66 sosman: problems of the oxides of iron 



If precipitated magnetite is oxidized by ammonium persul- 

 fate, or even by atmospheric air at a low temperature, the prod- 

 uct, although it contains almost no ferrous iron and is reddish 

 brown in color, is practically as magnetic as magnetite itself. 

 If this magnetic oxide is heated to 750° for a few minutes, its 

 highly magnetic character is lost, and it becomes like ordinary 

 ferric oxide. What is still more remarkable is that the magnetic 

 Fe 2 3 has a Curie inversion just above 500°, similar to that of 

 magnetite. 



Here, then, we have as fundamental a chemical change as it 

 is possible to get — no mere polymerism, but a change in the 

 formula weight with a reduction of the ferrous iron content from 

 33| per cent to zero; yet the magnetic susceptibility and inversion 

 point remain the same. Except for the color, it is as if we had 

 merely ground up the magnetite and mixed it with powdered 

 solid oxygen. 



A consideration of the preceding facts concerning the inversions 

 of iron and its compounds, and especially of the interesting 

 properties of the magnetic form of Fe20 3 , suggest that the mag- 

 netic properties and the inversions of iron and its compounds 

 are bound up with the spacing and arrangement of the iron atoms, 

 almost without regard to the other atoms present. 21 



What is this structure of the iron atoms which produces the 

 high permeability, and what is the nature of the intra-atomic 

 change at the inversion point? 



These problems are on the way to a solution, for Hull has 

 recently found by means of X-ray spectra that the lattice of 

 metallic iron is the centered cube. 22 He discovered also that 

 this structure seems to remain unchanged through both the 



21 This has been suggested for the Heusler alloys: Take. Faraday Soc. 

 Trans., 8: 177. 1912. Professor Pupin, speaking in the recent symposium on 

 the structure of matter, at the American Association meeting in New York, 

 December 27, 1916, stated that he had come to the conclusion from the magnetic 

 properties of very pure iron that its magnetism is intimately related to its 

 crystalline structure. On the other hand, as Hilpert points out, precipitated 

 oxides that are apparently amorphous and that pass readily into colloidal solu- 

 tion can be made, which are magnetic and possess magnetic inversions. Have 

 the particles of these colloids a microcrystalline structure, as yet unrecognized? 



22 Paper presented at the Cleveland meeting of the American Physical Society, 

 November, 1916. 



