April 28, 1916. The colliery guardian. 801 Results when Paper Diaphragm was Pulled Off. In order to determine whether the paper diaphragm at the top of the box had any appreciable effect in fixing the limits because of the slight pressure it would exert on the mixtures before it was disrupted by the force of the explosion, some experiments were made in which the diaphragm was pulled off the top of the box at the moment the mixtures were ignited. In this manner free expansion of the heated products of com- bustion was allowed, and compression of the burnt gases thereby avoided. The results showed that under the conditions of the tests the low limit was between 4-9 and 5 per cent, methane, and that the pressure in the first experiments caused by the paper diaphragm was not enough to influence the results. Results with Ignition from the Top. In the next experiments, ignition was effected from the top of the box, and the paper-covered diaphragm was placed at the bottom in some experiments, and at the top in others. The results showed the lower limit of inflammability of methane-air mixtures to’ be between 5-4 and 5-5 per cent, of methane with the aperture at the bottom, and between 5*7 and 5*8 per cent, with the aperture at the top. Results with Box Placed Horizontally. Tests were also made to determine the lower limit with the box in a horizontal position. The results showed that, under these conditions, the lower limit for horizontal propagation of flame was between 5-5 and 5-6 per cent, of methane. IO1_. Fig 2.—Hempel Explosion Pipette. Experiments with Hempel Explosion Pipette. The inflammability test (top ignition), in which a 100 c.c. Hempel glass explosion pipette (fig. 2) was used, mercury being the confining liquid, and two platinum wires being placed at the top of the vessel, whilst.the spark for ignition of the gaseous mixtures was obtained from an induction coil driven by four dry cells, showed that the lower limit of inflammation for the conditions used was between 5'5 and 5-6 per cent, of methane, corresponding to results obtained under the same con- ditions at different times at the Pittsburg station. In a test in the wood and glass box, ignition being effected at the bottom, and the proportion of methane throughout the box being 4*9 per cent., the flame widened as it travelled to the top, where it occupied almost the entire cross section. The result of the experi- ment duplicated closely the result previously obtained with a similar percentage of methane and air. Fig. 3.—Glass Explosion Vessel. Experiments with 2,800 c.c. Glass Vessel. Further experiments were made with a glass vessel of 2,800 c.c. capacity (fig. 3), 5 in. (12-7 cm.) wide, and 12 in. (30 cm.) high. The neck was sharply contracted to a diameter of 2|in. (6-3 cm.). A close-fitting paper diaphragm covered the top part. Two copper wires carrying a current of 7 amperes at 220 volts met at the bottom^of the vessel. Ignition was brought about by pulling the wires apart. In the first series of tests ignition was from the bottom upward, with the result that, with 5 and 5*1 per cent, of methane, a narrow flame 1 to 2 in. wide travelled to the top of the jar; with 5'2 per cent., the flame apparently filled the jar in its travel upward. General Comments on the Tests. In connection with the tests described above, no analyses were made of the products of combustion, the observations being confined to the results as seen by the eye. Unquestionably much methane remained Fig 4—Iron Pipe for Testing. from a Hempel that the methane- was about unburned in some of the mixtures through which the flame travelled, but the investigation was confined to the propagation of flame through the mixtures. The authors know of no reason why, in those mixtures, in the large box, that did propagate flame to the top, the propagation should not have been indefinite. Determining Upper Limit of Inflammability. Experiments to determine the upper limit of propaga- tion of flame in mixtures of methane and air have not been as numerous as those to determine the lower limit, principally because the lower limit is the dangerous one, and hence the one whose determination is most important. A methane-air mixture of 9-47 per cent, of methane and 90-53 per cent, of air contains enough oxygen for the complete combustion of the methane to carbon dioxide and water vapour. If more methane is present, carbon monoxide and hydrogen result. If the methane in excess of 9-47 per cent, remained inert like nitrogen, and had the same specific heat, it would, like nitrogen, act as a diluent; the speed of the reaction would be reduced by the presence of less and less oxygen until, in a mixture containing about 23 per cent, of methane and 77 per cent, of air, no explosion could be obtained. In such a mixture there would be 16 per cent, of oxygen. However, it is actually found that the upper limit is not so high as 23 per cent., principally because the excess methane forms carbon monoxide and hydrogen. Although the reaction produces heat, it also takes oxygen that would otherwise form carbon dioxide and water vapour. Much more heat is formed in the reaction that produces the latter substances, so that the net^ result is that much less heat is produced than would be if the excess methane remained inert. If the methane were mixed with oxygen, combustion should be complete in mix- tures containing up to 33-3 per cent, methane. With higher percentages, products of incom- plete combustion form, and explosions occur, until the volume of products formed is sufficient to show the reaction, so that not enough heat is developed to raise the gases to their ignition temperature. The authors found from their experiments in explosion pipette upper limit for oxygen mixtures 45 per cent. Experiments with Iron Pipe. Experiments on the upper limit of flame propagation were made in an iron pipe (fig. 4), 12 in. (30 cm.) in diameter, and 7 ft. (2-13 m.) high, with a capacity of 172 litres. The flame could be noted through narrow glass windows, one above the other. An asbestos cord lubricated with graphite, and passing through a mer- cury trap, was fastened on the inside to a weighted paper disc that could be moved up and down to mix the gases. Pairs of copper wires entered both the top and the bottom of the pipe. Each pair of wires could be pulled apart while an electric current was flowing, thereby producing a flash, and affording a means of igniting the mixtures either from the top downwards or from the bottom upward. An electric current of 7 amperes and 220 volts was used. A paper diaphragm covered an aperture 84 in. (22 cm.) in diameter at the top of the pipe, and its disruption allowed a vent for the burned gases after an explosion. The results obtained when ignition was effected from the top downward showed the upper limits of propaga- tion of flame to be between 13-5 and 13-9 per cent, of methane. The results obtained when ignition was effected from the bottom upward showed the upper limit of flame propagation to be between 15 and 15-4 per cent. Experiments in Hempel Pipette. Experiments were next made to determine the upper limit of complete propagation in the Hempel explosion pipette, shown in fig. 1. Ignition was brought about from the top downward by platinum wires, across which a spark was made to jump by means of an induction coil driven by four dry cells. These experiments showed the upper limit of flame propagation to be between 12-4 and 13-2 per cent, of methane in the Hempel explosion pipette. The result is somewhat lower than that obtained by ignition from the top in the iron pipe. Significance of Results. The value of the results obtained by the foregoing experiments lies in their application to the occurrence of methane-air mixtures in coal mines. The experi- ments showed that flame will travel upward in mix- tures too lean to cause it to travel downward, and will travel horizontally in mixtures containing a percentage of methane intermediate between that which will allow it to travel downward and. that which will allow it to travel upward. In mines it is seldom, if ever, that a uniform mixture of methane and air forms throughout a large area, or that the mixture in a particular place remains constant for a very long time. Hence the slightly different results obtained under different experi- mental methods are chiefly of scientific interest. So far as the coal miner is concerned, results obtained for hori- zontal propagation of flame are of more value than those obtained by ignition from the bottom upward. In a mixture of about 5 per cent, of methane, the flame would, if the mixture were ignited from below, travel to the top and perhaps then ignite a richer mixture, through which the flame would travel horizontally as long as the methane content was more than 5-5 per cent, methane, and as long as enough oxygen was pre- sent to burn the methane. It is generally recognised, of course, that the widespread devastation produced by explosions in mines, is usually not due to' explosive mixtures of methane and air extending throughout the workings, but to the ignition of dust-air mixtures by the explosion of a pocket of methane and air. The fact that methane may be brushed from places where it has accu- mulated, and, travelling down an airway usually fr’ee from gas, be ignited by an open flame or otherwise, has many times been demonstrated. Effect of Lowering Oxygen and Raising Carbon Dioxide Content. In coal mines, especially when the air is not being moved by a fan, the composition of the atmosphere changes rapidly at times. The principal changes that effect the explosibility of mine atmospheres are those having to do with the absorption of oxygen by the coal and the oxidation of coal to carbon dioxide. Both changes, if carried far enough, can result in reducing the oxygen content or in adding carbon dioxide until an explosion cannot take place, even if an. explosive pro- portion of methane be present. Therefore, knowledge of the propagation of flame in limit mixtures of methane, carbon dioxide, oxygen, and nitrogen is important. Haldane and Atkinson were the first to publish results of experiments with this phase, but the most exhaustive experiments on the subject were performed by J. K. Clement,* of the. Bureau of Mines, who found that the limits of inflammability were narrowed as the oxygen content was diminished, until, with 14 per cent, of oxygen, the low limit was 6-5 per cent, of methane, and the high limit 6-9 per cent. The inert gas present was nitrogen. When the oxygen content was kept constant at 20 per cent., and part of the nitrogen replaced by 10 per cent, of carbon dioxide, the low limit was raised from 5-8 per cent, of methane to 6-2 per cent. When the oxygen was again constant at 20 per cent., the replacement of part of the nitrogen by a carbon dioxide content of 62 per cent, was required to raise the low limit to 8-8 per cent, of methane. Even when the oxygen was reduced to 17 per cent., there was no change in the inflammability of methane-air mixtures from the limit observed with 20 per cent, of oxygen, which was 5-8 per cent. In another series, the low limit of methane, with 19 per cent, of oxygen, was 5'5 per cent., 0-3 per cent, lower than was found by the first method of experimentation. With .17 per cent’, of oxygen, the low limit was raised to 5-7 per cent., but even with 13 per cent, of oxygen, the mixture was explosive with any methane content between 6-6 per cent, as the low limit, and 6-8 per cent, as the high limit. The second series of experiments again brought out the fact that a large amount of carbon dioxide was necessary to affect appreciably the limits. Clement thinks that the action of carbon dioxide in reducing the explosibility of methane-air mixtures can be explained by the high specific heat of that gas; he found that carbon dioxide was more effective in reducing explosibility than was nitrogen. Distribution of Methane in a Mine. Methane is never distributed uniformly throughout the mine atmosphere, being in some places scarcely or never detectable by analysis, but in so-called still air the variation may be most marke$. The rate at which methane and air mix depends tO' some extent on how the gas enters the air. If the gas comes from the upper part of a coal bed, or from the roof, it will usually mix more slowly with the air than if it issues from the lower part, and if it enters a steady moving air current, subject to few obstructions, it may tend to move in a layer along the airway for a considerable distance. If there is, in the roof of a gaseous bed, a cavity through which the air current cannot pass, firedamp may accu- mulate there in dangerous quantities. In such a pocket the bottom layers of gas are in contact with fresh air, which works its way into the firedamp, thereby diluting the gas somewhat, but the bottom layers of the mixture contain more air than the upper layers, because of the air moving past the mouth of the cavity. In such places in a gaseous mine, if the ventilation has been inter- rupted, the variation in methane content at any parti- cular time will be marked, ranging from an atmosphere containing a non-explosive proportion of methane, through cue that is explosive, to one that is non-explo-. sive again, because of too much methane and of the absorption of oxygen by the coal. Such a range is to be expected only in a sealed mine or portion of a mine. The edges of atmosphere in an unsealed part of the mine will ’be in contact with air of high oxygen con- tent, and will be more or less inflammable. Samples collected in different parts of a room in a mine wherein ventilation had been suspended because of an explosion (principally gas) that had occurred therein about two weeks before, showed that explosive proportions . of methane were pre- sent in the wurking face near the top. Seventy feet from the face, near the top, a high percentage existed, and half-way to the floor a much smaller proportion was found. Composition of Air in Different Parts of Mines. The oxygen content of mine air is almost always less than that of normal air, usually only a little less in the moving current, but sometimes much less at places * Clement, J. K., “The Influence of Inert Gases on Inflammable 'Gaseous Mixtures.” Technical Paper 43, U.S.A. Bureau of Mines, 1913, 24pp.