HYDROGEN INTO ATOMS 201 



This agrees well with the diffusion coefficient calculated from the ex- 

 periments (Equation 83). We have previously remarked that the observed 

 rate of diffusion of hydrogen atoms through molecular hydrogen was 

 found to be 4.2 times greater than that calculated. Evidently hydrogen 

 atoms behave abnormally when diffusing through molecular hydrogen, but 

 not when diffusing through nitrogen. This is probably caused by the 

 hydrogen atoms going, we might say, through the hydrogen molecules by 

 combininsf with one atom in a molecule with which it collides and thus 

 liberating the other. This effect is probably entirely analogous to the 

 abnormal mobility of the H and OH ions in aqueous solutions. 



SUMMARY 



1. Previous calculations of the degree of dissociation of hydrogen, 

 from experiments on the heat losses from tungsten wires in hydrogen, 

 conflicted rather seriously with Bjerrum's determinations of the specific 

 heat of hydrogen by the explosion method. The weak point in the method 

 of calculation was a rather arbitrary assumption regarding the diffusion 

 coefficient of atomic hydrogen through ordinary hydrogen. 



2. This apparent discrepancy suggested that the actual amount of 

 dissociation might be somewhat less than that previously calculated, but 

 did not indicate that the abnormal heat conductivity of hydrogen at high 

 temperatures was not due to dissociation. 



3. The qualitative evidence of such a dissociation seems conclusive. 

 Thus, not only does hydrogen show an entirely abnormal increase in heat 

 conductivity, but this heat conductivity actually increases as the pressure is 

 lowered. 



A large amount of chemical evidence has also been accumulated. If 

 hydrogen at low pressures comes into contact with metallic wires at tem- 

 peratures above 1300° K., part of the hydrogen is converted into an active 

 modification which remains adsorbed on glass surfaces for long periods of 

 time. This active hydrogen can react at room temperature with oxygen or 

 with many reducible substances, or can dissolve in metals such as platinum. 

 There are good reasons for thinking that this active hydrogen is hydrogen 

 in the atomic condition. 



4. A series of experiments were made to measure, more accurately than 

 hitherto, the heat losses from a tungsten wire at temperatures from 800° 

 to 3500° K. and at pressures of hydrogen ranging from 0.0 1 mm. up to 

 atmospheric pressure. Similar experiments were made with nitrogen. 



5. At temperatures below that at which dissociation occurs the heat 

 loss from the wires decreases steadily as the pressure is lowered. Only at 

 pressures above 200 mm. do convection currents play any important part. 



