HYDROGEN INTO ATOMS 193 



wire is surrounded by atomic hydrogen, while the partial pressure of molec- 

 ular hydrogen is zero. In this case 



oji = — cim-x (74) 



and we thus find, from (26) 



K = a/5 (Lz:Ji>lPi. (75) 



This shows that at low temperatures 81 must be constant and equal to 

 unity, whereas at high temperatures it decreases and becomes proportional 

 to the pressure. Thus again, the third hypothesis offers a simple explana- 

 tion of the difficulty encountered in applying the second hypothesis. 



It is evident that ei and £0 completely lose their significance when botli 

 hydrogen molecules and atoms are present around the wire. We cannot 

 profitably look upon the reaction as being the difference between two op>- 

 posing reactions taking place upon the surface. 



In fact, the velocity of the reaction does not follow Gulberg and 

 Waage's Mass Law in the ordinary sense. Neither do diffusion phenomena 

 (at least at low pressures) seem to have any effect on the reaction velocity. 

 The velocity is determined simply by the rate at which the molecules and 

 atoms strike the surface of the wire and the equilibrium concentrations of 

 atomic and molecular hydrogen within the wire. 



Effect of Bulb Temperature. — In studying chemical reactions at low 

 pressures it has been pointed out ^^ that the temperature of the filament 

 and that of the gas molecules striking it can be varied independently of 

 one another. By changing the bulb temperature, the velocity (and internal 

 energy) of the hydrogen molecules striking the filament can be varied over 

 a wide range. If the reaction takes place entirely on the surface in ac- 

 cordance with the second hypothesis, we should expect that the relative 

 velocities of the hydrogen molecules and the tungsten atoms with which 

 they collide, would determine the velocity of the reaction. Because of the 

 small mass of the hydrogen molecules it happens that the average velocity 

 of hydrogen molecules at room temperature or even liquid air temperatures, 

 is much greater than that of tungsten atoms at 3500°. Therefore, the 

 amount of impact between the hydrogen molecules and tungsten atoms 

 depends vastly more on the bulb temperature than it does on the tempera- 

 ture of the filament. In case the impact' determines the reaction velocity, 

 we see that the latter should be greatly affected by bulb temperature. 



On the other hand, if the reaction takes place in the metal in accord 



^^ Langmuir, Jour. Amer. Chem. Soc, 55, 105 (1913). . 



