December 8, 1910] 



NATURE 



187 



combustion chamber tends to increase efficiency. Some 

 experiments were also quoted in which it was found that 

 lining an explosion vessel with bright tinfoil perceptibly 

 retarded the cooling of the products. More recently an 

 e.xplosion vessel has been plated with silver on the inner 

 surface, and the results have been compared after exploding 

 identical mixtures, first when the lining was highly 

 polished, and secondly when it was blackened over with 

 lamp-black. It was found that by highly polishing the 

 interior of the vessel the maximum pressure reached could 

 be inci eased 3 per cent., and the subsequent rate of cool- 

 ing during its earlier stages reduced by about one-third. 

 These experiments leave no doubt of the reality and of the 

 practical importance of radiation as a factor in determining 

 i the heat-loss in the gas engine.^ 



Reference may also be made to the part played by radia- 

 tion in determining the heat-flow in a boiler. Attention 

 was directed to this by Dalby in a recent report to the 

 Institution of Mechanical Engineers.^ The circumstances 

 in this case are widely different from those usually obtain- 

 j in the gas engine, but the instance serves to emphasise 

 importance to the engineer of the questions which will 

 discussed in this report. 



Amount of the Radiation fioni Flame. 



I R. von Helmholtz appears to have been the first to 



ntrempt the accurate measurement of the radiation emitted 



a flame. He found that a " solid " flame 6 mm. 



;neter, burning coal-gas, radiated about 5 per cent, of 



;i. total heat of combustion. A carbon monoxide flame 



radiated about 8 per cent., and a hydrogen flame about 



i--r cent. On account of the smallness of the flame his 



riments have not much application to the problem of 



gas engine. The size of the flame affects the matter 



rwo ways. In the first place, a large flame radiates 



■■>• per unit of area than a small one, because a flame 



■ J a great extent transparent even to its own radiation, 



hat radiation is received, not only from molecules at 



surface of the flame, but also from those at a depth 



lin it. This matter will be further dealt with in 



rher section of the report. The second point is that 



cooling of the gas is slower in a large flame than in 



-Tiall one. The radiation originates in the vibration of 



CO. and steam molecules, and the life of one of these 



!-cules as a radiating body extends from the moment 



;rs formation to the time when its vibrational energv 



been destroyed by radiation and by collision with colder 



•cules. such as those of the air surrounding the flame. 



smaller the flame the more rapid will be the extinc- 



of the vibrations, and the less, therefore, the total 



unt of radiation per molecule. The products of 



lesion in a closed vessel or in a gas engine differ 



-iderably in this respect from any open flame, how- 



• large, which it is possible to produce, for they are 



subject to cooling by mixture with the outside air. 



Mnreover, the density of the gas is very much greater. 



f'allendar has repeated some of Helmholtz 's experiments 



a larger scale, and has found that the radiation in a 



-luminous coal-gas flame 30 mm. in diameter may 



unt 10 15 per cent, of the whole heat of combustion. 



;her reference will be made to Callendar's work under 



heading of "transparency." 



lopkinson has recently made measurements of the 



ation emitted in the course of an explosion in a closed 



-el and subsequent cooling. .A bolometer made of 



kened platinum strip was placed outside a window 



fluorite in the walls of the explosion vessel. The 



I trical resistance of this bolometer was recorded by 



ins of a reflecting galvanometer throwing a spot of 



t on a revolving drum, and an optical indicator traced 



ultanpously a record of the pressure on the same drum. 



found that the total heat radiated during an explosion 



t 15 per cent, mixture of coal-gas and air, and the 



-equent cooling, amounted to more than 22 per cent. 



the whole heat of combustion. The radiation which 



! been received at the moment of maximum pressure 



lunted. to 3 per cent., and it continued, though at a 



inishing rate, for a long period. Radiation was still 



' Hopkin^on, Proc. Rry. Soc, A., vol. Ixxxiv. (1910), p. 155. 

 - Proc. Inst. Mech. Eng , October (1909). 



XO. 2145, VOL. 85] 



perceptible half a second after maximum pressure, when 

 the gas-temperature had. fallen to 1000° C 



Nature and Origin of the Radiation from Flames. 



In the gas-engine cylinder and in explosion experiments 

 we are usually concerned with flames in which there is 

 some excess of air. A mixture of similar composition 

 burnt at atmospheric pressure would give an almost non- 

 luminous flame ; in the gas engine there is more luminosity 

 on account of the greater density. There is, however, no 

 reason to suppose that the radiation in the gas-engine 

 cylinder differs materially as regards its quality or origin 

 from that emitted by an open flame. 



A very complete analysis of the radiation from different 

 kinds of flame was made by Julius, and his experiments 

 leave no doubt that the radiation is almost wholly due to 

 the CO, and steam molecules. He examined the spec- 

 trum of the flame by means of a rock-salt prism, and he 

 found that in all flames producing both COj and steam 

 most of the radiation was concentrated into two bands, 

 the wave-lengths of which are, respectively, 44 /i and 

 2-8 fi.. In a pure hydrogen flame the 44 band disappears 

 completely, but the other remains ; and in the pure CO 

 flame the 2-8 band disapj>ears, the other remaining. 

 These results are independent of the nature of the com- 

 bustible gas, the spectrum depending solely on the pro- 

 ducts of combustion." 



A confirmation of the statement that the radiation from 

 these flames originates in the CO, and H,0 molecules 

 only was furnished in the course of the work by R. von 

 Helmholtz, to which reference has been made above. He 

 measured the amount of radiation per litre of gas con- 

 sumed, emitted by flames of given size burning, re- 

 spectively, hydrogen, carbon monoxide, and certain com- 

 pound gases, such as methane, giving both CO, and 

 steam. The supply of air was adjusted in each case so 

 that the flame was just non-luminous. His results are 

 best given in his own words, but it should be stated that 

 he worked with a small flame about 6 mm. diameter 

 and measured the radiation with a bolometer, taking the 

 steady change of its resistance as a measure of the amount 

 of radiation falling upon it : — 



" According to the experiments of Julius described in 

 the first chapter, the quality of the radiation of flames 

 depends only on the nature of the burnt and not on that 

 of the burning gases. It is relevant to inquire whether 

 the quantity of radiation is also dependent on the mass 

 of the products of combustion. I have calculated in the 

 second and third columns below how many litres of 

 H,0 and CO,, respectively, arise theoretically from each 

 litre of combustible gas. I then assume that for every 

 litre of water produced as much radiation is sent out as 

 corresponds to the radiating power of a hydrogen flame — 

 for this gas yields one litre of H,0 per litre of combustile 

 — and that in a corresponding way the radiation from one 

 litre of carbonic acid would be determined by the radiating 

 power of the carbonic oxide flame, and I can then calculate 

 the radiation from the non-luminous flames of methane, 

 ethylene, and coal-gas. 



Litres F 



H20 



Hydrogen i 



Carbon monoxide .. o 



Mar.sh gas 2 



Ethylene 2 



Coal gas... .« ... I'2 



" The correspondence between the calculated numbers 

 with the radiation from a flame which has just been 

 rendered non-luminous surprised me the more since the 

 latter is conditioned, in some measure, by the volume of 

 air mixed with the gas, and this is very different for the 

 three non-luminous flames. On this account it cannot be 

 asserted that this agreement is not accidental. Moreover, 

 the number of observations is much too small. Neverthe- 

 less, the experiment seems worthy of record and will be 

 followed up further." 



• Proc. Key. Soc., A., vol. l.vxxiv. (igio), p. 155. 



2 " Die Licht- und Warmestrahlung verbrannter Case," Dr. W. H. 

 Julius. (Berlin, i8go.) 



