1889.] on Aluminium. 457 



boxes, which are built in brick. These boxes are provided with 

 openings or doors, and also with earthenware pipes connected with 

 a small flue for carrying off the uncondensed vapours to the large 

 chimney. At first the chlorine passed into each retort is all absorbed 

 by the charge, and only carbonic oxide escapes into the o])en boxes, 

 where it burns. After a certain time, however, dense fumes are 

 evolved, and the boxes are then closed, while the connecting pipe 

 between the box and the small flue serves to carry off the uncon- 

 densed vapours to the chimney. 



The reaction which takes place is as follows : — 



AlA + 2NaCl + 3C + 6C1 = 2AlCl3NaCl + 3C0. 



The chlorine is passed in for about 72 hours in varying quantity, 

 the boxes at the back being opened from time to time by the work- 

 men to ascertain the progress of the distillation. At the end of the 

 time mentioned the chlorine valves are closed and the boxes at the 

 back of the furnace are all thrown open. The crude double chloride, 

 as distilled from the retorts, condenses in the connecting pipe and 

 trickles down into the boxes, where it solidifies in large irregular 

 masses. The yield from a bench of five retorts will average from 

 1,600 to 1,800 lbs., which is not far from the theoretical quantity. 

 After the removal of the crude chloride from the condenser boxes, 

 the retorts are opened at their charging end, and the residue, which 

 consists of a small quantity of alumina, charcoal, and salt, is raked out 

 and remixed in certain proportions with fresh material, to be used 

 over again. The furnace is immediately re-charged and the same 

 operations repeated, so that from each furnace upwards of 3,500 lbs. 

 of chloride are obtained weekly. With ten of the twelve furnaces 

 always at work the plant is easily capable of i^roducing 30,000 lbs. 

 of chloride per week, or 1,500,000 lbs. per annum. 



Owing to the presence of iron, both in the materials used (viz., 

 charcoal, alumina, &c.) and in the fireclay composing the retorts, the 

 distilled chloride always contains a varying j^roportion of this metal 

 in the form of ferrous and ferric chlorides. When it is remembered 

 that it requires 10 lbs. of this chloride to j^roduce 1 lb. of aluminium 

 by reduction, it will be quite apparent how materially a very small 

 percentage of iron in the chloride will influence the quality of the 

 resulting metal. I may say that, exercising the utmost care as to 

 the purity of the alumina and the charcoal used, and after having the 

 retorts made of special fireclay containing only a very small per- 

 centage of iron, it was found almost impossible to produce upon a 

 large scale a chloride containing less than * 3 per cent, of iron. 



This crude double chloride, as it is now called at the works, is 

 highly deliquescent, and varies in colour from a light yellow to a dark 

 red. The variation in colour is not so much due to the varying per- 

 centage of iron contained as to the relative proportion of ferric or 

 ferrous chlorides present, and although a sample may be either very 

 dark or quite light, it may still contain only a small percentage of 



I 



