M. Deville on the various forms of Silicon. 271 



heat, and the operation is continued until no more vapours of 

 chloride of aluminium are formed. The needles of silicon which 

 are formed are purified from adherent matters by treatment with 

 aqua regia, hydrofluoric acid, and fusion with bisulphate of soda. 

 Sometimes there are formed small globules of a siliciuret of 

 aluminium which corresponds to the formula Si AP. 



The change in this reaction is thus : the chloride of silicon 

 is decomposed by the aluminium, which seizes upon the silicon 

 displaced, from whence a veritable solution results. Each mole- 

 cule of chloride which is decomposed, increases the concentra- 

 tion ; and when the saturation of the metallic bath is complete, 

 the silicon, being lighter, crystallizes on the surface, as does cam- 

 phor on the surface of an alcoholic solution. 



The crystals of silicon occur sometimes as long needles, some- 

 times as hexagonal prisms culminated by a very acute pyramid 

 with curved faces, and sometimes in forms resembling a rhombo- 

 hedron, but which are derived from the regular octahedron. 

 Senarmont* has shown that the regular hexagonal prisms of the 

 silicon are only faces of the rhomboidal dodecahedron developed 

 in a very simple manner. 



Silicon has a dark iron-gray colour with a reddish lustre, 

 and often resembles the specular iron ore of Elba, with all its 

 iridescence. The crystals are perfectly pure; they melt at a 

 point intermediate between the fusion of gold and that of cast 

 iron, and they then assume the form analogous to the diamond, 

 with curved faces, which appears peculiar to silicon obtained by 

 fusion. 



Deville is still engaged on an extension of this method to the 

 preparation of other bodies. He collates, in conclusion, the pro- 

 perties of silicon and of carbon, and shows in a very striking 

 manner their analogy. 



By substituting the fluoride for the chloride of silicon in this 

 preparation of silicon, Deville obtained along with silicon a 

 substance crystallizing in cubes, which has no action on polar- 

 ized light, is transparent and strongly refracting. It is fluoride 

 of aluminium, Al^ FF, quite free from silicon, as numerous ana- 

 lyses proved. It is unattackable by hydrofluoric and nitrofluoric 

 acid, and suljjliuric acid disengages only mere traces of hydro- 

 fluoric acid. It is only volatile at u bright red heat. It may be 

 prepared directly by treating calcined alumina with ])ure hydro- 

 fluoric acid in excess, drying the mixture strongly, and then 

 introducing it into a tube of carbon or platinum which is heated 

 to bright redness, and a current of hydrogen passed through. 

 The fluoride of aluminium sublimes, and is deposited in minute 

 cubes in the cold part of the tube. It is also prepared by ex- 



* Annates de Chimie et de Physique, Juuc 1866. 



