FLAMES OF ATOMIC HYDROGEN 117 



surfaces that otherwise converted atomic into molecular hydrogen. Thus 

 with moist h3-drogen the tube became filled with nearly pure atomic hydro- 

 gen and the diffusion of this to the catalytically active tungsten wire caused 

 the heating of the latter. Calculations based on an equation like (13) proved 

 that a pressure of only 0.16 mm. of atomic hydrogen at 500 deg. C. would 

 suffice to maintain a tungsten filament at 2400 deg. K. 



These conclusions were confirmed by Wood's observations that the 

 walls of the tube became only slightly heated if the hydrogen was moist, 

 whereas they were strongly heated with dry hydrogen. A tungsten wire was 

 heated red hot even when mounted in a side tube (of 5 mm. diameter) at a 

 distance of 4 cm. from the discharge tube, showing that the hydrogen atoms 

 could diffuse in relatively large quantities out of the discharge. 



It occurred to the writer that it should be possible to obtain even higher 

 concentrations of atomic hydrogen by passing powerful electric arcs 

 between tungsten electrodes in hydrogen at atmospheric pressure and by 

 blowing atomic hydrogen out of the arc by a jet of hydrogen directed 

 against it. 



ARCS IN HYDROGEN AT ATMOSPHERIC PRESSURES 



Studies had been made several years ago in this laboratory of arcs 

 between tungsten electrodes in various gases. ^* Arcs in hydrogen were 

 remarkable because of the high voltage drop and small cross-section. A 

 lo-amp., direct-current arc between heavy tungsten electrodes about 7 mm. 

 apart in a bulb containing pure hydrogen at atmospheric pressure appeared 

 as a sharply defined brilliant red line about 0.5 mm. in diameter along which 

 the potential gradient was 150 volts per cm., this being about 15 times as 

 great as in nitrogen or argon. This abnormal behavior of hydrogen was 

 attributed to the dissociation which carried energy so rapidly out of the 

 arc. 



A simple analysis shows that ordinary heat conduction (i.e., without 

 dissociation) and even convection are entirely incapable of causing such a 

 flow of heat as 1500 watts per cm. of length of arc. For the flow of heat 

 between concentric cylinders of radii a and r the shape factor 5" in equa- 

 tion (8) has the value 



^^'\ — r-r\ (14) 



log (fl/r) 



where / is the length of the cylinder. The surfaces of equal temperature 

 which surround the arc are concentric cylinders. Let us consider the two 

 surfaces which are the loci of points having the temperatures 2000 deg. 

 and 1000 deg., respectively. The value of (</)2 — </>i) corresponding to these 



** G. M. J. Mackay and C. V. Ferguson, /. Franklin Inst., 181, 209 (1916). 



