118 ELECTRICAL THEORY OF ADBORPTTON The writer considers the double layer as consisting of a swface of rigidly fixed atoms under continuous bombardment of positively and negatively charged ions, any particular point on the rigid surface becoming in turn negative, neutral and positive, these conditions arisdg in any order. The observed contact difference is the average effect of these conditions. Where several kinds of atoms are present in the solution the average number of any one of them at the surface will depend on their concentbration, valency and mobility. The variation of contact Werence from negative to neutral and positive was observed with cotton and aluminium sulphate near the neutral point. These variations occurred during the same experiment, the readings being direct measurements of E.1I.F.s developed by filtration under pressure.This point would be covered by putting n2 = 1 and = 2 or 3 in Mukherjee’s equation No. 13.118 ELECTRICAL THEORY OF ADBORPTTON The writer considers the double layer as consisting of a swface of rigidly fixed atoms under continuous bombardment of positively and negatively charged ions, any particular point on the rigid surface becoming in turn negative, neutral and positive, these conditions arisdg in any order. The observed contact difference is the average effect of these conditions. Where several kinds of atoms are present in the solution the average number of any one of them at the surface will depend on their concentbration, valency and mobility. The variation of contact Werence from negative to neutral and positive was observed with cotton and aluminium sulphate near the neutral point.These variations occurred during the same experiment, the readings being direct measurements of E.1I.F.s developed by filtration under pressure. This point would be covered by putting n2 = 1 and = 2 or 3 in Mukherjee’s equation No. 13.118 ELECTRICAL THEORY OF ADBORPTTON The writer considers the double layer as consisting of a swface of rigidly fixed atoms under continuous bombardment of positively and negatively charged ions, any particular point on the rigid surface becoming in turn negative, neutral and positive, these conditions arisdg in any order. The observed contact difference is the average effect of these conditions. Where several kinds of atoms are present in the solution the average number of any one of them at the surface will depend on their concentbration, valency and mobility.The variation of contact Werence from negative to neutral and positive was observed with cotton and aluminium sulphate near the neutral point. These variations occurred during the same experiment, the readings being direct measurements of E.1I.F.s developed by filtration under pressure. This point would be covered by putting n2 = 1 and = 2 or 3 in Mukherjee’s equation No. 13.118 ELECTRICAL THEORY OF ADBORPTTON The writer considers the double layer as consisting of a swface of rigidly fixed atoms under continuous bombardment of positively and negatively charged ions, any particular point on the rigid surface becoming in turn negative, neutral and positive, these conditions arisdg in any order.The observed contact difference is the average effect of these conditions. Where several kinds of atoms are present in the solution the average number of any one of them at the surface will depend on their concentbration, valency and mobility. The variation of contact Werence from negative to neutral and positive was observed with cotton and aluminium sulphate near the neutral point. These variations occurred during the same experiment, the readings being direct measurements of E.1I.F.s developed by filtration under pressure. This point would be covered by putting n2 = 1 and = 2 or 3 in Mukherjee’s equation No. 13.118 ELECTRICAL THEORY OF ADBORPTTON The writer considers the double layer as consisting of a swface of rigidly fixed atoms under continuous bombardment of positively and negatively charged ions, any particular point on the rigid surface becoming in turn negative, neutral and positive, these conditions arisdg in any order.The observed contact difference is the average effect of these conditions. Where several kinds of atoms are present in the solution the average number of any one of them at the surface will depend on their concentbration, valency and mobility. The variation of contact Werence from negative to neutral and positive was observed with cotton and aluminium sulphate near the neutral point. These variations occurred during the same experiment, the readings being direct measurements of E.1I.F.s developed by filtration under pressure.This point would be covered by putting n2 = 1 and = 2 or 3 in Mukherjee’s equation No. 13. IONISATION IN FLAMES OF VARIOUS ORGANIC SUBSTANCES. (InvestQations made at fh Air Minisfry Laboratory, Imperial College of Science and Technology, and publishd by permission of the Director of Scientiyfc Research, Air Miaistry.) BY J. A. J. BENNETT, CARNEGIE RESEARCH FELLOW. Received 31st March, 1927. The study of the phenomenon of detonation in internal combustion engines has recently given rise to many subsidiary problems of fundamental importance. I t has been suggested by many writers that ionisation is one of the chief causes of detonation, though no relation has yet been discovered between the former and the latter. Up till now attention has been devoted mainly to the study of ionisation in gaseous explosions.The object of the present experiments was to determine the correlation, if any, between the degree of ionisation of various substances in flames aqd the amount of detonation in an engine cylinder charged with fuels containing these substances. The range of investigation includes an experimental study of ionisation in flames of hexane, ether, alcohol, coal gas, etc., and the effect of adding iron carbonyl, amyl nitrite, etc.,-substances which alter the highest useful compression ratio (H.U.C.R.) of a fuel. All gases conduct electricity, but very dclicate apparatus is required to show that some gases in their natural state possess this property even to a small extent. Unless ions have been produced by raising the temperature, by exposure to active rays, or by some other method, all gases are relatively good insulators. One of the oldest methods of obtaining gases in an ionised state is by producing flame.Giese,2 by studying the electrical properties of gases coming from flames, discovered the theory which is now used to explain gaseous conductivity. The current passing between two electrodes at different potentials immersed in gaseous media is a t present considered to be due to the movement of the electrons towards the anode and the positive ions towards the cathode. The mobility of the former is of a higher order than the latter, except possibly when the electrons become attached to neutral atoms. At higher temperatures the neutral atoms (M) of gaseous elements dissociate into ions (&) and electrons manner :- + - M Z M + E - U, where U is the energy absorbed in the process.The H.U.C.R. is the highest compression ratio which it is worth while to employ with a given fuel. If the compression ratio is raised above this limit excessive detonation leading to pre-ignition and loss of power is obtained. For full information on this subject reference is made to the Aeronautical Research Committee publication, R. & M. No. 1013, on Ddpes and Detonation by Callendar, King, and Sims. Wied. Ann., 17, 519, 1882. 307308 IONISATION I N FLAMES OF ORGANIC SUBSTANCES - Saha,s assuming that the electron is a monatomic gas of atomic weight 1/1836 and the energy U to be the product of the ionisation potential and the electronic charge, has recently developed a theory of thermal ionisation of gaseous elements at high temperatures which has received confirmation by experiments with salt vapours in flames at particular temperatures by Noyes and Wilson,4 B a r n e ~ , ~ and for variation in temperature by the writer.d Garner and Saunders,' by using an Einthoven galvanometer for measuring the percentage of ionisation occurring in explosions of hydrogen and oxygen, found that Saha's theory holds approximately for gaseous explosions and therefore that the ionisation is mainly thermal as in flames.The similarity between the origin of ionisation in flames and explosions suggests that the present experiments may throw light or shadow on the views ex- pressed by other writers on the relation between ionization and detonation, concerning which there is much diversity of opinion.Measurement of Ionisation. Since the mobility of the electrons is so great compared with that of the positive ions, the current passing between two electrodes immersed in the excited flame medium is almost solely due to the electricity carried by the electrons. Accordingly, the electrical conductivity is proportional to the number of electrons present, and since the number of electrons is equal to the number of positive ions, the conductivity may be taken as a measure of the ionisation. The specific conductivity at a point in the region between the electrodes depends upon the potential gradient at the point as well as the current per unit area of one of the electrodes, guarded to ensure that the lines of force between the electrodes go straight across.The present experiments were not intended to measure the absolute value of the ionisa- tion, and no measurement of the potential gradient at various points between the electrodes was made ; it was therefore unnecessary to have one of the electrodes guarded. Precautions were taken to keep the electrodes at a constant distance apart, to keep their position constant in the flame and to keep the temperature of the flame constant ; the area of the electrodes and the difference of potential between the latter remained unchanged. The current under these conditions is practically a measure of the degree of ionisation. Descrz3tion of Apparaius.--Two platinum electrodes each 3 centimetre square were placed just over the inner cone of an ordinary bunsen flame.They were made of thick foil and welded to wires 3 millimetre in thickness. The wire connected to the cathode was made of platinum, but the anode was attached to the junction of two wires, one made of platinum, the other of platinum-rhodium. The temperature of the anode could therefore easily be kept constant, the wires leading to the anode being connected to a millivoltmeter supplied by the Cambridge Scientific Instrument Company. The horizontal distance between the electrodes was a little over a centimetre, and this distance was kept constant by passing the wires through quartz tubes of small bore fused together lengthwise. A battery of small cells supplied a constant potential difference of 30 volts between the electrodes and the current passing through the flame medium was measured by a Kelvin galvanometer, the sensitiveness of which :;Phil.Mag., 40, 478, 1920. Physical Rev., February, 1924. 7 Trans. Far. SOC., October, 1926. 4 Astroph. Jouv., 57, 20, 1923. 8 Phil. Mag., Vol. 3, January, 1927.J. A. J. BENNETT 309 could be adjusted by a control steel magnet. The deflection on the galvano- meter scale indicated a fairly steady current when the flame was shielded by a chimney. The effect of various substances on the conductivity of the flame was studied, the method of introducing each substance depending on its volatility. The more volatile substances were mixed with the air before entering the flame ; the less volatile substances were burned at the end of a quartz tube introduced at the mouth of the burner.Care must be taken to have the connecting wires well insulated and the electrodes free from deposit. The FIarne Constituents. Since the ionisation depends on (a) the nature, and (6) the concentra- tion of the flame constituents, (c) the temperature of the flame, (d) the distance of the electrodes from one another, and (e) the amount of deposit on the electrodes, it was necessary to eliminate all these variables except (a), as far as possible, if the true effect of each substance was to be deter- mined. By examination of the table of results it will be noticed that some substances decrease whilst others increase the conductivity of the flame. The action is not a catalytic one, but the variation of the conductivity of the flame is due to the variation of the concentration of electrons in the flame. ‘The presence of substances of low ionisation potential-which is the potential required to give an electron enough energy to ionise a neutral atom-increases the number of electrons present in the flame and there- fore increases the current ; substances of high ionisation potential decrease the electron concentration and therefore decrease the current.The following results with the Bunsen flame and the addition of various substances are typical of results obtained with other flames, as will be pointed out later. Substaxe. Nitrogen Compounds Aniline Methyl-aniline Di-methyl-aniline Benzyl-ethyl-aniline Xylidine Nitrogen-peroxide Ammonia Phenyl hydrazine Amyl-nitrite Picric acid Carbonyls Iron-carbonyl lU Nickel-carbonyl Organic oxides, etc.Benzoyl-peroxide Hydrogen-peroxide Ether Effect on Effect on Ionisation. Detonation.8 Slightly decreases Delays I , I 11 , I ,, slightly ,, increases ,, , t Decreases I , 9 Markedly induces I , Slightly delays Little effect Slightly induces Markedly decreases Greatly induces Greatly increases Induces Greatly increases Greatly delays Greatly increases Greatly induces , I ,, Little effect Cooling effect Induces I , Halogens and Halogen Iodine Greatly increases Delays Compounds Bromine , t ,, Slightly induces Ethylene dibromine I , ,, ,, delays Dichlorethylene , ,, Slightly induces Acetylene tetrabromide Increases Chloroform Greatly iricreases Lit?le effec;’ 8 Information obtained from the report on Dopes and Detonation by Callendar, King, 9 Dark flame between electrodes only.’0 Reddish-brown depxit on cathode only. and Simsl and from experiments made by the staff of the Air Ministry Laboratory.310 IONISATION IN FLAMES OF ORGANIC SUBSTANCES Substance. Effect on Ionisation. Effecct on Detonations- Aldehydes Valeraldehyde Slightly decreases Slightly delays Propylaldehyde Decreases Little effect Paraldehyde Slightly decreases Slightly delays Oenanthol I , Little effect Acetaldehyde I, ,, Slightly delays Miscellaneous Lead xylyl Greatly increases Markedly delays. Ethyl fluid l1 I , ,, Greatly delays Carbon-disulphide Decreases Delays Tin oleate Greatly increases Little effect Sulphonal decreases Delays Water Cooiing effect I , Ethyl alcohol ,I I , I Experiments with Different Flames.Various experiments were made to confirm the results tabulated above by using flames other than the ordinary Bunsen. The influence of the various substances on the ionisation in a hexane flame was examined and thereafter the ionisation in flames of the substances themselves was determined. The method of producing the hexane flame consisted in burning a mixture of air and hexane vapour at the ends of four small quartz tubes which were sealed with sealing-wax to a rubber stopper. A long flame was thus produced and the electrodes were placed in the flame just over the inner cones. The rubber stopper was fitted to a vessel containing cotton wool soaked in hexane through which a stream of air was passed. The air, which was supplied under slight pressure from a ten-gallon drum, was dried and freed from carbondioxide by passing through a tower of granular calcium chloride, then through soda-lime and then by bubbling through concentrated sulphuric acid.The flow of air was adjusted by means of a screw clip. The cathode became coated with deposit very easily and since the current varies very greatly with the amount of deposit on the cathode it was necessary to make each experiment as quickly as possible. The hexane was doped with each substance in turn and the results obtained were exactly similar to those given in the above table. To complete the investigation, the hexane vessel was replaced by a weighing bottle containing the substance in liquid form to be exlmined. The weighing bottle was surrounded by a beaker of water placed on an iron plate which could be heated at will by a Bunsen flame.In this way the less volatile substances were vaporised. The air in some cases. bubbled through the liquid and in others merely passed over the surface, the extent of the contact of the air with the liquid being varied to give the mixture strength desired. The amount of liquid used during each experi- ment was determined by weighing the bottle and contents before and after. In this way the mixture strength used in each case was kept constant. The temperature, of course, was kept constant as before by using the platinum- rhodium thermocouple. Those substances in the foregoing table which could be used in this way gave results which showed that flames of substances which increase or decrease the ionisation of a Bunsen flame possess electrical conductivity to a greater or less degree respectively.11" Ethyl fluid" consists of a mixture of lead-tetra-ethyl and ethylene dibromide in the proportion of 3 grams of theformer to 2 grams of the latter.J. A. J. BENNETT 31’ I n addition flames of u B.P.” petrol, benzene, pentane, acetone, pseudo- cumol, turpentine, and various mixtures were used; a mixture of two substances was found to give a flame the conductivity of which lies between the separate conductivities of the two substances, e . 6 , L L B.P.” petrol and carbondisulphide. The current obtained with (‘B.P. ” petrol alone was 0 - 7 microamp., with CS, alone 0.03 microamp., with a mixture containing 2 5 per cent.CS,, and 75 per cent. petrol 0.1 microamp. Conclusion. It is evident from these results that although in many cases knock- inducers increase and anti-knocks decrease the ionisation of flames, this is by no means generally so. Consequently, Wendt and Grimms’ theory IB that the advance of free electrons before the flame front ionises the unburnt gas, thus increasing the rate of flame propagation and causing detonation, does not seem at all possible in the light of the present results. If this theory were correct an increase in the ionisation of the flame would increase the rate of flame propagation and induce detonation. But the carbonyls. do not behave in this manner. Neither does the view of Charch, Mack, and Boord,lS that there is a direct relationship between the intensity of detonation and ionisation receive confirmation. Certainly for particular substances ionisation will increase with increase of knocking since there is a rise in temperature, but the degree of ionisation is not generally a criterion from which to estimate the intensity of detonation, for the carbonyls become exceedingly ionised in flames and the nitrites remain relatively in the neutral state whereas the former tend to delay, the latter to induce, detonation. We must therefore conclude that, although ionisation accompanies detonation as it does all flame phenomena, there is no simple relationship between them. Ionisation does not appear to be either a cause or effect of detonation, but mainly a temperature effect. The aboce work forms part of an investigation on the cause and prevention of detonation in the internal combustion engine made at the Air Ministry Laboratory, Imperial College of Science and Technology, London. I t was prosecuted under the supervision of Professor H. L. Callendar, C.B.E., LL.D., F.R.S., and with the help of Dr. E. W. J. Mardles, F.I.C., to both of whom the author is deeply indebted. lay. Ind. Eng. Chem., 1924, 16, Sgo. 13 Ibid., April, 1926.