INFL4MMABILTTY OF MIXTURES OF METHANE AND AIR 2591CCXLL-The Limits of Iri$arnmability of Mixturesof Methane and Air.By MAURICE JOHN BURGESS and RICHARD VERNON WHEELER.IN a recent communication (this vol., p. 1859) Coward and Brinsleysuggested that “inflammability mustl be regarded as a specificproperty, either present or absent; of any given gaseous mixtureunder definite conditions of temperature and pressure ” (p. 1884) ;(( independent of the shape and size of the vessel containing it, andindependent also of the nature of the means used for ignition inthe first place” (p. 1861).On the basis of this definition, a ‘(criterion” of inflammabilitywas put forward, which, stated briefly, demands that the true‘( limit-mixtare” must be regarded as that’ in which flame can bepropagated upwards-indefinitely.It is well known that mixtures of methane, for example, and airhave different.limits of inflammability, both higher and lower,dependent on the direction in which the flame has to travel.*Coward and Brinsley were probably led to their choice as t o whatconstitutes a true ( ( limit-mixture ” by consideration of the factthat a smaller percentage of inflammable gas is required (theydeal only with the lower-limit) for upward than f o r downwardpropagation of flame, and by the belief that the experimental condi-tions specified by them disclosed the least quantities of the com-bustible gases, hydrogen, methane, and carbon monoxide, that arecapable of forming with air mixtures in which self-propagation offlame can take place.It does not seem t o us desirable so t o restrict the use of the term‘( limit-mixtu.re.” The most important industry in which knowledgeof the limits of inflammability of gaseous mixtures is required iscoal-mining. The occurrence of firsdamp in mines constitutes, asis well known, one of the gravest dangers t o the industry; not somilch, i t is Selieved, because of the possibility of widespread explo-sions taking place in firedamp and air mixtures extending through-out the workings, modern legislation regarding the ventilation ofmines having rendered such occurrences exceedingly unlikely, butbecause a local accumulation of fire-damp, forming an explosivemixture with the air, may become ignited and transmit flame t o* This fact has been empliasised by Professor Enriqiie Hauser in a brochureentitled “ Leqons sur le grisou ” (Madrid, 1908).Hmser has summarised thedifferent results given by various experimenters for thc limits of inflammability ofmethane-air mixtures, and has offered an explanation of the differences2592 BURGESS AND WHEELER: THE LIMITS OFany fine coal dust that may be deposited on the roadways, and soproduce a widespread coal-dust explosion.Legislation has attempted to deal also with this danger, bystipulating that frequent analyses shall be made of the air of themine roadways and workings, and precautions taken t o preventthe percentage of methane contained in the air from exceeding acertain minimum. Despite precautions, accumulations of fire-damp,usually near the roof, sometimes occur; and, should sufficient fire-damp mingle with the ventilating current and by some mischanceencounter a sufficiently intense source of heat, what is te-chnicallycalled a “local ignition ” may occur.I n the majority of casesno’ damage or loss of life is caused by these “local ignitions,”which usually takc the form of a slowly-moving flame near theroof.Such cases of the propagation of flame in fire-damp-air mixturesare those moat freque’ntly reported. It will be realised that thedanger lies in the existence, over a considerable length of roadway,of a mixture in which a flame can travel horizontally. A similardanger arises when, as has been known t o happen, a fall of partof the roof liberates, and distributes in the ventilating-current, aquantity of fire-damp that has accumulated in cavities above theroof.,4 uniform inflammable mixture of fire-damp and air rarelyoccurs throughout any considerable area of a coal mine, but thereis an instance when a disused heading (or cul-de-sac), in which theair was practically stagnant, was slowly fed with fire-damp issuingfrom the neighbouring strata until, when the fact was discovered,the whole heading had become uniformly filled with a mixturecontaining 6 per cent.of methane. It has sometimes happened,also, that the slow ventilating current travelling through a roadleading from a goaf (or worked-out place) has been found to be aninflammable mixture, uniform in composition tliroughout itsextent.I n such cases the dangerous “ limit-mixtures” are such aswill allow of solf-propagation of flame throughout the wholemixture, whether ignition occurs a t the roof, floor, or centre of theroadway.The fire-damp of British mines consists of methane mixed withvarying proportions of nitrogen ; also, carbon dioxide, traces ofcarbon monoxide and ethylene may be present, and, occasionally,traces of ethahe have been detected.This paper records the results of determinations of the higher-and lower-limits of inflammability of mixtures of pure methanewith air, and shows how the compositions of the limit-mixtureFIG. 3.[ 7’0 ,face p . 2592FIG. 2.Prc:. 1. l ? I ( i . 4INFLAMMABILITY OF MIXTURES OF METHANE AND AIR. 2593differ, dependent on the manner in which ignition is effected.*These results are as follows :Methane, per cent.Lower limit. Higher limit.Central ignition in large globe .........5.6 14.8Vertical tube, closed at both ends :( a ) Ignition at bottom* ............ Not less than 5.40Horizontal tube, closed at both ends :Not more than 14.8'( b ) Ignition at top .................. 6.0 13-4I 5.4 (flame travels only( 5 . 6 (methane all burnt) j Ignition at one end ............ along top of tube) 14.3Coward aiid Brinsley obtained what appeared to be a self-propagating flame inone mixture containing, according to analysis, 5.3 per cent. of methane. They wereunable t o repeat the experiment.t The determinations made by IIauser (Zoc. cit.), using the Le Chntelier burette(downward propagation of flame), were : lower-limit 6-05 per cent ; higher-limit13-35 per cent.methane. Hwser obtained pure methane from alumininm carbide.The manner in which the flame travels in the different mixture'sis by no means the same, and an attempt has been made to indicatethe1 mere striking differences by diagrams. Fig. 1 represents theflam.e travelling upwards in a closed tube (6 cm. in diameter) coii-taining a 5.4 per cent. niethane-air mixture, aiid Fig. 2 the flametravelling horizontally in a like mixture. Figs. 3 and 4 illustratetwo stages in the progress of the flame downwards in a closed tubethrough a mixture containing 6.0 per cent. of methane.When an electric spark is passed at. the bottom of a closed vesselcontaining methane-air mixtures with 5.0, 5.1, 5.2, etc., per cent.of methane.flares of flame are produced which travel distancesincreasing with the percentage of methane, until with 5.4 per cent.of methane present the distance of travel reaches 2 metres in atube 2 metres long. It is possible that the flame in a 5.4 per cent.mixt'ure might travel upwards inoze than 2 metres; it might travelindefinitely, and, on the other hand, it might die out after adistance of 3 metres o r less. Since i t is obviously impossible t omake a crucial experiment' t o test this point, i t must suffice t orecord that the lower-limit mixture for upward propagation offlame contains not less than 5.4 per cent. of methane. The flamein mixtures containing 5.35 per cent.of methane, contained in aclosed glass tube 6 cm. in diameter, never exceeded a distance oftravel of 50 cm.; in 5.3 per cent. and 5-25 per cent. mixtures theflame travelled 40 cm. m d 30 cm. respectively.It will be seen in Fig. 2 that the flame travelling horizontally in* The influence of added nitrogen 011 the limits, as determined By centralignition in a large globe, is described in a subsequent paper2594 BURGESS AND WHEELER: THE LIMITS OFa 5.4 per cent. mixture occupies only the upper quarter or thirdof the containing vessel, and analysis of the mixture left in a tube6 cm. in diamet'er after the flame had travelled along i t showed itto contain 3-25 per cent. of methane. When a tube 10 cm. indiameter was used, 3.9 per cent. of methane remained in theproducts of combustion.Presumably, if a large room were filledwith such a mixture and a light applied a t some point near thefloor, a column of flame would travel upwards from the point ofignition to the roof and spread along the upper portion only ofthe room.It may not be generally known that when a slow current of aircontaining between 4 and 5 Fer cent. of methane passes over a smallflame, such as that of an oil lamp, the cap or aureole that ordin-arily forms round the lamp-flame may become detached and floatalong with the current. Similarly, if a slow current of air contain-ing 3 or 4 per cent. of coal gas is allowed to ascend a vertical tube,and a succession of electric spark3 passed a t the bottom, small capsof flame can be caused to pass from end t o end of the tube, onlya fraction of the mixture being burnt.Of a like nature aro such mixtures of hydrogen and air as allowI' balls " of flame to travel (perhaps indefinitely) upwards throughthem-with the convection-current-in the manner described byCoward and Brinsley.From the practical point of view such mixtures are not in them-selves dangerous, but they are potentially dangerous in that theycould convey flame to richer mixtures.From the scientific point of view it is or' interest €0 know thatsuch curious flames can be produced, but the doubt must alwaysremain whether actually '' indefinite " propagation of flame cantake place in the mixtures that exhibit them; and the fact thatchemical reaction proceeds in but a small portion of t'he wholemixture, renders it difficult to employ f o r comparative measure-ments the " criterion " of inflammability adopted by Coward andBrinsley.%* It should be noted that, making use of their criterion, Coward and Brinsleydecided upon 4'1 per cent.as the lower-limit percentage of hydrogen in air(3'8 per cent. of hydrogen remaining in the mixture after the flame had passed).Dixon and Crofts (this vol., p. 2047) have obtained the following figures for theignition-temperatures of various mixtures of hydrogen and oxygen :2H,+ 0, ............... 526'2H,+ 202 ............... 5112H,+ 80, ............... 4782H2+160, ............... 4722H,+3202 - ...............The last niixture, which contained 5.88 per cent.of hydrogen, could not beignited, although five expcrirnents were made in which the temperatures reacheINFLAMMABILITY OF MIXTURES OF METHANE AND AIR. 2595It may be remarked that if the object of their criterion ofiiiflammability is to specify the least quantity of inflammable gasin air that is capable of propagating flame, that object is notattained. For when a mixtare of methane and air containing5.0 per cent. of methane, enclosed in a 4-litre globe, is stronglyagitated by revolviqg a small fan therein, and an electric sparkis passed a t the' centre of the globe, flame travels rapidly throughoutthe mixture, all the methane being burnt. The lower-limit ofinflammability of " agitated " mixtures of methane and air couldtherefore be stated t o be 5-0 per cent., or possibly less with moreviolently agitated mixtures.The flame travelling downwards in a 6.0 per cent.mixture,depicted in Figs. 3 and 4, and the flame started centrally in a5.6 per cent. mixture contained in a large globe, both burn themethane completely. I n both cases the line of demarcation betweena mixture that will propagate flame and one that will not is sharplydefined.*For the latter reason, as well as for their convenience, thecriteria of inflammability for downward propagation, used byLe Chatelier and universally applied by French mining engineersf o r the1 measurement of fire-damp ; and for propagation throughouta globe, adopted by several investigators, would seem to commendthemselves f o r such comparative rneasurements as a study of thepropagation of flame in gaseous mixtures may demand.EXPERIMENTAL.The methane used was prepared from aluminium carbide, andwas purified from traces of acetylene by passing through ammmia-cal cuprous chloride, and from hydrogen by passing slowly over" oxidised '' palladium precipitate heatemd at 90°.It contained99.8 per cent. of nxhhane.The explosion-vessels were glass cylinders, sealed a t each end,were 500", 600", 700", lOOO", and 1700" respectively. Dixon and Crofts consideredthat the limit of inflammatdity had been passed in the last mixlure, but they notedthe fact that combuvtion occurred of part of the mixture.* The following deterrninations of the lower-limits, downward propagation,were made by Alr. A. Whitaker. With mixtures containing, according Lo aiiallysis,6.03, 6.02 and 6.01 per cent. of methane, flame was propagated throughout thelength of a closed tube 7 cni. in diameter and 160 cm. long. With 5-99, 5-98 and5.97 per cent. of methaiie the flame travelled 12 or 15 cm. only.A volume of mixture sufficient for two experiments was made containing 6.00per cent. of methane. After one experiment, i n which propagation of flame wascomplete, sufficient air was added to thc remainder of the mixture to reduce thepercentage of methane to 5'99; flame travelled downwards only 16 cm. in thismixture2596 BURGESS AND WHEELER: THE PROPAGATION OF FLAME TN6 cm. in diameter and 2 metres long. They were fitted withplatinum firing points a t one end and a t the' other with a three-way tap, through which the mixtures t o bO experimented withwere intsoduced after the cylinders had been exhausted of air. Themixt.ures were made over glycerol and water in graduated glassgas-holders, and were all analysed before use.Es I; YE A LS,CUMBERLAND