100 THE NITROGEN PROBLEM. 



generally constructed for 600 to 1,000 kilowatts. Compared 

 with the Birkeland-Eyde furnace, it is much smaller, and a 

 larger installation is therefore required to give the same 

 output. On the other hand, the apparatus is simjjle, durable 

 and comparatively inexpensive. Tlie further treatment of the 

 nitrogen oxides and the final products obtained are the same 

 as in the former process. 



1 c. — The Pauling Process. 



In this arcs are employed, which are produced between 

 two electrodes curved like those of the so-called horn lightning 

 arresters, the air being blown through and along them from 

 the parts of the electrodes which are nearest together, the 

 minimum distance being large enough to allow of the air 

 passing through. TJie preheated air current is blown over the 

 arc in such a manner that it passes along the whole length of 

 the electrodes, which are made of iron and are water cooled. 

 Cooling of the exit gases is effected by passing cold air into 

 the upper part of the flame from the side at a slower rate 

 than the main air current, wliich has the effect of broadening 

 the flame. Two arcs are contained in one furnace, and for a 

 400-kilowatt furnace 600 cubic metres of air are used per 

 hour. The gases leave the furnace at a temperature of 800° C. 

 In 1909 it was stated that 60 grams of nitric acid were 

 produced per kilowatt hour by this process, which was first 

 established on the large scale at Patsch, near Innsbruck, in 

 Tyrol, and afterwards at La Eoche de Eame, Hautes Alpes, 

 France. The furnaces are ^im])le in construction, about 

 3ft. by 4ft. in horizontal section and 10ft. liigh, and about 

 600 to 1,000 kilowatt capacity. The air is preheated before 

 entering the furnace by utilising the heat of the gases leaving 

 the furnace, and this is also utilised to concentrate the acid, 

 which is marketed as such or may be neutralised to form salts. 



/ (1 . — The Moscicici Process. 



This has been worked hitherto on a semi-commercial scale 

 and differs from those previously mentioned in the method of 

 obtaining the flame. Two concentric ring electrodes are used, 

 the arc passing from one ring to another in a radial direction 

 and being magnetically deviated as in the Birkeland-Eyde 

 process, which causes it to revolve continuously arf)uncl in the 

 annular space between the rings, hence tlie designation which 

 has been given to this, namely, tlie revolving flame metliod. 

 It is stated to liave given a yield of 60 grams nitric acid per 

 kilo'\\att hour and to be well adapted for direct current o^^jera- 

 tion, It has, however, now been discontinued. 



In addition to tlie simple reactions already stated to occur 

 when air is passed through an arc flame, there are others 

 which militate against a complete theoretical extraction of the 

 oxides of nitrogen produced. The gas wliich leaves the oxida- 

 tion tanks for the absorption system consists of 98' 5 to 99 per 

 cent, of air and 1 to T5 per cent, of the mixed oxides of 

 nitrogen from XO to XoOf which when dissolved in water 

 yield nitrous and nitric acids. In the absorption towers other 



