12 BULLETIN 1179, U. S. DEPARTMENT OF AGRICULTURE. 
The furnace consisted of a water-cooled crucible, with the cooled section extend- 
ing no higher than the region of the molten slag. It was lined with fire-clay brick, 
but silica brick would prove more satisfactory. The portion not exposed to the 
action of slag was lined with a fire-clay brick. All gas mains and the cooling tower 
had a fire-clay brick lining. The heat from the gases served to harden the exposed 
surface and thus improve the service of the brick. The electrodes entered through 
the top of the furnace, but below a line where the charge entered. Care should be 
taken in the design of such furnaces that the angle of the electrodes conform with 
the angle of repose of the charge. Thus as the charge falls in a natural pile, the 
breakage of electrodes is eliminated. Electrodes may be conveniently controlled by 
hand, or mechanically. Hand control was used in this experiment, with the control 
so located that the switchboard and instruments could be observed. Six-inch and 
four-inch graphite electrodes were used. The life of a 4-inch electrode was about 7 
days, while the 6-inch electrodes lasted on an average of 10 days under favorable 
conditions. Thus it may be seen that with such a low consumption electrodes may 
be operated by hand, since the chief movement of electrodes is when they are con- 
sumed. In this experiment the charge was fed by hand; however, this is obviously 
impractical in a laige installation, where mechanical apparatus should be used. 
During regular operation about 2,000 pounds of rock were consumed per 12-hour 
period. 
A slag pit filled with water was used to quench the molten slag as it flowed from 
the furnace. The sJag thus chilled slid to one end and was removed mechanically. 
The P 2 5 content of the slag was approximately 2 per cent, although it is possible 
to reduce it to 1.5 per cent or even 1 per cent for regular operation. The Pj0 5 con- 
tent of the slag is largely a matter of the mixing of the charge and using the proper 
proportions of rock, sand, and coke. 
The average pioduction was 0.3 pound H 3 P0 4 per kilowatt hour absorbed; 
however, there were periodic yields, during times of good operating conditons, in 
which 0.4 pound H 3 P0 4 per kilowatt hour was produced. Judging from the average 
results of this experiment it seems reasonable to assume that a production of 0.6 
pound H 3 PG 4 per kilowatt hour is possible. Of course, the production is entirely 
dependent upon the efficiency of the furnace. In the case of this work no means 
were adopted to utilize the heat absorbed by the water surrounding the crucible or 
in the gases carried over from the charge. Also there were heat losses from the 
oxidation of phosphorus to phosphorus pentoxide (P 2 5 ) and carbon monoxide (CO) 
to carbon dioxide (C0 2 ) which, if utilized, would materially have increased the 
efficiency of the process. 
As the gases were removed from the furnace they passed through a cooling tower 
before entering the treater. This tower was installed to afford sufficient radiation, so 
that the gases entered the treater at 250° to 300° C. Above these temperatures in 
the treater electrical and mechanical difficulties arise which make higher temper- 
atures undesirable. 
The treater consisted of a header of common brick, with a reinforced concrete top 
to support the pipes, and 20 treater tubes of vitrified sewer pipes, 12 inches in 
diameter and 15 feet in length. All joints were packed loosely with silica to prevent 
air from entering at these points. The pipes were inclosed to prevent cracking, due 
to heat differences, and to maintain an even flow of gas. 
All pipes at the top were inclosed in a common hood. Supports for the conductors 
rested on insulators within the treater hood. Complete clearance was given to 2,000 
cubic feet of gas entering at 300° C, with a velocity of 3 linear feet per second. * * * 
Power was supplied the treater from a 150-volt motor-generator set, and transformed 
to higher voltages by a 7.5 kilovolt ampere transformer. A 5-point switch on the 
low-tension side of the transformer, connecting the various turns of coils, made a 
variation of voltages possible. It was found that 70 kilovolts was sufficient to give 
complete precipitation of the gases, at the above stated volumes and velocity. 
As the acid fell from the pipes it was caught in a receiving basin of vitrified brick 
set in acid-proof cement. From this basin the acid flowed out and was disposed of 
by pumping to a receiving vat. The concentration of the acid collected was con- 
trolled by the temperature of the gas in the treater. At a temperature of less than 
100° C. the concentration is not likely to exceed 50 per cent H 3 P0 4 , while a temper- 
ature of 250° to 300° C. will yield an' acid of 85 to 93 per cent H 3 P0 4 . In one case 
an acid of 97 per cent was produced. An acid above 85 per cent H 3 P0 4 will prob- 
ably solidify when it reaches atmospheric temperatures, and therefore the pumping 
apparatus and pipe lines should be so constructed as to prevent clogging. 
From the results obtained in this plant Carothers estimated the 
yield of phosphoric acid (P 2 5 ) in a 3,000-kilowatt furnace operating 
