1024 
PROFESSOR J. W. MALLET ON A REVISION 
Pemberton, Esq., Vice-President of the Pennsylvania Salt Manufacturing Company 
at Natrona, Pennsylvania, and to W. N. Richard, Escp, of the same works, for an 
abundant supply of aluminum hydrate, such as is thrown down by a stream of carbon 
dioxide from solution of sodium aluminate in the process of making soda from cryolite. 
This was not absolutely free from iron, but one lot contained traces only of this metal, 
insufficient, as it turned out, to contaminate the aluminum to be made in contact with 
it. The hydrate was strongly heated in well covered crucibles until it ceased to give 
off water; the alumina which was left required then to be guarded from the air, as it 
readily took up moisture again. All attempts to use it mixed with water and any 
unobjectionable cementing material to a plastic mass failed from excessive shrinkage 
and crumbling, but by mixing it in the dry state with dry sodium aluminate better 
results were obtained. The sodium aluminate had to be specially prepared, as that 
made (for soap boilers’ use) at Natrona from cryolite contained too much iron. The 
dry mixture was pressed into wooden moulds, and three or four crucibles thus made 
having been very slowly and cautiously heated up in a gas furnace, stood fairly the 
necessary temperature of the reduction of aluminum, although they were very fragile. 
On the whole, however, it was found best to use a highly aluminous Beaueaye 
crucible, with a thick lining of this mixture of dry alumina and sodium aluminate well 
rammed in and very gradually heated. There were some failures from cracking of the 
crucibles or linings, and whenever the slightest contact of the metallic aluminum with 
the outer crucible occurred, silicon and generally iron were sure to be found in the 
metal. With successful prevention of this by an adequately thick and perfectly 
continuous lining there was much difficulty in securing a sufficiently high temperature 
in the interior, since the conducting power for heat of the alumina linings seemed to 
be quite low. This led to much loss of sodium by combustion, and but a very small 
yield indeed of really pure aluminum was secured. When the crucible had been 
heated up ready for the reduction, a small quantity of the pulverised mixture of 
potassium and sodium chlorides was thrown in; soon afterwards the aluminum bromide 
(which had been fused with the alkaline chlorides) with the proper quantity of sodium 
was introduced,* and as soon as the violent reaction was over a further portion of the 
mixed alkaline chlorides. 
Only the large, well-fused globules of aluminum were picked out; these were re¬ 
fused once or twice before a blowpipe flame upon a support of alumina, to free them 
from any possible remains of the flux; any trace of oxide was detached by acting 
slightly upon the surface with pure hydrochloric acid, and the globules were then well 
washed with water, and dried by a gentle heat. Specimens cut from different portions 
of the globules were carefully tested, particularly for silicon, iron, sodium, and 
* Aside from the question of expense there is some advantage in using aluminum bromide for making 
the metal, on account of the low melting-point of the sodium bromide which is left. Cabnelley’s recent 
determinations (Chem. Soc. Journ., July, 1878, pp. 279, 280) make the melting-point of sodium bromide 
708° C.> while that of the chloride is 772°, and that of the fluoride above 902°. 
