304 PHENOMENA, ATOMS, AND MOLECULES 



The values for 8 given in the above table are to be looked upon as lower 

 limits. No correction was made for the cooling effect of the leads, which 

 was very considerable with the short filament used. This correction would 

 be much more important in the case of the combustion of carbon monoxide, 

 since with this reaction the maximum velocity occurs only over a narrow 

 range of temperature. Furthermore, traces of carbon dioxide were found 

 to slightly poison the surface of the platinum, so that the catalyzer only 

 slowly recovered its full activity. It may well be that minute traces of 

 impurities have made the values of 8 lower than they should be under 

 ideal conditions. 



From these considerations it seems probable that the values of 8 cor- 

 responding to perfectly pure gases and a filament at uniform temperature, 

 might approach unity as a limit in the case of each of three gases. 



These experiments indicate that the reflectivity of oxygen, hydrogen, 

 and carbon monoxide molecules is small even at temperatures of 900° K. 

 There is every reason to believe that the reflectivity at room temperature 

 is not greater than at higher temperatures. 



A large number of other heterogeneous reactions are being studied by 

 similar methods and the coefficients 8 are being determined. All of this 

 work so far seems to bear out the general conclusion that values of 8 much 

 less than unity occur only where the surface of the solid is largely covered 

 by some inactive material. The velocity of the reaction is thus determined 

 by that fraction of the surface which is active. 



(b) Homogeneous Reactions. — Although homogeneous reactions furnish 

 no direct evidence regarding the reflectivity of molecules from solid sur- 

 faces, yet several cases that have been studied have indicated that re- 

 actions are by no means rare in which every collision between unlike 

 molecules results in combination. In this case we may consider that the 

 reflectivity between the molecules is negligible, and are thus led to suspect 

 that in many gas reactions, when two molecules collide, they stick to 

 each other (condense), although they may subsequently separate again 

 (evaporate). 



When a tungsten filament is heated to 2800° K. in vacuum, it evaporates 

 at the rate of 0.43 X io~^ grams per sq. cm. per second.-^ If the filament 

 is surrounded by nitrogen at a pressure less than 500 bars, the rate of 

 evaporation remains unchanged, but each atom of tungsten vapor com- 

 bines on its first collision with a nitrogen molecule to form the definite 

 chemical compound WN2, which collects on the bulb as a brown deposit. 

 The nitrogen therefore disappears at the constant rate (independent of 

 pressure) of .057 cubic mm. of nitrogen (at i megabar) per sec. per sq. 

 ^^ Langmuir, Phys. Rev., 2, 340, 1913. 



