36 PHENOMENA, ATOMS, AND MOLECULES 



phase, even by introducing caesium vapor as a getter, does not remove the 

 oxygen film on the tungsten surface after it has once been formed. This 

 means that the oxygen from such a film does not evaporate appreciably at 

 i,500°K. Measurements at higher temperatures show that about half of 

 the adsorbed oxygen evaporates in 27 minutes' time at i ,860° and in about 

 20 seconds at 2,070° (12), From the temperature coefficient at this rate 

 of evaporation it can be concluded that it will take about 3 years at 1,500° 

 to remove half the fihn by evaporation and that the heat of evaporation is 

 of the order of 160 K. cal. per gram atom. 



This heat of evaporation is far greater than the heat of dissociation 

 of oxygen into atoms ; so we have very direct evidence that the forces that 

 hold oxygen to a tungsten surface are comparable to the most powerful 

 chemical forces known. This gives reason for believing that the oxygen 

 film which so greatly decreases the electron emission from tungsten consists 

 of a single layer of oxygen atoms which are chemically combined with the 

 underlying tungsten atoms. 



The electron emission from the tungsten filament in the presence of 

 oxygen at temperatures below 1,800° is found to be independent of the 

 pressure of oxygen, provided that more than about io~^ barye is present. 

 This must mean that the surface is practically completely covered with 

 oxygen and that an increase in oxygen pressure does not cause the thick- 

 ness of the layer to increase beyond that of a single layer of atoms. 



Curve III in Fig. i indicates the way in which the pressure of oxygen in 

 the bulb decreases with time when the tungsten filament is maintained at 

 i,500°K. The rate of disappearance of the oxygen is proportional to the 

 oxygen pressure and there is no fatigue effect, such as was observed during 

 the clean-up of hydrogen. 



Interaction of Oxygen with Hydrogen in Contact with Tungsten Fila- 

 ment. Mixtures of oxygen and hydrogen at low pressures in a bulb contain- 

 ing a tungsten filament behave ui an extraordinary manner (See Ref. 13 

 p. 2271; Ref. 14; and Ref. 15 p. 608). Typical results obtained with a 

 filament temperature of 1,500° are shown in Curves IV and V of Fig. i. 

 These curves were obtained with a mixture of 3 parts of hydrogen and 5 

 parts of oxygen. When the filament was lighted, the gas disappeared at 

 exactly the same rate as shown in Curve III when 5 parts of oxygen alone 

 were present. After about 14 minutes practically all of the oxygen had dis- 

 appeared. This was confirmed by analysis of the residual gas which was 

 found to be pure hydrogen. This hydrogen, however, did not dissociate in 

 the usual way into atoms and disappear by adsorption on the glass walls 

 as shown in Curve I, but the pressure remained practically constant until 

 after about 24 minutes, when the pressure suddenly began to decrease as 

 indicated by Curve V. This Curve V, however, is identical with Curve I, 



