all the coals presented in Table 16 can be recognized as coking. Moreover, according to these data, the coals marked by the mines as K appear very heterogeneous, differ by great instability in their caking characteristics, etc. In contrast to this unappealing picture given by plastometric analysis, the characteristics by “equivalents” are in complete agreement with the mine markings, with the exception of coals Nos. 24, 25, 29, and 32, and establish, on the contrary, the constancy of the coals and the absence of sharp fluctuations in the numerical characteristics.

Such a sharp contradiction in the characteristics of coals by one method and the other is apparently explained by the fact that, in the plastometric analyses, deviations from the strict requirements established by the instructions were permitted. The method of L. M. Sapozhnikov permits no deviations from the requirements set out in the instructions with respect to uniformity of heating, coal moisture, degree of crushing, etc. Under conditions of mass laboratory control, these requirements cannot always be fully satisfied.

As Table 17 and the diagram of Figure 6 show, all the charges fell, in degree of softening, within the interval 70–50. The “equivalent” 50 evidently approaches the lower limit of caking under the coking regime conditions at the KMZ installation.

If one looks closely at the arrangement of the charges on the diagram,

Fig. 6.

Table 16

Table 17

one may note a curious feature of this arrangement. Namely, the charges are located mainly in a band at an angle of approximately 45° to the abscissa, so that, as the degree of softening increases, their thermal stability should increase. Such an arrangement of charges giving more or less constant results in the drum test of coke is quite natural, since the negative influence on the strength of the coke from a negative difference of “equivalents” increases with increasing degree of softening of the coals. This fact exhaustively explains to us the peculiarity in the behavior of the coals of the Osinniki deposit when coked by themselves, as coals of a high degree of softening with large negative differences of “equivalents.”

In contrast to the Osinniki coals, all coals of the Prokopyevsk deposit with a degree of softening of 50 “equivalents” and higher (coking coals) are characterized either by an insignificant negative or by a positive difference of equivalents, i.e., they are thermally more stable coals. Among the coals of the Prokopyevsk deposit with a degree of softening below 50 equivalents, both thermally stable and thermally unstable coals are encountered. In the present case, on the diagram there is not a single coal with a degree of softening below 50 “equivalents” and with a positive difference of equivalents (the Prokopyevsk and Anzhero-Sudzhensk mines); but, as tests of several samples taken by me at the KMZ coke plant in the summer of 1935 showed, such coals do occur among the Prokopyevsk coals.

From what has been set forth above it follows that successful coking of the Osinniki coals is possible on condition that such coals be added to them which would, first of all, lower for the blend the negative difference of equivalents, i.e., thermally stable coals.

Until 1935 the KMZ coke plant received grade K coals from Pro-

of the Kopeisk mine in such an amount as could ensure their participation in the charge at 25–30%; Osinovka coals went into the charge on average at 30%, and 40–50% consisted of PS-grade coals, so that the mixture was dominated by PZh and PS coals (70–75%). Since among PS-type coals thermally unstable coals are often encountered, such a charge did not provide sufficient thermal stability in the mixture, and the coke obtained was of low quality (drum 270–290 kg). The situation changed sharply when, in 1935, the plant was able to supply 45–50% of K-type coals and reduced the participation of PS to 20–25%.

A charge composed of coals of different degrees of softening and thermal stability, and coking coal, even if they have entirely identical indices and are coked under exactly identical conditions, may yield cokes that differ considerably in strength. Moreover, the differences in the properties of the coke will be the greater, the sharper the difference in the properties of the individual components entering into the charge. In this case the cause of the differences should be sought above all in unsatisfactory mixing of the charge, in the uneven distribution of the individual components in the coal charge. The coke here will constitute not a homogeneous mass, but pieces and fragments of coke from the individual components, bonded to one another.

To the same extent, what has been said also applies to the degree of crushing of the coals before mixing, and to the uniformity of crushing. A charge and coking coal with identical coking indices can yield cokes more or less similar in properties only under the condition of such pulverization and such mixing of the components in the charge as will sufficiently ensure uniformity of the composition of the coal charge into the oven.

Conclusions.

1. The degree of softening of coal must be recognized as the principal indicator of coking ability, since without the transition of coal into the plastic state the formation of coke is impossible.

2. Of all the methods so far proposed for the quantitative characterization of the degree of softening of caking coals, the method of Professor I. V. Gebler best meets the conditions of current factory laboratory control in its simplicity, reliability, and speed of carrying out the test.

3. The inconveniences of operating with the “softening numbers” obtained by the method of Prof. I. V. Gebler are eliminated by converting these “numbers” into “equivalent” quantities of substances known to be fusible, expressed as percentages of the content of these substances in a mixture with a completely infusible substance. The latter possess the property of additivity in mixtures and thus make it comparatively simple to perform calculations when composing mixtures.

4. A single quantitative characteristic of the degree of softening, however, is insufficient for characterizing coking ability, since coals are encountered with an identical quantitative characteristic of the degree of softening but with sharply different coking ability. In the present work it is shown that in this case the differences in coking ability are connected with the different character of decomposition of coals during the period of their softening. The different character of decomposition is due to the unequal thermal stability of the substances formed during the initial decomposition of the coal.

5. The difference in “equivalents” at experimental temperatures of 800° and 600° quite clearly characterizes the thermal stability of coals during the period of the plastic state. Since the “equivalents” possess the property of additivity at both of the named experimental temperatures, the new

“long” difference of equivalents (positive or negative) possesses the same property. Thus, no difficulty is created in calculating charges in this case either.

1. Verification of the proposed method for determining caking capacity on coals and charges of the coke plant of the Kuznetsk Metallurgical Works named after I. V. Stalin gives grounds for concluding that it is suitable for daily laboratory control.

LITERATURE

1) P. Damm. Glückauf No. 32, 1928. (“Properties of coking coals and the processes during their coking.”)

2) G. Agde and L. Linker. The Process of Formation of Lump Coke. Translation from the German, 1931 edition.

3) According to G. L. Stadnikov. Chemistry of Coking Coals, 1934 edition.

4) K. M. Sapozhnikov and L. P. Bazilevich. Chemistry of Solid Fuel, issue 2–3 for 1932. (“Plastometric study of the coking process.”)

5) I. V. Gebler. Coke and Chemistry, No. 10 for 1934. (“Method for determining the degree of softening of coals during heating and new numerical characteristics for caking coals.”)

6) S. A. Gusinskaya and E. I. Beskina. Coke and Chemistry, No. 1 for 1934. (“On the variability of coal bitumens during storage of coals in air.”)

7) P. Fischer, Broche, Strauch. Brennstoff-Chemie. 3, 33–43. 1925. (“On the composition of the components of coal bitumen and the role of its individual parts in the caking and swelling of coals.”)

8) Fuchs. Die Chemie der Kohle. Berlin. 1931.

9) Pott, Broche, u. an. Glückauf. 1933. S. 903.

10) According to Zammersbach and Schneider. Fundamentals of Coke Chemistry. Translation from the German; 1933 edition.

11) P. A. Mott and R. V. Wheeler. Coke for Blast Furnaces. Translation from the English, 1934 edition.

12) Experimental Coking of Coals of the Kuznetsk Basin. Report of the operational group of Soyuzkoks, in preparation. S. I. Panchenko, edited by Prof. M. I. Kuznetsov. 1932 edition.

13) G. L. Stadnikov: a) Chemistry of Solid Fuel, issues 11–12 for 1932 (“Theory of the Coking Process” and “The Question of the Melting of Coals during Coking at the Third International Conference on Bituminous Coals in Pittsburgh [USA]”); b) Chemistry of Solid Fuel, issue 6 for 1934 (“Initial propositions for compiling an industrial classification of fossil coals of the USSR”); c) Chemistry of Solid Fuel, issue 4 for 1934 (“New data on the question of the coking process of coals”).

14) I. A. Orlov. Essays on Coal Chemistry. (General chemistry of coal.) 1934 edition.

15) Works of L. M. Sapozhnikov and his collaborators in the journal Chemistry of Solid Fuel for 1935, issues 1, 2, 3, and 4, on the study of parameters of plastometric classification.

16) Strache-Lant. Chemistry of Coal, 1931 edition.

17) V. A. Vekhov, S. I. Zolotov, and G. I. Podbelsky. “Classification of Kuzbass coals by caking ability.”

18) N. P. Chizhevsky and G. N. Dmitriev. Chemistry of Solid Fuel, issue 6 for 1935. (“An attempt at classifying bituminous coals by caking ability.”)

19) M. B. Kamensky. Coke and Chemistry, No. 7 for 1935. (“Spherometric express method for determining the caking capacity of coals and their mixtures.”)

20) K. N. Shishkin. Coke and Chemistry, No. 4 for 1934. (“Bitumens and caking properties of coals.”)

21) Collection of works by staff members of the Kharkov Scientific Research Institute for Coal Chemistry and the Study of the Coking Process, published in 1932. (Collection of Works on Coal Chemistry, 1929–1931. Kharkov, 1932.)

22) P. Michaelis. Glückauf. 18, 413–423. 1935. (“Experience in coking fine lean coal.”)

23) K. Bunte, H. Bräckner, W. Ludewig. Glückauf. No. 34, 1933. (“Apparatus for determining the behavior of coal during softening and coke formation.”)

24) H. Winter. Glückauf, No. 34 for 1932. (“Analytical decomposition of coal.”)

25) B. Neumann, L. Kremser. Glückauf, No. 36 for 1933. (“Behavior of petrographic components of Upper Silesian coal during coking, semicoking, and extraction.”)