Gaseous Hydrocarbons

Amongst the earliest experiments carried out with a view to the quantitative determination of the limits of inflammability of combustible gases were those of Davy with fire damp, which is mainly methane, CH₄. Owing to the importance of this gas in connection with gob fires and explosions in coal mines, several other workers have also investigated it. The value of the results, however, is restricted by the fact that firedamp, like most natural products, is subject to very considerable variation in composition. Even Davy recognised that it was not pure methane; indeed, perfectly pure methane is not easy to prepare in quantity. The gas, as obtained from sodium acetate, may contain as much as 8 per cent, of- hydrogen, as well as ethylene. No doubt this variation in composition is one contributory cause of the very varied results.

Of these results the most reliable are those of Burgess and Wheeler, and of Coward and his co-workers, which may now be briefly considered.

Apparatus used by Burgess and Wheeler (1911).

Several methods of attacking the problem were devised, namely, central ignition in a large glass globe by means of an electric spark; employment of vertical tube, closed at both ends, and ignited either at the bottom or at the top; and a horizontal tube, closed at both ends, ignited at one end.

1. Central Ignition in a Large Globe. - The apparatus consisted of a glass globe, of capacity about 2 litres, fitted with platinum electrodes having looped ends. The electrodes passed along a diameter of the globe through ground stoppers. The gases could be admitted and withdrawn through the tap. In all but a few special experiments a little distilled water was placed in the globe to saturate the gaseous mixture with water-vapour at the room temperature. The methane was obtained in a state of high purity by the action of an alu minium - mercury couple on a well-cooled mixture of methyl alcohol and iodide. The product was freed from traces of hydrogen by passage through "oxidised" palladium sponge heated to 98° C. and by subsequent liquefaction with liquid air.
   
   The manner of determining the lower-limit mixtures was that of "trial and error"; for example, a mixture of methane and air containing 6.1 per cent, of methane having been tried and found to propagate inflammation on the passage of an electric spark, a second mixture was prepared containing 5.9 per cent, of methane. This also propagated flame. The percentage of methane was therefore further reduced by 0.10 in a new mixture, and so on, until two mixtures were obtained, differing in their content of methane by 0.20 per cent., one of which enabled flame to be propagated, whilst the other did not. The lower- limit mixture was taken to be that containing the mean percentage of methane contained in these two mixtures.
   
   

   Lower Limit of Inflammation of Methane in Air (Burgess and Wheeler, 1911).

   The lower-limit mixture could be distinguished with certainty from that just containing sufficient combustible gas; for the momentary passage of the electric spark sufficed to promote the inflammation of all the gas contained in the globe in the former case, and on further sparking no signs of combustion could be observed. Whilst in the latter case, although the flame of the burning gas might appear to travel nearly through the whole mixture on the first passage of the spark, and some doubt might exist as to whether it had not, in fact, travelled throughout, on causing the spark to pass a second time, a "cap" appeared above it, showing that the mixture still contained combustible gas. This cap remained whilst passage of the spark was continued, growing gradually smaller in size, until all the gas had been burnt.
   
   All the experiments were made in a darkened room, so as to enable the appearance of the flames to be readily observed.
   
   The appearance is beautifully shown in fig.
   
   The following results were obtained:
   
   Lower-limit mixtures
   
   

   Gas or Vapour.	Lower Limit Mixture in Air. Per cent, by Volume.
   Methane	5.50 to 5.70
   Ethane	3.00 to 3.20
   Propane	2.15 to 2.30
   Normal-Butane	1.60 to 1.70
   Normal-Pentane	1.35 to 1.40
   iso-Pentane	1.30 to 1.35

   
   
   The value for methane agrees quite well with that obtained both by Coquillon and Le Chatelier; for firedamp there are at least five factors concerned in the magnitude of the lower-limit mixture. These are:
   1. The heat of combustion of the gas;
   2. The relative volumes and specific heats of the diluent gases;
   3. The ignition temperature of the mixture;
   4. The initial pressure; and
   5. The initial temperature.
   As Burgess and Wheeler point out, the first factor must undoubtedly exercise the preponderating influence. The last two factors can easily be kept constant.
   
   As a first approximation, therefore, it would appear probable that the lower limit of inflammation should vary inversely as the calorific value of the gas; that is to say, if L is the proportion of the combustible gas necessary to form a lower-limit mixture, and C its calorific value,
   
   L = f(l/C) or L = k/C
   
   where k is a constant.
   
   The value obtained for the lower limit of inflammation of methane when mixed with air is 5.6. The calorific value of methane is 189.1. Substituting in the above equation a value of 1059 is obtained for k, and the relative values for L for other gases can then be calculated.
   
   Lower-limit mixtures
   
   

   Gas or Vapour.	Heat of Combustion	L observed	L calculated.
   Methane	189.1	5.60	standard
   Ethane	336.6	3.10	3.15
   Propane	484.2	2.17	2.19
   n-Butane	631.7	1.65	1.68
   n-Pentane	779.2	1.37	1.36
   iso-Pentane	779.2	1.32	1.36

   
   
   The agreement between the observed and calculated values of L is very striking, and seems to point to a definite and dominating relationship between the calorific values of the combustibles named in the table - the paraffin hydrocarbons - and their lower limits of inflammation when mixed with air.
   
   But when this method of calculation is applied to other combustible gases, the agreement is not so close. This, however, is scarcely surprising, for there is no reason why L should be actually in direct proportion to C, and this is the basic assumption of the calculations.
2. Vertical Tube closed at Both Ends. - In these experiments glass cylinders 6 cm. in diameter and some 2 metres in length are employed, and the combustible mixture is fired by sparking between platinum terminals at either the upper or the lower end. Bottom ignition tends to give a low value for the low limit and a high value for the higher, limit of inflammation, in consequence of convection currents. Top ignition behaves in an exactly opposite manner. This is evident from the results in the table below.
   
   The quantitative values for the higher and lower limits are greatly affected by the diameters of the tubes unless these exceed 5 cm. Narrow tubes raise the lower limit and depress the higher limit, thereby reducing the total range of inflammability. This is illustrated in the case of mixtures of acetone and air in the accompanying table. If the diameters are very small indeed, no combustion will take place.
   
   Influence of Diameter of Tube upon the Limits of Inflammability
   
   

   Diameter, cm.	Lower Acetone Limit.			Higher Acetone Limit.
   	Upward.	Downward.	Horizontal.	Upward.	Downward.	Horizontal.
   2.5	2.30	2.75	2.40	7.5	6.5	6.7
   5.0	2.20	2.40	2.25	9.5	8.3	9.3
   10.0	2.15	2.35	2.20	9.7	8.5	9.5

3. Horizontal Tube, closed at Both Ends. - The ignition is effected at one end as already indicated, and the flame creeps along the upper portion of the tube in a similar manner to a small air bubble in a spirit-level.
   
   The diameter of the tubes, if less than about 5 cm., exerts an analogous influence upon the results, as was found to be the case with vertical tubes. This is shown by the data in the above table.

In carrying out some of these a bottle of capacity 11 litres was chosen, and fitted with a rubber stopper conveying gas and water tubes and insulated leads for the sparking wires. The gases were admitted by displacement of water, the air being first added in nearly sufficient amount, followed by a measured volume of inflammable gas. Finally, the total volume was brought to 10 litres by the addition of sufficient air. The water remaining in the vessel, amounting to about 1 litre, enabled the gases to be thoroughly admixed by shaking. This procedure necessitated the experiments being confined to gases saturated with moisture. In the case of gases of but slight solubility mixtures could be made accurately to one part per thousand. For upward ignition, the bottle was supported upside down on a tripod stand and the sparking gap extended about 2 cm. above the surface of the water. When downward ignition was required, the bottle stood on its base. As the beautiful vortex rings of flame rising through the mixture were soon extinguished owing to the limited capacity of the bottle it was impossible to decide whether such rings were capable of travelling indefinitely, or if they would tend to break up and produce a general inflammation, or eventually become extinguished. A long, rectangular tube was therefore constructed, two sides of which were of wood and two of plate glass. Square in cross-section, and of total length 1.8 metres, its capacity was 170 litres. Its top was of wood, and its bottom open, but water-sealed during experimentation by immersion in a tank. The gases were admitted by displacing water, and ignited electrically. An analysis of the gaseous mixture was made, as a check on the accuracy of mixing, just prior to the test of its inflammability. The methane was prepared from aluminium carbide and water, with subsequent removal of acetylene and liquefaction to separate hydrogen.

Limits of Inflammation of Methane-Air and Ethylene-Air Mixtures

	Methane per cent, by Volume.			Ethylene per cent, by Volume.
	Lower Limit.	Upper Limit.	Lower Limit.	Upper Limit
Central ignition in large globe	5.6	14.8	. . .	. . .
Vertical tube closed atboth ends: a. Bottom ignition b. Ignition at top	not < 5.4 6.0	not > 14.8 13.4	3.3 3.6	25.6 13.7
Horizontal tube closed both ends	5.4 flame travels only along top of tube	14.3	3.4	14.1
Ignition one end	5.6 all CH4 burned	. . .	. . .	. . .
Under slight shock	5.6	. . .	. . .	. . .
Perfectly tranquil	5.3	. . .	. . .	. . .
Do	. . .	15.4	. . .	. . .

With a 5.1 per cent, methane mixture in this large box, a ring of flame was formed which travelled about 30 cm., broke quickly, and formed a tongue of flame which travelled another 30 cm. before extinction. With 5.3 per cent, methane a stout ring of flame travelled a few cms. and resolved itself into a steady flame nearly as wide as the box, travelling right to the top with a swaying motion. This experiment could not be satisfactorily repeated, and even a 5.5 per cent, mixture failed to yield a flame sufficiently strong to traverse the whole box; but with 5.6 per cent, of methane, a steady flame with a convex front passed throughout the whole mixture. The authors therefore conclude that the flames of mixtures containing 5.3 to 5.6 per cent, of methane are very sensitive to extinction by shock, and that a 5.6 per cent, mixture will invariably propagate flame when the shocks are no greater than those occasioned by the somewhat violent bubbling of gas through water. When, however, circumstances are such that a tranquil passage is assured, 5.3 per cent, is the lower limit of in flammability.

The foregoing data, in so far as methane and ethylene are concerned, may be summarised as shown on in the table.