OF THE ATOMIC WEIGHT OF ALUMINUM. 
1023 
With a view to render the bromide more manageable while in contact with the 
air, it was decided to unite it with an alkaline chloride, as the aluminum and sodium 
chloride is used in the ordinary process of Deville, and in order to keep down 
the melting point, a mixture of both potassium and sodium chlorides was employed. 
The two alkaline salts were separately purified, carefully tested, well dried in plati¬ 
num vessels, mixed in the proportion of one molecule of each, the mixture fused in 
lots of about 250 grms. in a platinum dish over a gas furnace, poured out into 
another such dish standing on a block of cold iron, the thin cake gently crushed to 
a coarse powder while still warm in a mortar of hard glazed porcelain (from which, 
it was afterwards proved by examination, no silica was taken up), and the pul¬ 
verised product kept until needed in a well stoppered bottle. Separate portions 
of the purified aluminum bromide having been in the last distillation collected in tared 
glass flasks and weighed, a quantity of the above mixture of alkaline chlorides was 
weighed off for each corresponding to 1 molecule (K-fNa)Cl for 1 molecule AlBr 3 , 
and the flask having been heated until the bromide was fused, the potassio-sodic 
chloride was cautiously added. Very marked rise of temperature occurred, so that 
in a first attempt, using considerable quantities of the materials and mixing them 
abruptly, since such an effect had not been foreseen, the flask was violently cracked, 
and torrents of aluminum bromide vapour were driven off. This evolution of heat 
is interesting as evidence of chemical combination taking place, not only between two 
chlorides or two bromides, but between a chloride and a bromide. The fused mass 
was on cooling crushed to small fragments and coarse powder and preserved in a 
well-stoppered bottle. The precaution was not omitted of testing for any evidence 
of impurity derived from the flasks or mortar, but with negative result. The material 
thus prepared did not fume in the air, was sufficiently slow in taking up atmo¬ 
spheric moisture to be managed without difficulty, and fused on re-heating to about 
130° C. The sodium to be used in decomposing it was in large ingots, which when 
needed were wiped to remove naphtha, the outer crust cut off with a knife, the large 
pieces roughly weighed, and the proper quantity rapidly cut up into small fragments 
without again moistening with naphtha. 
The great difficulty in the way of obtaining pure metallic aluminum consists in 
obtaining crucibles of suitable material, especially such as shall not yield either iron 
or silicon. M. Tissier"" seems to have been more fortunate than I in the use of 
carbon vessels. The crucibles of hard carbon which I had on hand, purchased in 
Germany, contained both the above-named impurities, which were taken up in no 
small quantity by the aluminum; and I failed in sundry attempts to make crucibles 
of purer carbon, or to use this substance as a lining, the carbon either burning away, 
crumbling up, or permitting the fused materials to pass through its pores or through 
cracks in the mass. I at last succeeded in adapting to my purpose alumina itself, 
sufficiently cemented together by sodium aluminate. I was indebted to Henry 
* ‘ Comptes Rendus,’ loc. cit. 
