9 The Reaction between Gas and Pole in the Electrical Ignition of Gaseous Mixtures .
By Prof. W. M. Thoknton , D.Sc .
, D.Eng .
, Armstrong College , Newcastle-upon-Tyne .
( Communicated by Sir Charles Parsons , K.C.B. , F.R.S. Received June 21 , 1915 .
) 1 .
Introduction .
Every inflammable gaseous mixture has a definite lower electrical limit , of current or voltage or capacity , at which ignition fails .
Previous measurements on this have been confined to the observation of least sparking distance or lowest gas pressure at which inflammation ceases to be possible , and in almost every case the sparks used have been disruptive discharge between platinum or aluminium points .
The results of Emich * who made the first systematic observations , led to the view that ignition is independent , or nearly so , of the metal of the poles between which the igniting spark is produced , and this is generally accepted .
There appear to be no previous observations of ignition by low voltage condenser discharge , or by break-sparks , as being affected by the metals of the poles .
The present paper is an examination of typical reactions : in the case of break-sparks , at atmospheric pressure and with the percentage of combustible gas in air varied ; in the case of long and short jump-sparks , with a constant mixture and varied gas pressure .
The two gases chosen for the purpose , using break-sparks , were ethane and carbon monoxide , both of which could be obtained pure .
Ethane is a better representative of the paraffins than methane , which appears to be in some respects abnormally difficult to ignite .
Carbon monoxide was chosen because of the relative simplicity of its combustion .
The limits of inflammability of ethane are 3*1 and 10*7 per cent , in air , of carbon monoxide about 14 and 71 .
The latter have not been worked out closely , for it is in the intermediate mixtures that most difference is observed .
2 .
Continuous Current .
The results of the trials of different metals using .continuous current are given in fig. 1 for ethane .
The curves have the same relative positions in * F. Emich , Vienna , 'Akad .
Wiss .
Sitzber .
, ' vol. 106 ( Abth .
2b ) , p. 10 ( 1897 ) ; 'Naturw .
Rundsch .
, ' vol. 12 , p. 575 ( 1897 ) ; ' Monatshefte Chem. , ' vol. 18 , p. 6 ( 1897 ) ; vol. 19 , p. 299 ( 1898 ) ; vol. 21 , p. 1061 ( 1900 ) .
10 Prof. W. M. Thornton .
The Reaction between Gas and weak and rich mixtures , and are inclined to the horizontal in the manner previously found for continuous current ignition of the paraffins.* Ignition is most difficult with platinum poles , then with copper , nickel , iron , and aluminium in this order , f S 4 5 6 7 8 9 IO II 12 13 14 PER CENT OF GAS IN AIR Fig. 1- Least igniting currents in mixtures of ethane and air .
Continuous current , .
100 volts .
Break-sparks .
( 1914)Tte EleCtrlCal Ignition of G~ Mixtures , " ' Roy .
Soc. Proc. , ' A , vol. 90 , p. 272 .
+ See ^'^bs !
\lt ; PbiL MaS- ' ( 6)\gt ; voL P- 617 , 1905 , or J. J. Thomson , The Conduction of Electricity through Gases , ' 2nd edit .
, fig. 127 , p. 459 .
I amperes Pole in the Electrical Ignition of Gaseous Mixtures .
11 In carbon monoxide ( fig. 2 ) the curves are more of the horizontal type observed with condenser spark-ignition , * which is atomic in the sense that ignition is determined by difficulty of choice between atoms free to combine .
The order of difficulty in weak mixtures is platinum , copper , nickel , iron , and aluminium , the same as in ethane .
In the rich mixtures copper and platinum change places .
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 PERCENT OF GAS IN AIR .
Fig. 2.\#151 ; Least igniting currents in mixtures of carbon monoxide and air .
Continuous current , 100 volts .
Break-sparks .
3 .
Alternating Currents .
With ethane ( fig. 3 ) the usual rounded type of curve is found , though the lower cuives are flattened .
The order of difficulty is now copper , iron , nickel , aluminium , and platinum , and the same order is found in carbon monoxide ( fig. 4 ) .
The Ignition of Gases by Condenser Discharge Sparks , " vol. 91 , p. 17 ( 1914 ) .
1 Roy .
Soc. Proc. , ' A , 12 Prof. W. M. Thornton .
T Reaction betiveen Gas and 9 10 II 12 3 1 PERCENT OF GAS IN AIR .
Fig. 3 .
Least igniting currents in mixtures of ethane and air .
Alternating current , 200 volts , 40 ~ .
Break-sparks .
i AMPERE Pole in the Electrical Ignition of Gaseous Mixtures .
to 15 20 25 30 35 40 45 50 55 60 65 70 75 80 PERCENT OF GAS IN AIR Fig. 4 .
Least igniting currents in mixtures of carbon monoxide and air .
Alternating current , 200 volts , 40 ~ .
Break-sparks .
Comparing continuous and alternating values , the currents for platinum are much the same in both , but the others have changed greatly .
Taking the midway mixtures , that is , 7 per cent , in ethane , 42*5 per cent , in carbon monoxide , we have the following ( Table I ) , the currents being in amperes:\#151 ; Prof. W. M. Thornton .
The Reaction between Gas and Table I. Continuous current , 100 volts .
Alternating current , 200 volts , 40 ~ .
Metal .
Ethane , 7 per cent. Metal .
CO , 42 * 5 per cent. Metal .
Ethane , 7 per cent. Metal .
CO , 42 *5 per cent. Pt 1 -98 Pt 0*90 Cu 9*7 Cu 9 *7 Cu 1 *65 m 1*05 Fe 7*5 Fe 7*5 Mi 1-47 Cu 1 TO Ni 3*7 Ni 4*7 Fe 1 -33 Fe 0*85 A1 2-9 A1 3 2 A1 1T5 A1 0*82 Pt 2*5 Pt 2*6 0 With alternating pressures the magnitudes of the least igniting currents are little different in the two gases , and it is therefore necessary to qualify a previous conclusion* that " ignition by alternating current break-sparks appears to depend more upon the nature of the gas , that by continuous current on the nature of the poles " The reaction is too intimate for this to be true for all gases .
It applies to the paraffins when ignition is made with iron poles .
4 .
Jump-sparks .
This name describes the single disruptive discharge through a gas , from condensers or from the secondary of an induction coil the primary circuit of which is made or broken with the trembler locked .
The latter spark is sometimes called in chemical papers a " break-spark , " but this may be confused with the extra current arc formed at the point where any circuit is broken , which has been always called a break-spark .
In the present case the spark-gap was fixed in the explosion vessel at a length of 1^ mm. If the current broken is below a certain value no secondary spark passes , and in most inflammable gaseous mixtures even when a jump-spark is first obtained it does not cause ignition .
There is , for every mixture , pressure , and coil , a least primary current which , when broken , causes ignition by the single secondary spark .
The disruptive discharge in this case lasts longer than condenser discharge and is at atmospheric pressure a thin spark .
At lower pressures it spreads to a much larger volume .
Most work on the electric strength of gases has dealt with sparks of this nature , a full discussion of which , from the point of view of ionisation by collision , is given by Prof. J. S. Townsend , in 'Electricity in Gases/ Chapters 9 and 10 .
It is shown there that in general the sparking potential is a linear function of the product of gas pressure and sparking distance , which means , as shown by de la Rue and * The Ignition of Coal Gas and Methane by Momentary Electric Arcs/ ' 4 Trans. Inst. Min .
Eng./ Part I , vol. 44 , p. 173 .
i Pole in the Electrical Ignition of Gaseous Mixtures .
15 Muller , that the passage of the spark depends on the number of molecules of gas between the poles.* 3 Under this head may be also classed the exceedingly short sparks obtained at the discharge of a condenser , charged at a low pressure , from 50 to 250 volts , by bringing together a platinum wire and rod .
Just before metallic contact a spark passes , a few thousandths of a centimetre long , sometimes with large sparks welding the poles together .
Ignition by this very rapid discharge has features quite distinct from that by break-sparks .
The most definite of these is that at certain mixtures there is a sudden increase in difficulty of ignition as the percentage of combustible gas is raised .
A series of steps is formed , the lowest starting from the lower limit of inflammability , each having a flat tread , the last at the upper limit having an infinite rise .
In order to examine the electrical conditions of ignition by jump-sparks under conditions comparable with those for which the laws of sparking potential have been worked out , mixtures of constant proportion were used and the gas pressure varied .
In this case the heat of combustion q of the gas per unit volume is proportional to the pressure .
The relation between voltage Y , pressure p , and spark length s , is of the form Y = aps + b , where both a and b differ with the kind of gas.f Writing q = Cp , the voltage required to produce a spark of constant length which will liberate unit quantity of heat in the gas is Y/ q = A + B/ jp .
If this is the critical igniting voltage there should be a relation of the form ( y \#151 ; A)p = a constant , where y is the electrical ordinate .
This is sometimes found with jump-sparks , but more usually with break-sparks , when the pressure is varied .
On account of the large volume of gas required for these observations it was only possible to work with coal-gas used for lighting .
The mixture to give perfect combustion was 11 per cent , of gas in air .
This was made up in a gasholder , and the same used throughout .
The results are given in figs. 5 to 7 , and show that the material of the poles has a great influence on ignition both in magnitude and type of variation , even with jump-sparks .
To distinguish between the two kinds used single induction coil discharge has been called a jump-spark J , a condenser discharge spark being called a condenser spark K. In the latter case the capacity was varied , the voltage maintained constant .
The same number of molecules was , therefore , acted upon in every trial .
In the former the secondary voltage is proportional to the primary current , the energy of the secondary spark to the square of the current .
There is evidence that it is the voltage and not * W. de la Rue and H. W. Muller , 'Phil .
Trans. , ' vol. 171 , p. 109 ( 1880 ) .
+ See Townsend , loc. cit. , p. 357 .
16 Prof. W. M. Thornton .
The Reaction between Gas and energy of the spark that determines ignition in certain cases .
The ordinates of the J or long jump-spark curves are the primary currents which when broken cause ignition by the single secondary spark .
Those of the K or condenser discharge curves are the capacities which when charged to 100 volts just cause ignition by discharge between points a few thousandths of a centimetre apart .
amperes microfarads ATMOSPHERES NICKEL ATMOSPHERES r _ I RON ig .
, ) .
Ignition by disruptive discharge .
Gas pressure varied .
Mixture , 11 per cent. of illuminating gas in air .
i * Pole in the Electrical Ignition of Gaseous Mixtures .
ATMOSPHERES ALUMINIUM VOL. XCII.\#151 ; A , C 18 Prof. W. M. Thornton .
The Reaction between Gas and * MICROFARADS 1-4 O atmospheres COPPER 8 10 12 14 O CARBON Fig. 7 .
ATM05PHERES 5 .
Ignition by Jump-sparks .
The material of the poles has here as much influence on ignition as the nature of the gas .
The following Table gives the relative order of difficulty of ignition , with the break-spark results for comparison .
Pole in the Electrical Ignition of Gaseous Mixtures .
Table II .
Coal-gas .
Ethane .
Carbon monoxide .
Jump .
Condenser .
C.C. break .
A.C. break .
C.C. break .
A.C. break .
Ni Cu Pt Cu Pt Cu Fe Fe Cu Fe Cu Fe Al Ni Ni Ni Ni Ni Pt Al Fe Al Fe ah Cu Pt Al Pt Al pt| Condenser and alternating break-sparks have the same order of metals .
In all three cases ignition is easiest with platinum poles .
With jump-sparks copper is lowest , and the inversion of copper in the first two columns and of platinum in the next pairs shows that the mechanism is profoundly different .
6 .
Causes of Difference .
The physical properties of the poles which may influence ignition are density , specific heat , melting point , thermal conductivity , and , for jump-sparks , tenacity .
The following values are taken for these .
Table III .
Metal .
Density , A. ' Specific heat , \lt ; r. Melting point , M. Thermal conductivity , d. Tenacity .
xl0'- M0 Aluminium 2*61 0*212 ' C. 625 0*344 1 *75 x 109 2*5 Copper 8*9 0*093 1084 0*918 2*9 8*3 Iron 7*8 0*110 1515 0*115 4*6 4*9 Nickel 8*6 0T09 1435 0*142 5*3 4*6 Platinum ... 21 *5 0*032 1900 0*166 3*3 2 *1 The order of difficulty of ignition by continuous-current break-sparks is that of density of the poles , and would appear to arise from difficulty of accelerating bodies of atomic dimensions set free at the cathode .
Continuous-current ignition has been previously observed to have features of ionisation , * by a-rays , for example .
The difference between continuous and alternating break-spark ignition depends mainly upon time , that is , upon the relative rates of reaching a state at which there is discharge from the poles .
It is therefore directly proportional to density and specific heat , and inversely to thermal conductivity .
If , however , the surface liquefies at a relatively low temperature * ' The Electrical Ignition of Gaseous Mixtures , ' loc. cit. , p. 279 .
20 Prof. W. M. Thornton .
The Reaction between Gas and there is both an absorption of heat from the spark and a lower temperature available for corpuscular discharge .
The lower the melting point the greater the difficulty of ignition by radiation from the poles .
We take , then , the ratio Aa/ M0 as a measure of the difficulty of ignition when it is affected most by rate of heating .
The order of magnitude of the figures in the last column is copper , iron , nickel , aluminium , and platinum , which is that of the metals in condenser and alternating break-spark ignition .
This kind of ignition can thus be said to have a thermal origin , * in the sense that it is proportional to the rate of heating of the poles .
The rate and its influence are the same in ethane and carbon monoxide at the mid-point chosen , but not elsewhere .
No direct comparison can be made with condenser ignition , but the relative positions of the curves are the same .
Disruptive discharge , apart from ignition , has been shown to depend upon the rates of ionisation by collision in the gas .
When , however , there is chemical combination , the spark may be accelerated or retarded in its igniting power ; there is evidence of both .
Aluminium and carbon ignitions are very sensitive to change in collision frequency in the gas , and no two materials give the same curves .
The order of the metals in Column 1 of Table II is that at half an atmosphere pressure , the currents at atmospheric pressure being too small for discrimination .
The magnitudes are as follows:\#151 ; Table IV .
Jump-sparks .
Condenser discharge .
Metal .
Primary current .
Metal .
Capacity .
amperes .
microfarads .
Ni 11 Cu 6 Fe 7 Fe 6 Al 4*5 Ni 5 Pt 2-4 Al 2 Cu 1*5 Pt 0*9 Carbon 1*9 Carbon 0*45 The spectrum of spark discharge shows that atoms are torn from the poles and carried over in the disruptive rush .
Difficulty of starting a spark might therefore be expected to depend upon the tenacity of the material.f * * " The Least Energy required to Start a Gaseous Explosion , " ' Phil. Mag. , ' vol. 28 , p. 737 ( November , 1914 ) .
t See also J. J. Thomson , loc. cit. , p. 457 .
) Pole in the Electrical Ignition of Gaseous Mixtures .
21 With the exception of aluminium and carbon , the order in the first column is that of tenacity .
Both of these substances offer more difficulty than might be expected from their sensitiveness to gas conditions .
Carbon atoms in gases are difficult to ionise , and the pioneering ionisation before the spark passes may be retarded by this .
The influence of the metal and state of the poles on disruptive discharge in different gases requires fuller examination .
7 .
Carbon Poles .
In many observations it has been found that when the limit was approached ignition took place not at the first spark but after a few had passed , with a wait of several seconds between each .
To account for this the surface must be changed .
Under the microscope it appears slightly roughened , and when working with carbon compounds and most metal poles a faint deposit of carbon was made around the point of contact , which after many sparks had been passed formed a regular ring .
It became necessary , therefore , to examine whether carbon helps ignition .
For this purpose graphite B drawing pencils were used , one sharpened to a point , the other rounded , with results given in fig. 7 .
Ignition by condenser spark discharge with carbon poles is about twice as easy as with platinum .
It has a very well-defined curve , a curious medley of steps and cusps .
That by jump-sparks has an equally singular shape not unlike that of aluminium .
In both carbon cases the first increase of difficulty begins about 0*8 atmosphere .
At 0'7 there is a sharp rise which with jump-sparks ends in a step with a further rise at 0*62 , reaching a maximum at 0*55 .
It then falls to an ill-defined minimum which appears to be an attempt to renew the stepped formation , and proceeds to a lower limit at about 0T8 atmosphere .
In the condenser curve , after the rise at 0*7 there is a fall to a minimum at 0*55 , from which the curve goes by steps to a lower limit a little above 0*2 atmosphere .
The magnitudes of the latter curve are about midway between copper , the lowest , and platinum .
The steps at 0'6 and 0*7 are found in methane but not in hydrogen .
There is a fall at 0*55 in hydrogen and 0*5 in methane .
Steps are found in methane more than in any other gas ; the rise at 0*35 is peculiar to carbon monoxide .
As might be expected , all the constituent gases take part in the reaction , carbon monoxide carrying it on below pressures at which , in both methane and hydrogen , ignition fails .
The condenser curve follows hydrogen in type , the jump-spark curve methane .
Thus when , as in condenser discharge , we see from SS 5 and 6 that the order of ignition depends on the rate of heating of the poles , hydrogen carries on the action of the spark ; when , as in jump-sparks , ionisation by collision is the decisive feature , methane is Electrical Ignition of Gaseous Mixtures .
first affected .
These are pre-combustion conditions , and the above conclusions do not bear upon the question of preferential combustion m an established inflammation .
The influence of carbon as an accelerator may have bearing on ignition in internal combustion engines .
So long as the deposit does not short-circuit the spark-gap it would appear to be beneficial .
The fact that with condenser discharge carbon gives easiest ignition of all the materials tried may help to explain why carbonisation of the sparking plug does not hinder and may even help the efficient working of the Lodge condenser system of gas-engine ignition .
8 .
Summary .
It is shown that the metal of the sparking points has a great influence upon ignition , at pressures up to an atmosphere , whether the sparks are disruptive or formed by separation of the poles .
The order of difficulty of ignition by continuous-current break-sparks is that of the density of the poles .
That by alternating or condenser discharge sparks is proportional to the rate of heating of the points of contact .
That by jump or long disruptive discharge sparks varies with the tenacity of the material of the poles .
With carbon poles condenser discharge in coal gas has most effect upon the hydrogen , longer disruptive discharge upon the methane .

