1026 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1954 



G. W. Clark from the Liiide Air Products Company. Similar studies 

 have been performed pre^'iously at, room temper-ature on Fe304 ' by 

 the author, and in a preliminary way on (Xi())(i.9i (FeO)o.,,9 FeoOs ^ by 

 the author in eollaboration with J. A. Andrus and H. G. Hopper. 



THE EXPERIMENTS 



Preparation of Samples 



The key to the success of experiments of this sort is of course obtain- 

 ing the domain pattern shown in Fig. 1, so that we ha\'e only one mo\'able 

 wall in the sample. The achie\'ement of this pattern, and the obser\'a- 

 tion of it when achieved, depend in turn on success in producing a 

 perfect or almost perfect sample. It therefore seems worth while to 

 describe the process which has finally emerged as a satisfactory way of 

 producing these samples. 



The rough crystal is first oriented by means of X-rays (Laue and X-ray 

 goniometer techniques) to an accuracy of a few minutes while it is 

 mounted in such a position that afterwards we can grind a flat on it 

 w^hich is coincident with the (110) plane. This flat is ground simply with 

 a belt grinder. A cut is then made with a diamond saw parallel to this 

 flat, so that we have a disc whose faces are (110) crystal planes. This 

 disc is usually made about 2}/^ to 3 mm thick. The second face is ground 

 accurately parallel to the first in a paralleling block. A flat coincident 

 with the (100) plane is ground on the edge of this disc for later use in 

 orienting the sample in the plane of the disc. This flat is also ground with 

 a belt grinder after orienting the disc wdth Laue and X-ray goniometer 

 techniques. 



The disc is ground as smooth as possible with 3031.2 emery, and then 

 very carefully polished. The polishing is done first on a lap surfaced 

 with No. 0000 french emery paper, with Linde A abrasive loose on top 

 of the emery paper. After this the lap is surfaced with a sheet of \'ery 

 smooth paper and then Linde B abrasive is used. The polishing process 

 takes four to eight hours per disc, and removes all pits ^'isible under 

 50 X magnification, except those inherent in the crystal. Only on such a 

 smooth surface can the very small holes which sometimes occur in these 

 crystals be seen. Sometimes, however, the polishing process conceals 

 fine (Tacks. It also cannot reveal variations in chemical composition 

 which sometimes occur from point to point in the crystals. Such com- 

 position variations are presumably variations in the concentration of 

 divalent iron from point to point in the crystal. In order to reveal these 

 latter imperfections, the disc is etched for three hours by boiling it in 



