298 PHENOMENA, ATOMS, AND MOLECULES 



HEAT CONDUCTION IN GASES AT LOW PRESSURES 



Kundt and Warburg in 1875 predicted from the kinetic theory that 

 there should be a discontinuity in the temperature close to the surface of 

 a solid body which is dissipating heat by conduction into a gas at low 

 pressure. Smoluchowski, in 1898, observed and studied this phenomenon 

 and developed the theory of it. He found that in some gases, particularly 

 hydrogen, the amount of heat given up to the gas by the solid was only a 

 fraction of that which should be delivered if each molecule striking the 

 surface reached thermal equilibrium with it before leaving. 



Knudsen ^ gave the name "accommodation coefficient" to the ratio 

 of the heat actually carried away by the gas, to that which would be 

 carried if thermal equilibrium were reached. Kudsen measured the accom- 

 modation coefficient (a) for several gases and several diflferent kinds of 

 surface. The lowest value found was with hydrogen in contact with a 

 polished platinum or glass surface, in which case Knudsen obtained 

 a = 0.26. With heavier gases, such as nitrogen and carbon dioxide, he 

 found a = 0.87. 



Soddy and Berry ^ in a similar study of the heat conductivity of gases 

 at low pressures, found that a for hydrogen in contact with platinum 

 varied from 0.25 at — 100° C. to 0.15 at -|-200°. For argon they found an 

 average value a = 0.85, while for helium they found 0.49 at —100° C. and 

 0.37 at -]-i50° C. The accommodation coefficients obtained with heavier 

 gases were always close to unity. 



The writer "^ has determined the accommodation coefficient of hydrogen 

 in contact with tungsten at 1500° K. and obtained a = 0.19, while nitrogen 

 under the same conditions gave a = 0.60. 



The evidence thus far available indicates that the accommodation co- 

 efficients of the ordinary gases range from 0.19 up to unity. Only in the 

 cases of hydrogen and helium have accommodation coefficients less than 

 0.8 been found at room temperature. It is interesting to note that all 

 observers find that the coefficient for a given gas is independent of the 

 nature of the solid, so long as it has a polished surface. This is an indica- 

 tion, for which other evidence will be given below, that the surfaces under 

 the conditions of the experiments are covered with adsorbed layers of gases. 



Another conclusion from the experiments is that the accommodation 

 coefficient varies only slightly with the temperature, there being a tendency 

 for it to decrease at higher temperatures. Thus for hydrogen the coefficient 

 seems to decrease from about 0.35 at —190° to 0.19 at 1500° K. With 



' Ann. Phys., 34, 593, 191 1. 



^ Proc. Roy. Soc, 84, 576, 1911. 



''Jour. Amer. Chem. Soc, ,?/, 425, 1915. 



