THE COLLIERY GUARDIAN AND JOURNAL OF THE COAL AND IRON TRADES- Vol. CV. FRIDAY. JANUARY 10, 1913 No. 2715. THE FLAMELESS OR 11 CONVERGENT ” COMBUSTION OF CASES. By Dr. Jean Meunier. [Specially Contributed.] An explosive mixture composed of suitable proportions of any combustible gas and air does not necessarily give rise to an explosion on ignition, but may burn quietly, in which case the combustion is flameless and the gaseous strata burn in contact with a hot, solid body, which is thus maintained in a state of incandescence. Various names have been applied to this type of combustion—“ flameless combustion,” “ combustion by incandescence,” “ surface combustion,” &c.; but in view of the molecular mechanism by which it Is produced, a more suitable name would be “ convergent combustion.” This molecular mechanism, in fact, enables a simple conception to be formed of the laws and properties of the phenomenon, which are of a most remarkable and unexpected character. Actually, this new mode of combustion is the exact antithesis of the explosive method in which ignition, starting from a certain point in the gaseous medium, is propagated with violence throughout the whole of the mass by divergence. It is curious to observe that the dynamics of combustion exhibit the same antithetical duality found in the majority of natural phenomena, which occur in both a positive and negative sense—as, for instance, electricity, magnetism, &c. The action of incandescent electrical conductors in firedamp mixtures, which has long formed the subject of investigation by the author in collaboration with M. Couriot, merely forms a particular instance of convergent combustion. The laws of the phenomenon can be most easily studied by the aid of platinum wire. For the investiga- tion, it is necessary to use a Bunsen burner, adjusted in such a manner as to burn with a perfectly blue flame. The platinum wire is inserted into the flame, and when it has become heated to redness, the rubber tube supplying the burner with gas is squeezed so as to extinguish the flame, and is then immediately released, thus re-admitting gas to the burner. In these circum- stances, the platinum wire glows with added vigour, and remains in this state of vivid incandescence without the gas becoming ignited. About 50 years ago, St. Claire Deville and Debray pointed out that a current of gas directed on a still hot platinum crucible raises it to red heat, and, in certain circumstances, itself becomes ignited ; but a point which entirely escaped these workers is that the proportion of air mingled with the gas plays a leading part in the phenomenon. At that time the ideas prevailing on the limits of inflammability, and on the properties of explosive gaseous mixtures, were of only a very vague character, and no one had any inkling of the mechanism of combustion. Incandescence depends essentially on the composition of the gaseous mixture concerned. To arrive at a clear conception of the matter it is necessary to examine the curve representing the properties of mixtures of gas and air in relation to their composition. Take, as an example, coal gas of average composition, the lower limit of inflammability of which is 10 per cent., and the upper limit 33 per cent., and mark on the axis of the abscissas the percentage proportions of gas from 1 to 100, the corresponding velocities of ignition being plotted on the axis of the ordinates. These velo- cities will evidently be represented by zero for propor- tions from 0 to 10 per cent., and the curve will coincide with the axis of the abscissas, but will rise suddenly when 10 per cent, is exceeded, indicating a rapid increase in the explosive energy. It attains a maximum at 18 to 20 per cent, (proportion of complete combustion) and thereafter diminishes, intersecting the axis of the abscissas at an angle of 45 degs. when 33 per cent, is reached, and preserves this direction in a straight line, of negative value, up to 100 per cent. This dip at an angle of 45 degs. is accounted for by the fact that, in order to bring, say, a 40 per cent, mixture to ignition (33 per cent.), there must be added to it an equal volume of a 26 per cent, mixture, the velocity of ignition of which is thereby neutralised. Mixtures containing more than 33 per cent, of gas have an extinctive action, and behave exactly like those in the flame and in convergent combustion. 50 60 0 iO u. o > Fig. 1. Fig. 2. 33 This affords a very simple means of tracing the curves of inflammability (fig. 1). At the upper limit of inflammability mark the straight line at an angle of 45 degs. (fig. 2), prolonging it above and below the axis of the abscissas, and set up the corresponding ordinate for the proportion of complete combustion—that is to say, 19 per cent, for coal gas—then joining the point of intersection to the point corresponding to the lower limit of inflammability, which is 10 per cent, for coal gas. Since in the vicinity of a maximum the variation of the function is slight, the apex of the triangle is rounded off, and the figurative curve thus obtained is in perfect harmony with experimental facts. Mixtures of combustible gas and gaseous medium of combustion may thus be grouped into three categories: (1) uninflammable mixtures, in which the proportion of combustible is too small and the properties of the medium are retained; (2) explosive mixtures; (3) extinctive mixtures, in which the properties of the com- bustible medium are masked. To plot the curves of inflammability of a com- bustible, it is sufficient to ascertain the two limits of inflammability, and to calculate the proportion which gives complete combustion. The author has devised a very simple apparatus (fig. 3) which enables the limits of inflammability of any combustible gas or vapour to be determined with ease. The field of explosibility lies between the positive portion of the curve and the axis of the abscissas. It is very small in the case of firedamp, the limits of inflammability—6 and 16 per cent. (?)—of which are very close together; but, on the other hand, is very wide in the case of hydrogen, oxygen, carbon monoxide and acetylene, which exhibits an anomaly in the lower section. In an ordinary flame, the mixture is of the extinctive or explosive category, according to its position; but at present we are not concerned with flames. With the preceding data, the peculiarities of con- vergent combustion can be explained without difficulty. Successive mixtures of the explosive type are pre- pared in a glass test tube into which, after inversion, is f®. 90 i8P 300 Ji 11 Fig. 3. Apparatus for Measuring the Limits of Inflamma- bility of Combustible Gases. Fig. 4. Exhausting and Trans- ferring the Gas. introduced a hot platinum wire (fig. 5). The incan- descence of the wire is the more vivid in proportion as the explosive character of the mixture is increased. Since this incandescence is persistent and the gas is stagnant, it follows necessarily that the different strata of the gas must be powerfully attracted toward the incandescent wire. All the phases of incandescence can be obtained simul- taneously by operating in a different manner. In this case an ordinary Bunsen burner (fig. 6) is provided with a glass cylinder, which plays the part of a lamp chimney and is fitted with a mica rim to protect the glass from breaking. The lamp is lighted to give a blue flame, and as soon as the platinum wire is hot it is lowered into the glass cylinder, and the flame is extinguished by com- pressing the rubber tube supplying the burner. The platinum wire continues to glow, and by adjusting the air supply the brilliance of the incandescence can be increased or diminished. This incandescence is in harmony with the above-mentioned curve of inflamma- bility, thus demonstrating that the intensity of con- vergent combustion is proportionate to the explosive intensity. To stop the incandescence, all that is required is to prevent the further admission of air, without touching the gas tap at all. It should be noted as an essential point that, for an explosive mixture of any given proportions, the incan- descence is greater in proportion as the wire is thinner. To demonstrate this, a number of platinum wires of different gauge (1 mm. to 01 mm.) are attached to a