Investigation of the Structure of a Water-Gas Flame

Ya. N. Norkin and G. D. Spetsii

Introduction

In the rather extensive literature describing the methods and results of experimental and theoretical investigation of the chemical composition of a gas flame, we have found no materials concerning the flame of water gas; in most cases such studies have been carried out on flames of illuminating gas or of artificially prepared mixtures of combustible gases. This was one of the reasons for undertaking experiments to investigate the flame of water gas. In addition, it should be pointed out that the question of the greater or lesser combustibility of individual gases in a complex gas mixture cannot yet be regarded as finally resolved; more precisely, there are different points of view on the question of the preferential burning-out of the various gases present in the mixture. Therefore, the results of the investigation obtained by us may serve as additional factual material supporting a definite point of view. This circumstance also served as one of the reasons for undertaking experiments to investigate the flame of water gas. Finally, we believe that the results of our investigation may have significant practical importance in all cases where two-cone combustion of water gas occurs, i.e., under conditions in which atmospheric burners are used, when the amount of primary air in the combustible mixture constitutes only a certain fraction of the theoretically necessary amount of air for complete combustion—for example, when water gas is burned at factories producing electric lamps, where flame treatment of parts and assembly of electric lamps are carried out with the aid of atmospheric burners, and also in laboratories using water gas, etc. As is known, when a gas mixture consisting of combustible gas and primary air, present in the mixture in an amount insufficient for complete combustion, issues from burners, a stationary two-cone flame is formed, with two flame fronts.

On the surface of the first, inner cone, part of the gas burns in accordance with the amount of oxygen in the primary air; in doing so, all the oxygen of the primary air burns out. On the surface of the second, outer cone, the remaining part of the gas burns owing to diffusion of oxygen from the atmospheric air. What has been said above is illustrated by the diagram in Figure 1.

The height of the combustion cones depends chiefly on the velocity of flame propagation, the composition of the mixture, the outflow velocity, the burner diameter, and other factors. This question will be considered in detail in one of the following communications in connection with the results of an experimental investigation of combustion stability, depending on the relation between the outflow velocity and the flame-propagation velocity. In the present communication, the question of investigating the chemical composition of the flame gases passing through the surface of the inner combustion cone is considered.

Until 1890 there existed a generally accepted opinion that combustion develops in the space between the inner and outer cones, having begun on the surface of the inner cone. At that time the idea of the dual character of the flame was not generally accepted, according to which the gas burns only partially on the surface of the inner cone and completely on the surface of the outer cone.

In 1891–92 N. Teclu, A. Smithells, and H. Ingle, simultaneously and independently of one another, were the first to carry out experiments on separating the flame cone; they separated the inner cone from the outer one and investigated the composition of the interconal gases; these investigations served as the basis for modern ideas on the structure of the flame.

N. Teclu separated the flame of coal-gas illuminating gas and described the qualitative aspect of the separation process; in his report no information is given on the composition of the initial gas, on the composition of the interconal gases, or on the quantity of primary air in the initial combustible mixture. However, despite the absence of these quantitative characteristics, what is valuable in Teclu’s investigation is the very fact of the separation of the flame, as well as Teclu’s definite conclusion that “in the intermediate space between the two cones no combustion takes place.”

Fig. 1

Labels in Fig. 1:
Outer cone: combustion at the expense of oxygen diffusing from atmospheric air.
Inner cone: combustion at the expense of the oxygen of the primary air.
Combustible mixture.

A. Smithells and H. Ingle carried out experiments on the separation of the flames of illuminating coal gas and of various separately taken hydrocarbons (ethylene, methane, pentane, heptane, benzene). They investigated the composition of the interconal gases, which makes it possible to draw a number of conclusions concerning the quantitative side. Thus, for example, it is important to note that in the experiments of these investigators, when illuminating gas was burned, 4.8 volumes of primary air in the initial combustible mixture corresponded to 1 volume of gas, i.e. combustion was carried out with such an amount of primary air as is almost equal to that theoretically necessary for complete combustion (it is impossible to state this more definitely because of the absence of data on the chemical composition of the initial gas). Despite the large quantity of primary air, no oxygen was found in the interconal gases; all the oxygen had burned out in the first cone. The composition of the interconal gases was as follows:

\[ \mathrm{CO_2}—4.2\%;\quad \mathrm{CO}—8.8\%;\quad \mathrm{H_2}—9.3\%;\quad \mathrm{H_2O}—16.0\%\ \text{and}\ \mathrm{N_2}—62.0\%. \]

The calorific value of the interconal gases is approximately \(500\ \dfrac{\text{kcal}}{\text{m}^3}\) (calculated by us from the composition of the gases). The results of the investigation by Smithells and Ingle considerably influenced the solution of the question of what burns out first from the combustible mixture: hydrogen or hydrocarbons? From the materials of Smithells and Ingle’s investigation it follows that, both in the combustion of illuminating gas and even in the combustion of pure methane, hydrocarbons are absent from the composition of the interconal gases. This circumstance would seem to give grounds for asserting that hydrocarbons (for example, \(\mathrm{CH_4}\)) burn out preferentially in comparison with \(\mathrm{CO}\) and \(\mathrm{H_2}\). However,

as will be seen from what follows, the disappearance of hydrocarbons (for example methane) hardly signifies their complete combustion to CO₂ and H₂O, but most likely signifies only the initial conversion to CO and H₂ and only then, in the presence of oxygen, the subsequent burning out of CO and H₂ to the products of complete combustion. If in the experiments of Smith and Ingle there had not been such a large amount of primary air, but considerably less (for example, as much air as is required for the conversion of CH₄ to CO and H₂, or a little more), then there would have been sufficient material for deciding the question of what occurs first: complete combustion of hydrocarbons or only their conversion to CO and H₂?

Extensive investigations by Bunsen, Dixon, Horstmann, Landolt, Mistele, Bone, Drummond, Bone together with Davis and Greem are also devoted to resolving the question of the preferential burning out of individual gases, and of the greater or lesser strength of the chemical affinity of gases for oxygen in the process of combustion. We consider it necessary to dwell briefly on the elucidation of this question.

In 1853 Bunsen carried out a series of experiments on the combustion in a eudiometer of gaseous mixtures consisting of 100 volumes of O₂, 200 volumes of H₂, and an amount of CO varying in each experiment from 793.8 to 74 volumes. The results of Bunsen’s experiments are presented in Table 1.

The results of the experiments given in Table 1 provide grounds for drawing the following conclusions:

1. The distribution of oxygen during combustion between hydrogen with the formation of H₂O and carbon monoxide with the formation of CO₂ does not remain constant

\[ \frac{(O_2)_H}{(O_2)_{CO}} \ne \mathrm{const}, \]

but changes depending on the ratio in the gas

\[ \frac{CO}{H_2}. \]

1. When the concentration of one of the gases in the mixture decreases (for example CO), the fraction of O₂ going to the combustion of CO also decreases, but this decrease in the fraction of O₂ is not proportional to the decrease of CO in the initial mixture. If the decrease in the consumption of O₂ for the combustion of CO were proportional to the decrease of CO in the mixture, then in coordinates \(\alpha=\frac{CO}{H_2}\) and \(\beta=\frac{CO_2}{H_2O}\) one would obtain a straight line; in reality this is not the case, as is seen from Fig. 2 (constructed by us from the data of Table 1).

V. Bone, analyzing the results of Bunsen’s experiments, believes that Bunsen put forward an erroneous conception that, during the combustion of gases, the Berthollet principle—the principle of mass action—is not obeyed, and that combustion is accompanied by an abrupt (per saltem) violation of the principle of proportionality to the acting masses. In 1874 Meyer carried out experiments similar to Bunsen’s, and considered the results obtained by him to confirm Bunsen’s results. In 1890, 13 years after Guldberg and Waage had finally established the law of mass action, Dixon and Horstmann independently carried out experiments on the combustion of mixtures of CO, H₂, and O₂ and found that, during the combustion of these mixtures, the distribution of oxygen between the combustible CO and H₂ obeys the law of mass action.

V. Bone considers Bunsen’s experiments defective in that Bunsen worked with a moist eudiometer and at different temperatures of the initial mixtures; consequently the saturation of the gaseous mixtures with water vapor was different. Thus, experiments II, IV, VI, VII, IX were carried out at a mixture temperature from 2.3 to 7°C; experiments I, III, V, VIII—at a temperature

Table 1

Results of Bunsen’s experiments on the combustion of mixtures of CO, H₂, and O₂

22–23°C. Owing to the different content of water vapor in the initial mixtures, Bunsen obtained results on the distribution of oxygen which allegedly undermined the principle of proportionality to the active masses. Without fully agreeing with V. Bone’s assessment, it is nevertheless necessary to note that from any of Bunsen’s experiments there follows the conclusion that there is a preferential combustion of \(H_2\) as compared with CO. For example, the first experiment gives the ratio in the combustible mixture

\[ \frac{\mathrm{CO}}{\mathrm{H}_2}=3.97 \]

and in the combustion products the remaining \(H_2\) and CO are in the ratio

\[ \frac{\mathrm{CO}}{\mathrm{H}_2}=4.37; \]

in the last, ninth experiment the ratio in the combustible mixture is

\[ \frac{\mathrm{CO}}{\mathrm{H}_2}=0.37 \]

and in the combustion products

\[ \frac{\mathrm{CO}}{\mathrm{H}_2}=0.53, \]

etc.

Fig. 2

The question of the order of combustion of individual gases in a complex combustible mixture is also studied in the work of H. Landolt, carried out in 1856. Coal illuminating gas was taken as the fuel. The gas was burned in an open burner by diffusion of atmospheric air, i.e., without the participation of primary air; consequently, the combustion was single-cone. From the flame cone, along the axis of the burner, gas samples were taken for analysis at various heights from the burner mouth, namely at heights of 0, 10, 30, 20, 40, and 50 mm; the flame height was kept constant and equal to 100 mm. Analyzing the results of Landolt’s experiments, the following conclusions may be drawn:

1. As the gas-sampling point is moved away from the burner mouth, the amount of air drawn in from the atmosphere increases: thus, when sampling at a height of 0 mm from the burner mouth, the ratio

\[ \frac{\text{air}}{\text{gas}}=\frac{27.08}{100}, \]

whereas when sampling at a height of 50 mm the ratio is

\[ \frac{\text{air}}{\text{gas}}=\frac{381.66}{100}, \]

and it should be noted that this ratio corresponds approximately to 50% of the air theoretically required for complete combustion.

1. The combustion of hydrogen along the height of the flame is nonuniform. In the lower layers of the flame, where the inflow of air from the atmosphere is smallest, the combustion of hydrogen is fairly rapid; in the upper layers of the flame, where the inflow of air from the atmosphere increases as the point of gas sampling recedes from the mouth of the burner, the combustion of hydrogen slows down; this is illustrated by the data in Table 2.

Table 2

Combustion of hydrogen according to Landolt’s experiments

1. The combustion of methane along the height of the flame is also nonuniform. In the lower layers of the flame the combustion (disappearance) of methane is slow, but more or less uniform; in the upper layers, however, there is an abrupt and rapid disappearance; thus, the ratio

\[ \frac{\text{residual } CH_4 \text{ in the flame}} {\text{initial amount of } CH_4 \text{ in the mixture}} = \underbrace{0.83 - 0.74 - 0.74 - 0.56}_{\text{slow combustion of methane}} - \underbrace{0.31 - 0.09}_{\text{rapid combustion of methane}} . \]

1. The CO content in the combustion products in the flame at all points of gas sampling along the axis of the flame exceeds the amount of CO in the initial combustible mixtures. An especially sharp increase of CO in the flame occurs at the 6th sampling point (50 mm from the mouth of the burner), where the combustion products contain the smallest amount of methane in the flame. At this point the CO content increases from 5.56 volumes in the initial mixture to 25.14 volumes in the flame gases.

On the basis of his experiments, Landolt concluded that “gases disappear in the flame according to the degree of their combustibility. Of all the other gases, hydrogen is removed fastest of all; methane disappears somewhat more slowly, and, finally, the heavy hydrocarbons, whose combustion occurs predominantly in the upper half of the flame.” With regard to this conclusion, W. Bone remarks that the true course of the combustion reaction in Landolt’s experiments was masked by the catalytic action of the platinum tube through which the gas was withdrawn for analysis and which, during sampling, was heated red-hot. Meanwhile it is known that in the presence of a platinum catalyst (tube) hydrogen burns selectively.

To this conclusion of W. Bone it should be added that the sharp increase of CO with the decrease of \(CH_4\) in the flame indicates that, in the flame, when there is a deficiency of air, what occurs is not the combustion of \(CH_4\), but the conversion of methane into CO and \(H_2\), and that this conversion, apparently, takes place in the per-

in the first instance, and only then the combustion of CO and H₂. This idea is also suggested by the experiments of A. Smithells and H. Ingle, as was mentioned above (p. 5).

In 1861 Kersten published the results of his experiments on the explosion of a mixture of ethylene with electrolytic gas of the following composition: C₂H₄—5.48 cm³, H₂—15.40 cm³, O₂—7.70 cm³, N₂—0.22 cm³. As a result of the explosion the following products were obtained: C₂H₄—0.38 cm³, CO—10.50 cm³, H₂—20.70 cm³, and N₂—0.22 cm³. From this it is seen that approximately 93% of the ethylene was converted into CO and H₂

\[ \left(\frac{5.48 - 0.38}{5.48} \cdot 100 = 93\%\right); \]

as for H₂, about 68% of it remained unburned

\[ \left(\frac{20.70 - 5.48 \cdot 2 - 0.38 \cdot 2}{15.40} \cdot 100 = 68\%\right). \]

Kersten concluded that, before the hydrogen burns out, all the carbon of the ethylene is burned to carbon monoxide and then the excess oxygen is distributed between CO and H₂. In 1874 Meyer obtained results which, in the main, confirmed Kersten’s conclusions. Mistele expressed in 1905 the idea that in a flame oxygen reacts not with hydrocarbons as such, but with the products of their decomposition (C and H), and that, in the first instance, the carbon burns out.

In recent times, detailed investigations were carried out by W. Bone and his collaborators to resolve the question of the relative combustibility of hydrocarbons, hydrogen, and carbon monoxide in a gaseous flame. In 1906 Bone and Drugman (Bone and Townend, Flame and Combustion in Gases, 1927) exploded mixtures of the composition—C₂H₆ + O₂; C₂H₄ + H₂ + O₂; C₂H₂ + 2H₂ + O₂. These mixtures contained equal quantities of carbon, hydrogen, and oxygen, i.e. 2C, 6H, 2O. As a result of the explosion it was found that, for the ethane–oxygen mixture, the explosion products contained free carbon, carbon monoxide, methane, hydrogen, and water vapor in rather considerable quantities. For the ethylene–oxygen mixture, the explosion products contained no free carbon at all, and only 10% of the oxygen was consumed in the formation of water vapor. For the acetylene–oxygen mixture, neither free carbon nor water vapor was found in the explosion products; all the oxygen was consumed in the formation of CO. The investigators concluded that hydrocarbons receive in the flame the greater part, or even all, of the oxygen, while hydrogen remains free; that is, in the flame the affinity of hydrocarbons for oxygen considerably exceeds the affinity of hydrogen for oxygen.

Then, in 1915, Bone, Davis, and Gray exploded a mixture of methane and electrolytic gas of the composition CH₂—O₂—2H₂. The explosions were carried out at various initial pressures of the mixtures (18.33—19.62—48.85—17.12—18.31—26.12 atm). As a result it was found: 1) deposits of free carbon were not observed, 2) the amount of unreacted methane was small (2.80—5.25), 3) the distribution of oxygen between methane and hydrogen was as follows: from 95.7 to 97.2% O₂ went to methane, and from 2.8 to 5.25% O₂ went to hydrogen.

On the basis of these research results, Bone and his collaborators conclude that in an explosive flame the affinity of methane for oxygen is 20–30 times stronger than the affinity of hydrogen for oxygen. With regard to such a conclusion, Wheeler objected, saying that oxygen must necessarily combine with methane in order to form such a methane–oxygen mixture that would have the same flame-propagation velocity as the flame-propagation velocity of the hydrogen–oxygen mixture (it is known that the content of fuel in a mixture determines the flame-propagation velocity).

In subsequent experiments Bone, Davies, and Gray exploded methane-hydrogen-oxygen mixtures of variable composition, $\mathrm{CH_4 - O_2 - xH_2}$, where $x = 2, 4, 6, 8$. It was found that the percentage of unburned methane is proportional to $x^2$.

Thus, for the 1st mixture, $x = 2$: $2.9\%$ $\mathrm{CH_4}$ did not burn;

“ 2nd mixture, $x = 4$: $9.0\%$ $\mathrm{CH_4}$ “;

“ 3rd mixture, $x = 6$: $27.4\%$ $\mathrm{CH_4}$ “.

The authors named also carried out an investigation involving the explosion of mixtures of the composition $\mathrm{CH_4 + O_2 + 2CO}$ in order to establish the order of combustibility of methane and carbon monoxide. It was found that no free carbon was formed, and the distribution of oxygen between $\mathrm{CH_4}$ and $\mathrm{CO}$ was as follows:

from $4.3$ to $5.1\%$ $\mathrm{O_2}$ goes to $\mathrm{CO}$, and

from $94.9$ to $95.7\%$ $\mathrm{O_2}$ goes to $\mathrm{CH_4}$.

In conclusion, W. Bone writes that, when the quantity of oxygen is limited, competition arises between hydrocarbons and hydrogen, as a result of which the hydrocarbons receive almost all the oxygen, i.e., in the flame hydrocarbons are more combustible than hydrogen.

In undertaking the investigation of the structure of a water-gas flame, we clearly realized that, by using for the investigation water gas produced under factory conditions, we could not obtain strictly constant research results because of fluctuations in the composition of the water gas; of course, it would have been more correct to prepare artificially a gas mixture of constant composition, in which case the results obtained would have had greater scientific accuracy. However, at the cost of some accuracy, we found it possible to work with water gas as produced at the plant, so that the results obtained would more closely approximate actual conditions.

Water gas is produced in the gas shop of the Tomsk electric-lamp plant in “Pinch” gas generators with manual servicing. The fuel used is hard coal from the Aralichevsk deposit. The coal is lean, partly semi-anthracitic, and does not cake.

The average composition of the coal is: $W^p — 9.0\%$; $A^c — 16.0\%$; $S^0_{\text{total}} — 0.8\%$; $C^2 — 88.1\%$; $H^2 — 4.0\%$; $N^2 — 1.7\%$; $O^2 — 5.3\%$; $V^2 — 12.0\%$.

The technological regime of gasification is not strictly constant and usually varies within the following limits: the period of air (hot) blast is from 2 to 3 min, the period of steam (cold) blast is from 3 to 4 min, sometimes being lengthened to 5–6 min. The principal indication used for stopping the steam blast is the pressure of the water gas before discharge into the scrubber; at the beginning of the steam blast the pressure in the scrubber is usually 190 mm water column, and the steam blast is stopped when the pressure falls to 80 mm water column. Until November 1947 saturated steam was used; beginning in November 1947, superheated steam at a pressure of 4 atm and a temperature of $325^\circ\mathrm{C}$ was used. The blast is only from below. As a result of using superheated water vapor, the calorific value of the water gas changed considerably, from 2500 kcal/m$^3$ on average to 2760 kcal/m$^3$ on average, owing to an increase in the $\mathrm{CH_4}$ content.

From this point of view, the water gas produced could be called double water gas, although no structural changes of a constructive nature were introduced into the design of the generator.

The fluctuations in the composition of the water gas according to analyses made in January, February, and March 1948 are given in Table 3.

Table 3

Fluctuations in the composition of water gas

Method of Investigation

The investigation of the structure of a water-gas flame was carried out by separating a two-cone flame; in this, the chemical analysis determined the composition of the initial gas, the composition of the gas–air mixture supplied to the burner, and the composition of the intercone gases. The following five series of experiments were performed:

1st series (A)—the flame of a Bunsen burner; the amount of air in the initial mixture was varied;

2nd series (B)—the flame of a Meker burner with a platinum mesh on the burner; the amount of air in the initial mixture was likewise varied;

3rd series (V)—the flame of a two-tube injection burner with forced supply of gas and air to the burner;

4th series (G)—the flame of a two-tube injection burner with a platinum mesh;

5th series (D)—experiments on the combustion of water gas in capillary tubes.

The arrangement of the apparatus for the 1st and 2nd series of experiments is shown in Fig. 3.

In the first series of experiments the outlet opening of the burner (1), 9 mm in clear diameter, was open; in the second series of experiments a cap (2) with copper tubes 13 mm in diameter was inserted into the outlet opening of the burner, covered on top with a platinum mesh having 625 openings/cm². On the burner, by means of a stopper, an outer glass tube (3), 35 mm in diameter and 400 mm high, was mounted. At a distance of 2/3 of its height this glass tube had a glass branch (4), 7 mm in diameter. The intercone gases were directed through this glass branch and a rubber tube (5) into an aspirator (6). Control experiments were carried out each time at the same pressure, which was measured by an alcohol U-shaped manometer (7) on the gas pipeline; the amount of air entrained was regulated by opening the air holes, for which purpose a pointer (9) was attached to the sleeve with the air holes (8), by means of which the angle of rotation of the sleeve was set. Ignition of the combustible mixture was at first carried out with the aid of a nichrome spiral (10), heated by an electric current to red heat, but subsequently the spiral was removed in order to avoid the catalytic influence of nichrome on the combustion process, and ignition of the combustible mixture was carried out from above through a wide glass tube. In this case only gas was supplied through the burner and tube, access of primary air being closed. Then, with the aid of the sleeve (8), the air holes were opened slightly. Then, at the base of the upper cone above the tube (3), a second inner cone of bluish color appeared, very unstable. The height of this cone was very small in comparison with the diameter of the base. As the opening …

from the opening for the primary air, the initial orange-yellow coloration of the outer cone becomes less intense, and the inner cone descends along the glass tube to the mouth of the burner and is stabilized there. Stable separate two-cone combustion is formed. The upper cone is predominantly colored violet-blue and has a bluish border along the edges; the lower cone is blue with a slightly greenish tint.

In carrying out the second series of experiments, a Meker cap with an inserted platinum gauze was placed on the burner. Ignition was carried out in the same way as described above. It should be noted that, in the presence of the platinum gauze, combustion in the inner cone was stable even

Diagram of the apparatus for investigating separate two-cone combustion in an injection burner.

Fig. 3

when a considerable amount of primary air was added; the inner cone merged with the platinum gauze, so that it was not visible to the eye; the platinum gauze was heated to a bright-white glow. Since the primary air was ejected in an amount smaller than that required for complete combustion, an outer combustion cone still formed above the glass tube.

Gas analysis was carried out with the aid of a gas analyzer with afterburning. The calculations were made by the usual method.

The third and fourth series of experiments were carried out on the apparatus shown in Fig. 4.

Water gas from the main line (1), under a definite pressure, is supplied to the gas line (2), 15 mm in diameter; in the gas line there is installed a valve (3) for regulating the flow rate and a sharp-edged orifice, 7.5 mm in diameter (4); the pressure drop across the orifice is recorded by a differential manometer (5). Through a nozzle 1.5 mm in diameter (6), the gas enters the burner mixer (7). Air, delivered by a blower through an air duct (8) 25 mm in diameter, is also supplied there. A valve is installed in the air duct

(9) for regulating the flow rate, and a sharp-edged orifice (10) with a differential manometer (11). To obtain separate two-cone combustion, as in the preceding experiments, a glass tube (12) of diameter 40 mm and height 380 mm is used. To prevent melting and deformation of the glass tube as a result of the high temperature, the inner surface of the tube is carefully insulated with asbestos (13), with two openings

Fig. 4. Schematic of the apparatus for investigating separate two-cone combustion in a two-pipe burner.

in the asbestos: one opening opposite the glass branch (14), and the other opening (15) for visual observation of the behavior of the inner cone. The intercone gases, as well as the initial combustible mixture, were drawn off by an aspirator (16) through the glass gas-discharge tube (14). Ignition of the combustible mixture was carried out in the same way as in the preceding two series of experiments.

Results of the Investigation

A. Experiments with an Injection Atmospheric Burner

Experiments on the combustion of water gas in an injection atmospheric burner were carried out on the apparatus shown in Fig. 3. Water gas from the main was supplied to the burner under a pressure in the gas line of 135 mm water column. The gas pressure was regulated, according to an alcohol manometer, by a valve on the gas line; the entrainment of primary air was set by opening the air holes. Samples of water gas, combustible mixture, and intercone gases were taken into the aspirators. The experiments were carried out with the addition of various amounts of primary air to the water gas.

The results of the experiments are presented in Table 4.

Results of experiments of series A. Atmospheric injection burner. Gas pressure 135 mm water column.
Schematic of the apparatus in Fig. 3

From the data given in Table 4 it follows that, when water gas is burned by means of an atmospheric injection burner with a deficiency of primary air, hydrogen burns out preferentially in the inner cone of the flame; the relative decrease of hydrogen in the first experiment is

\[ \frac{\mathrm{H_2}}{\mathrm{CO}+\mathrm{CH_4}}=\frac{1.28}{0.88}=1.45 \]

and in the second experiment

\[ \frac{1.49}{1.05}=1.42. \]

With an increase in the fraction of primary air from 0.345 to 0.475, the relative “combustibility” of hydrogen decreases, but the relative “combustibility” of methane increases, although preferential burning remains with hydrogen. This is also confirmed by the data given in column 8—the ratio

\[ \frac{\mathrm{CO_2}}{\mathrm{CO}+\mathrm{CH_4}}, \]

which increases from

\[ \frac{0.33}{0.11}=3 \]

to

\[ \frac{0.41}{0.11}=3.7. \]

B. Experiments with a platinum gauze in an atmospheric injection burner

Experiments on the combustion of water gas on a platinum gauze in an atmospheric injection burner were carried out on the apparatus shown schematically in Fig. 3. Water gas was supplied to the burner from the mains under a pressure of 150 mm water column. The gas pressure was regulated by a manometer with the aid of a valve; the entrainment of primary air was set by opening the air holes. Combustion on the platinum gauze made it possible to increase the entrainment—

... primary-air suction without the formation of reverse motion of the flame (without flashback). The results of the experiments are presented in Table 5.

Table 5

Results of experiments of series B. Atmospheric injection burner with a platinum gauze.
Gas pressure 150 mm water column (scheme in Fig. 3)

The results given in Table 5 show that, when water gas is burned in the presence of platinum, the relative decrease of hydrogen during combustion is: in the first experiment

\[ \left(\frac{H_2}{CO+CH_4}\right)^{\text{gas}} : \left(\frac{H_2}{CO+CH_4}\right)^{\text{i.c.g.}} = \frac{1.4}{0.73} = 1.92 \]

and in the second experiment

\[ \frac{1.47}{0.65}=2.27. \]

If these results are compared with the data of Table 4 (series A, without catalytic influence), then it follows from the comparison that, under the influence of the catalytic action of platinum on the combustion process, the relative “combustibility” of hydrogen increases in comparison with carbon monoxide and methane. Moreover, despite the increase in the fraction of primary air, the relative combustibility of hydrogen not only does not decrease, as occurs in series A experiments, but, on the contrary, increases. If the results of gas combustion in the presence of the platinum gauze are compared with the results of combustion without the gauze but with an increased quantity of primary air (Table 6), one may come to the conclusion that, despite the increase in the fraction of primary air in the combustible mixture during combustion on the gauze, the preferential burnout remains with hydrogen.

B. Experiments with a Two-Wire Burner

Experiments on the combustion of water gas by means of a two-wire burner were carried out in order to make it possible to know not only the ratio between air and gas in the combustible mixture (from the chemical...

analysis), but also the quantities of air and gas by their direct measurements. These experiments were carried out on an apparatus shown schematically in Fig. 4. The flow rate of air and gas was determined from the pressure drop created by diaphragms installed in the gas and air pipelines. The gas pipeline has an internal diameter of 15 mm, the diaphragm 7.5 mm. The calculation formula for determining the gas flow rate is

\[ V_g = 0.54\sqrt{h}\ \text{m}^3/\text{h} \]

The air pipeline has an internal diameter of 25 mm, the diaphragm 12.5 mm. The calculation formula for determining the air flow rate is

\[ V_a = 1.11\sqrt{h}\ \text{m}^3/\text{h}. \]

The outlet opening of the burner has a diameter of 9 mm. The velocity of outflow of the combustible mixture from the burner was equal to 10.6 m/sec. The results of the experiments are presented in Table 6.

Table 6

Results of experiments of series B. Injection two-pipe burner
Diagram in Fig. 4

Gas pressure 100 mm water column.
Air pressure 25 mm water column.
Pressure drop in the diaphragm 3.6 mm water column.
Gas flow rate 1.027 m³/h.
Pressure drop of air in the diaphragm 1.6 mm water column.
Air flow rate 1.4 m³/h.
Theoretically required quantity of air 2.37 m³/m³.

From the experiments of series B on the combustion of water gas with the aid of a two-pipe burner, which makes it possible to measure the gas and air flow rates, it follows: 1) that the relative “combustibility” of hydrogen is less than the relative “combustibility” of the sum of carbon monoxide and methane, and 2) that the relati-

the relative “combustibility” of methane is greater than the relative “combustibility” of the sum of carbon monoxide and hydrogen. Comparing these conclusions with the results of the investigation given in Table 4, obtained under analogous conditions, one might arrive at the conclusion that they are contradictory; however, attention should be paid to the fact that in the present case combustion was carried out with a large proportion of primary air: the proportion of primary air was increased to 0.58, as compared with that in the preceding experiments, where it amounted to 0.345—0.475. The experiments of series A (Table 4) and series B (Table 6) show that with an increase in the proportion of primary air in the combustible mixture the relative “combustibility” of hydrogen decreases and, moreover, that when the proportion of primary air is increased to 0.5 and above the theoretically required amount, preferential burnout after hydrogen is not retained.

G. Experiments with a platinum gauze in a two-tube burner

These experiments were performed on the apparatus shown schematically in Fig. 4. The outlet opening of the burner was covered with a Meker cap with a platinum gauze having 625 holes/cm². The experiments with the platinum gauze made it possible to increase the gas and air flow rates and to reveal the catalytic influence of the flame on the combustion process of the individual gases. The outlet opening of the burner cap has a diameter of 12 mm. The free cross-section of the platinum gauze, with a wire thickness of 0.09 mm, was 0.000053 m². The velocity of the combustible mixture flowing through the gauze is expressed by the values: 15.5 m/sec for the first experiment and 25.8 m/sec for the second experiment.

Table 7

Results of experiments of series G. Injection two-tube burner with platinum gauze

Schematic diagram in Fig. 4

Gas pressure 160 mm water column.
Air pressure 60 mm water column.
Pressure drop of gas in the orifice: 1st experiment 4.5 mm water column; 2nd experiment 7.6 mm water column.
Gas consumption: 1st experiment 1.15 m³/hour; 2nd experiment 1.49 m³/hour.
Pressure drop of air in the orifice: 1st experiment 3.7 mm water column; 2nd experiment 9.8 mm water column.
Air consumption: 1st experiment 1.8 m³/hour; 2nd experiment 3.43 m³/hour.
Theoretically required quantity of air: 1st experiment 2.55 m³/m³; 2nd experiment 2.4 m³/m³.

From the experiments of series G on the combustion of water gas on the platinum gauze of a two-wire burner, it follows that hydrogen burns out preferentially, since the relative amount of hydrogen decreases; in the first experiment the ratio

\[ \left(\frac{\mathrm{H}_2}{\mathrm{CO}+\mathrm{CH}_4}\right)^{\text{gas}} : \left(\frac{\mathrm{H}_2}{\mathrm{CO}+\mathrm{CH}_4}\right)^{\text{l.c.g.}} = \frac{1.46}{0.59} = 2.47 \]

and in the second experiment this ratio is equal to

\[ \frac{1.8}{0.55}=3.3. \]

If these experiments are compared with the experiments of series B, also carried out on a platinum gauze but with a smaller amount of primary air (Table 5), then the comparison again shows that, under the catalytic action of platinum, simultaneously with an increase in the fraction of primary air the “combustibility” of hydrogen increases. The combustion of methane lags far behind.

The next experiment was also set up with the aid of a two-wire burner covered with a platinum gauze, but with a considerably increased flow rate of air and gas. The gas flow rate was 2.5 m$^3$/h and the air flow rate 9.5 m$^3$/h.

For water gas of the composition indicated in Table 8, the theoretically required amount of air is 2.5 m$^3$/m$^3$; consequently, combustion proceeded with an excess of air equal to

\[ \frac{9.5}{2.5\cdot 2.5}=1.52. \]

It would therefore be expected that there should be no unburned gases in the combustion products, and still less unburned hydrogen; however, in reality, CO, CH$_4$, and hydrogen all proved to be unburned in the combustion products, as is evident from Table 8.

The results of the investigation of the combustion of water gas with an excess of primary air in a two-wire burner show: 1) that, despite the considerable excess of air, complete burnout of the gas components is not achieved, owing to imperfect mixing of the gas and air; 2) that with an excess of air the burnout of hydrogen occurs faster than the burnout of methane.

It should be noted that in this experiment the flame front had no definite shape; the unburned gases, together with the excess oxygen, “broke through” the flame front owing to the high velocity of efflux. Apparently, owing to the high velocity of passage of the gas through the flame front and the imperfect mixing, despite the large excess of air, combustion proved incomplete. The velocity of efflux of the combustible mixture through the gauze was 42.5 m/sec.

Table 8

Results of burning water gas with an excess of primary air in the burner according to the scheme in Fig. 4

Gas pressure: 200 mm water column.
Air pressure: 80 mm water column.
Gas pressure drop in the orifice: 10 mm water column.
Gas flow rate: 1.7 m³/hour.
Air pressure drop in the orifice: 34 mm water column.
Air flow rate: 6.45 m³/hour.
Theoretically required amount of air: 2.5 m³/m³.
Air-excess coefficient: 1.52.

D. Experiments on the combustion of water gas in capillary tubes

All the preceding experiments were carried out under conditions of separate two-cone combustion during the outflow of the combustible mixture, with the formation of a stationary flame front; in this case the composition of the intercone gases was analyzed, but not their quantity. In order to obtain a clear picture of the distribution of the oxygen of the air for the combustion of CO, H₂, and CH₄ separately, it was necessary to carry out experiments that would make it possible to measure the quantity of the products of incomplete combustion formed and the ratio between them. Such an investigation was performed by burning a mixture of water gas and air in a definite ratio:

when burning in a porcelain tube, the ratio

\[ \frac{\text{gas}}{\text{air}} = 0.852 : 1, \]

when burning in a platinum tube, the ratio

\[ \frac{\text{gas}}{\text{air}} = \frac{1}{1}, \]

in both cases the initial quantity of combustible mixture was 100 cm³. The combustible mixture was repeatedly passed through tubes heated from the outside by the flame of a gas burner until the volume of the combustion products obtained became constant; this indicated the cessation of the combustion process owing to the complete disappearance of the oxygen of the air. Then the composition of the products of incomplete combustion obtained was analyzed by the usual method. The arrangement of the apparatus is shown in Fig. 5. The results of the investigation are presented in Table 9.

Table 9

Results of experiments of series D. Combustion of water gas in capillary tubes

Arrangement in Fig. 5

The results of the experiments of series D on the combustion of water gas in a capillary porcelain tube, i.e., with a weak catalytic action of the walls on the combustion process, make it possible to conclude that, just as

and in the experiments of series B, where combustion was carried out with primary air, the fraction of which is greater than 0.5 (Table 6), the predominant burn-out remains with the sum of carbon monoxide and methane, while the “combustibility” of hydrogen is less than the “combustibility” of the other gases; here it should be assumed that methane burns out predominantly, if one compares the distribution of oxygen among hydrogen, carbon monoxide, and methane during combustion.

Fig. 5

During the combustion of water gas in a platinum capillary tube, the relative decrease of hydrogen during combustion is equal to the ratio

\[ \left(\frac{\mathrm{H}_2}{\mathrm{CO}+\mathrm{CH}_4}\right)^{\mathrm{gas}} : \left(\frac{\mathrm{H}_2}{\mathrm{CO}+\mathrm{CH}_4}\right)^{\mathrm{m.k.g.}} = \frac{1.66}{0.98} = 1.7 \]

when the amount of air is equal to 0.39 of the theoretically required amount.

Comparison of these results with the data obtained in a porcelain capillary reveals a greater “combustibility” of hydrogen in comparison with carbon monoxide and methane as a result of the catalytic influence of platinum. In addition, if these data are compared with the results of the investigation of combustion also in the presence of platinum, but with a larger amount of air (Tables 5, 7), one may conclude that, when the amount of primary air is decreased, the “combustibility” of hydrogen decreases.

Conclusions

1. The article presents a description of experimental work on the investigation of the structure of a water-gas flame by the method of two-cone separate combustion, when two stationary conical combustion surfaces are formed, and also in capillary tubes.

2. It has been shown that, in the absence of an active catalyst, the individual gases constituting water gas burn out in the following order—

Thus: in the presence in the combustible mixture of primary air in an amount not exceeding \(\frac{1}{2}\) of that theoretically necessary for complete combustion, hydrogen burns out preferentially in comparison with carbon monoxide and methane; if the amount of primary air is greater than \(\frac{1}{2}\) of the theoretically necessary amount, then the relative “combustibility” of hydrogen decreases, and preferential burnout after hydrogen is not retained, yielding its place to methane.

1. It has been shown that in the presence of a platinum catalyst, at any amounts of primary air, hydrogen burns out preferentially before carbon monoxide and methane; moreover, with an increase in the fraction of primary air, the relative “combustibility” of hydrogen increases.