36 PAYMAN AND WHEELER THE PROPAGATION OF III.-The Pqropagation of Flarrie through Tubes of Smull Diameter. Part 11. By WILLIAM PAYMAN and RICHARD VERNON WHEELER. IT is a common practice a t collieries to test’ the safety of the miners’ flame lamps before they are taken underground by intro-ducing them into an inflammable mixture of coal-gas and air. It is known that the speed of propagation of flame in mixtures of coal-gas and air can be considerably faster than in any mixture oIf methane and air. Since any inflammable mixture into which a miner’s lamp may accidentally be introduced in the practice of coal mining is produced exclusively by firedamp and since it is a rare occurrence f o r firedamp to contain even traces of any in-flammable gas other than methane the use of mixtures of coal-gas and air for testing the security of a lamp for use underground is justifiable oaly on the grounds of providing an adequate “margin of safety.” The use of coal-gas becomes unjustifiable if the margin of safety thereby provided is excessive; for every additional pro-tective device embodied i n the construction of a miner’s flame safety-lamp militates against the proper ventilation of the lamp, and theref ore diminishes its lighbgiving power.It is thus of importance to be ablel to make an exact comparison between the speeds of propagation of flame in mixtures of coal-gas and air and fire-damp or methane and air under similar conditions of experiment. Furthermore since the rapid speed of flame in coal-gas-air mixtures is presumably due mainly to the hydrogen contained therein and since different qualities of coal-gas contai FLAME THROUGH TUBES OF SMALL DIAMETER.PART 11. 37 different proportions of hydrogen i t is necessary to obtain infoma-tion regarding the effect of varying the proportions of the con-stituent gases in coal-gas on the speed o€ propagation of flame in its mixt’ures with air. Following the same methods of experiment as with mixtures of fire-damp and air (T. 1918 113 656) the speeds of the uniform movement of flame in mixtures with air of coal-gas hydrogen and a 1 1 methane-hydrogen mixture have been determined in glass tubes of different small diameters for comparison with the results obtained with fire-damp-air mixtures in similar tubes. Compara-tive expe’riments have also been made on the projection of flame through brass tubes of small diameter.For the experiments with mixtures of coal-gas and air a supply of gas from the main was stored over alkaline water in a metal gas-holder of 70 litres capacity; and the mixtures with air were made in smaller glass gas-holders from this supply. I n this manner variations in the composihion of the coal-gas such as would have occurred had the gas for each mixture been drawn direct from the main were avoided. Rather more gas was required to complete the series of experi-ments than was anticipated so that it was found necessary t o re-charge the storage-holder before all the information desired was obtained . From one point of view this was unfortunate for the second charge of gas differed slightly in composition from the first and mixtures with air of the one could not be directly compared with mixtures with air of the other.From another point of view how-ever the enforced use of samples of coal-gas of different composi-tions was not to be regretted for there were found to be marked differences in the speeds of propagation of flame in mixtures with air of the two qualities of gas. This observation led a t once to the determination of the speeds of flame in mixtures with air of what may be termed a “synthetic coal-gas,” containing equal parts by volume of methane and hydrogen. The results obtained taken in conjunction with the known values for methane-air and hydrogen-air mixtures under the same conditions of experiment are of con-siderable theoretical interest.whilst’ they should also prove of practical value. According to Le Chatelier (“Le Carbone,” p. 266. Paris 1908), if several combustible gases are mixed together with air the follow-ing relation exists a t the lower limit$ of inflammability of the mix-ture between the limits of inflammabilit,y N and N f of each of two gases and their proportions n and n’ in the limit mixture: 7 2 p f T L f p V ’ = 1 38 PAYMAN AND WHEELER THE PROPAGATION OF Coward Carpenter and Payman have shown (this vol. p. 28) that this formula can be applied with considerable accuracy to a numbes of mixtures of gases and that it holds also a t the upper limit of inflammability. The formula implies that if a limit mixture with air of one inflammable gas is mixed in any proportion with a limit mixture with air of another inflammable gas a limit mixture results.Another way of stating the relation in an expanded form is as follows : By means of this equation the limiting percentage L for a mixture of gases can be found directly from the known limits of the individuals; a b . . . being the percentages of the individuals in the mixed inflammable gases and La Lb . . . their respective limits. The subject of the calculation of the limits of inflammability of mixed combustible gases is introduced here because a similar formula holds with remarkable accuracy (considering the nature of the phenomena under investigation) for calculating the speeds of flame in mixtures with air of a composite combustible gas like coal-gas the speeds in mixtures of the individual gases with air being known.The formula is: . . . . . (ii) a + b + . . . . . s - a/S,+b/Sb+ . . . . in which S is the speed required; a b . . . t.he percentages of the different combustible gases in the mixed gas (coal-gas for example) ; and Sa Sb . . . the corresponding speeds of flames in mixtures of the individuals with air. This formula necessarily finds its readiest application in the calculation of the speeds in distinctive mixtures namely (1) the limit mixtures upper and lower in which the speed of flame is slowest; and (2) the mixtures in which the speeds of flame are fastest. For such mixtures the agreement between calculat.ed and observed speeds is close. I n the table that follows are given (1) the limits of inflamma-bility with horizontal propagation of flame in a glass tube 9 mm.in diameter for hydrogen methane and a mixture of equal parts of hydrogen and methane; and (2) the speeds of the uniform move-ment of flame in a ho'rizontal glass tube 9 mm. in diameter with the lower- and upper-limit mixtures and in the mixtures with the fastest speeds o t flame for hydrogen methane and the 1 1 hydrogen-methane mixture. The calculated limits and speeds fo BLAME THROUGH TUBES OF SMALL DIAMETER. PART II. 39 the hydrogen-methane mixture as determined by equations (i) and (ii) respectively are also given. Speeds of uniform movement of flame. Cm. per second. I Lower-Limits. Per cent. limit Lower. Upper. mixture. Hydrogen ............ 6-7 65.7 8-3 Methane ............7.8 11.6 32.6 Hydrogen- 1- (Obs.) 7.2 19.6 13.7 methane mixture J (ca~c.) 7.2 19.7 13.2 Mixture with Upper-fastest limit speed. mixture. 430 49 96 17.1 (;;!a 90 (42) The value obtained for the speed of propagation of flame in the upper-limit mixture of hydrogen and air is iiot the true value, which should approximate t o that of the speed in the lower-limit mixture. The probable reason for the discrepancy is explained later. Omitting this value and the calculated value f o r the speed of flame in the upper-limit mixture of hydrogen-methane-air based on it it will be seen that there is a close correspondence between the calculated and the observed values for the limits and speeds. With coal-gas the gases that preponderate are hydrogen and methane which in the two samples A and B used foc these experi-ments totalled 83.5 and 85 per cent’.respectively. Ignoring the other gases calculation according to equation (ii) gives 106 and 96 an. per second respectively as the maxirnum speed obtainable during the uniform movement’ of flame in a tube 9 mm. in diameter in mixtures of each sample of coal-gas with air. The speeds as detekmined by chronographic means were 106.2 and 94 an. per second. The proportion of mixed gases to be added to air to give inixtures with thel fastest speed of the uniform movement of flame can also be calculated knowing the corresponding values for each individual gas. The fastest speed with mixtures of methane and air is obtained over the range 9*5-10*0 per cent.of methane; with mix-tures of hydrogen and air the range is 38-45 per cent. of hydrogen. Using the same type of formula as for calculating the limit mixtures the “fastest speed mixtures’’ of air with mixed combustible gases are found t o be as follows: Mixtures with air in which the speed of the uniform movement of flame is fastest. Per cent. o f combustible gas. Combustible gases. Calculated. Observed. H y drogen-methane, Coal-gas A ............... 17.5-18.8 184-19.0 Coal-gas H ............... 16.3-17.9 16-5-17-5 (1 1) .................. 15*2-16*3 15*0-16* 40 PAYMAN AND WHEELER THE PROPAGATION OF If equation (ii) is expressed in the form - a+b+ . . . . a/S,+b/Sb+ . . . . - -#a,+b+ . . . . * i t is a t once apparent that the inverse of the speed of flame in mixtures of a composite gas with air is a simple additive property of the inverse of the speed of flame in each constituent gas with air.I n other words the time taken for flame t o spread through a given volume of a mixt'ure of combustible gases with air under the conditions of combustion during the uniform movement is the mean of the times taken for flame t o spread through the same FIG. 1. j I Combustible gas in air. Per cent. volume of mixtures of each constituent gas with air if present alone. No doubt this relatioln which has been shown to hold true for the fastest and the slowest speeds of the uniform movement of flame is true also as suggested by the generalisation just stated, for intermediate speeds. So that given the necessary data respect-ing the individual combustible gases the behaviour of flame in any mixture of seve'ral with air can be deduced.The exposition of the validity or otherwise of this assumption when three or more combustible gases in varying proportions are used will form the subject of a subsequent communication. I n Fig. 1 are shown plotted to the same scale the speed-per FLAME THROUGH TUBES OF SMALL DIAMETER. PART 11. 41 centage curves for the uniform movement' of flame in a horizontal glass tube 9 mm. in diameter for hydrogen and methane for a 1 :1 mixture of hydrogen and methane and f o r coal-gas (sample A ) . The curve for hydrogen is constructed from Haward and Otagawa's determinations (T. 1916 109 SS) with additional figures obtained near and a t th0 limiting percentages.I n this connexion it should be noted that Eaward and Otagawa though they made no attempt to determine accurately the limits of ill-flammability f o r horizontal propagation of flame considered that in a tubs 9 nun. in diameter flame would not travel horizontally in mixtures containing less than 11.8 or more than 63.5 per cent. of hydrogen. Actually the 1imit.s under the conditions thus specified are 6.7 (lower) and 65.7 (upper) per cent. It was found that when igniting mixtures near the limits of inflammability, great care had to be exercised to avoid disturbance a t the mouth of the tabe and for that reason a lighted taper such as was employed by Haward and Otagawa which answered admirably over the range of mixtures studied by them was unsuitable.* * The details of the determinations made to locate the limits of inflam-mability of hydrogen-air mixtures in a horizontal tube 9 mm.in diameter are as follow the tube was 1.5 metres long and the mixtures were ignited by a secondary discharge across a 5 111111. gap 4 cm. from the open end of the tube. Lower limit. Hydrogen per cent. Result. 9.4 Flame travelled throughout. 7.5 Y Y I9 ?, 7-1 99 Y 9 , 6.8 9 9 97 ,? 6.6 Incomplete propagation of flame. These results place the lower limit at 6-7 per cent. hydrogen. A mixture of this composition when tested failed three times to propagate flame but on five occasions flame travelled throughout the length of the tube. The flames travelled very slowly and were only visible when the room was in complete deskness.Upper limit, Hydrogen per cent. Result. 63.5 Rapidly moving flame throughout. 64.5 9 9 9 9 ?? 65.0 I # ?? Y 9 65.3 9 9 9 97 With mixtures containing 66 per cent. or more of hydrogen a sharp report occurred on sparking due to the rapid combustion of a mixture made poorer in hydrogen by diffusion between the point of ignition and the open end of the tube. Flame also travelled rapidly over short distances towards the closed end of the tube; thus with 67.5 per cent. hydrogen the flame travelled 25 cm. and with 67.8 per cent. 10 cm. With 68 per cent. no flame could be C 42 PAYMAN AND WHEELER THX PROPAGATION OF Attention should be directed to the slow speed 8.3 cm. per second a t which flame could travel in a mixture of hydrogen and air a t the lower limit a fact which illustrates the well-known persistence of hydrogen flames.I n conformity with the results obtained f o r other gases it was expected that the flame in the higher-limit mixture would be equally slow. The fact that so high a speed as 50 cm. per second was recorded was due to the difficulty experienced in igniting the mixture before diffusion a t the mouth of the tube could decrease the percentage of hydrogen there. The coal-gas was made a t the carbonising plant- of the Experi-mental Station and was not diluted with water-gas. The analyses of the two samples were: Sample A. Per cent. I3enzene and higher olefines ......... 1.2 Carbon dioxide ........................... 0.1 Carbon monoxide ........................ 7 . 3 Hydrogen .................................50.6 Methane and higher parafins ...... 32.9 Nitrogen (by difference) ............... 5.0 Ethylene ................................. 2.9 Sample B. Per cent. 1.6 nil 2.8 7- 1 47.0 38.0 3.5 It will be seen that the difference between the two samples of gas lay almost entirely in the proportions of hydrogen and paraffins that they contained; and from a comparison between the speed-percentage curves for coal-gas A and the hydrogen-methane mixture given in Fig. 1 it is evident that the slower speed of flame obt,ained with coal-gas B as compared with coal-gas A is due to t.he higher methane-content of the former. For the highest speed obtainable with the hydrogen-methane mix-ture which contained 50 per cent.of each constituent is consider-ably slower than the highest? speed obtainable with coal-gas A , which also contained 50 per cent. of hydroger but only 33 per cent. of methane. The results obtained on the propagation of flame in horizontal glass tubes of smaller internal diameter than 9 mni. are recorded in the tables that follow. The determinations were made in the same manner and the numbers in the table have the same signifi-cance as in the experiments with mebhane and air (Zoc. cit. p. 658), with which they should be compared. observed t o travel away from the open end of the tube. These results place the upper limit for self-propagation of flame between 05.3 and 66.0 per cent. hydrogen. The determination was completed as follows : Hydrogen per cent.Result. 65.9 65.7 ¶ * 3 9 65.6 Flame travelled rapidly about 40 cm. Complete propagation of ff ame. 9 FLAME THROUGH TUBES OF SMALL DIAMETER. PART 11. 43 Internal diameter of tube, nun. 2.0 3.0 4.2 5.0 6.0 7.1 8.0 Internal diameter of tube, mm. 2.0 3.0 4.2 5.0 6.0 7.1 8.0 Internal diameter of tube, 111111. 8-45 2-0 nil 3-0 -4.2 -5.0 -7.1 (30) 8.0 (40) TABLE I. Coal-gas Sample A . Coal-gas in mixture. Per cent. - 13.1 16.4 17-5 18.4 19.2 20.0 22.7 24.5 nil nil nil nil nil nil nil nil (50) - - 81.3 - - nil* -49.6 - - 87.3 - - nil* -35.3 - 53.5 - -55.5 - - 99.1 - - 40.6 nil* 58.7 86.8 97.7 103.7 102.7 90.3 50.6 nil* 98.1 - -59.2 87.4 98.8 t104.2 103.9 91.8 57.0 29.8 7 9.45 nil nil 38.7 39.4 -(28) TABLE 11.Coal-gas Sample B. Cod-gas in mixture. Per cent. 13.1 14.0 14.8 16.5 16.8 nil nil nil nil nil 60 nil (35) (35) (41) (52) (57) - - -- - - -- - - - - - - -62.2 71.4 84.1 80.1 63-2 77-4 85.8 82.3 TABLE 111. Bydrogen-Hethane 1 1, Hydrogen-methane in mixture. Per cent. A v 11.90 13.90 14-35 14.95 15-95 17.20 18.10 18.65 19.55 nil nil nil nil nil nil nil nil nil 56.3 74.9 - 74.5 73.0 nil* - - -57-9 75.9 - - - 46.1 nil* - -59.4 83.1 - 86.0 84.0 53.5 nil" - -63.0 85.1 88.0 95.8 94.1 74.1 43.3 nil* -69.5 87.1 90.7 - - 75.8 49.4 34.2 nil* * Flme travelled towards the open ends of the tubes a distance of 3 cm. The results recorded in table 111 are shown as smoothed curves in Fig. 2 which illustrates the extent to which the "limits" are dependent on the environment of the inflammable mixtare.It will be seen that the range of mixtures over which continued (hori-zontal) propagation of flame was possible became gradually restricted as the diametesr of the tube was decreased until with a c" 44 PROPAGATION OF FLAME THROUGH TUBES OF SMALL DIAMETER. 3 mm. tube it was less than half of that obtaining in a 9 mm. tube. With a tube 2 mm. in diameter no flame could travel away from the point of ignition whatever the percentage of combustible gas present an observation that applies also to the mixtures of coal-gas and air. With all mixtures of methane and air a diameter of 3.6 mm. prevented tGhe propagatdon of flame; whereas with hydrogen and air (30 per cent.hydrogen) Mallard and Le Chatelier (Ann. des Mines 1883 [viii] 4 320) have recorded’ the propagation of flame in a glass tube only 0-9 mm. in diameter. Complementary to these results are the results of experiments loor-401. 1 8 FIG. 2. I -T 0 12 Combustible gas in air. Per cent. made on the passage of flame in mixtures of coal-gas and air through brass tubes either open at’ both ends or arranged as extensions to a larger vessel a t the closed end of which the mix-ture was ignited. It is unneceesary t o give the debails of these experiments which were conducted in the1 same manner and ex-hibited the same general features as the experiments with methane. It is sufficient to record that the flame in an 18 per cent. mixture of coal-gas B and air passed through 15-18 crn.length of brass tube of 4.4 mm. internal diameter placed horizontally and open a t both ends and was projected from a closed veesel (20 cm. long) through 13-15 cm. length of the same brass tubing. The corre MIXTURES OF NITROGEN PEROXIDE AND NITRIC ACID. 45 sponding distances when a 10 per cent. mixture of methane a d air was used were 7.5 and 3-4 cm. respectively. The general conclusion to be drawn from these experiments as regards the testing of miners’ safety-lamps is that coal-gas ” is an unsuitable gas t o employ for that purpose for the following reasons : (1) Comparatively small variations in the composition of coal-gas affect! the speed a t which flame can travel in its mixtures with air. I n particular a reduction in the proportion of paraffins which it contains such as is usually accompanied by an increase in the proportion of hydrogen when as generally carburetted water-gas is employed to dilute the coal-gas enables a much higher speed of flame to be attained than can be given by mixtures of methane and air. (2) Even with gas produceld solely by the carbonisation of coal a t normal retort temperatures the speed of propagation of flame attainable is more than double that possible in mixtures of methane and air. (3) It would seem that the ability of flame to pass through tubes or holes of small diameter is not dependent alone on its speed, although this is the main factor but is t o a certain extent a quality of the inflammable gas concerned. Flame in mixtures of hydrogen and air possesses the property of being able to pass through holes of very small diameter and the presence of hydrogen in coal-gas confers this property in a certain degree on the flame in mixtures of the latter with air. ESKMEALS , CUMBERLAND. [Received October 30th 1918.