MO'I'IOX OF l\I)IVinr\L DOMAIN WALLS 1027 



50 per cent H2S()4 uiuler a reflux condenser. An asl)esto.s pad is placed 

 between the flame and the bottom of the flask containing the II2SO4 

 ill order to pre\-ent sharp temperature fluctuations in the bath. The rate 

 of the etching attack, and the ([uaUty of the surface it leaves are sharply- 

 dependent upon temperature. It should be mentioned that if this etch 

 is used on the discs before polishing, the surface remains rough or may 

 c\ (Ml be made rougher, so that it is impossible to detect the imperfec- 

 tions in the disc. The etch must start on a smooth surface. 



Once the disc is cut, polished, and etched, if it is found to be sufiiciciitly 

 U-cc of imperfections, a sample is cut from it in such a way as to avoid 

 those which there are, as they are revealed by the polishing and etching 

 processes. First the diamond-shaped hole is cut by means of a jig whose 

 rotational posiiion with respect to the disc is determined from the (100) 

 flat on the edge of the disc. This jig is a piece of steel which is driven in 

 \ibiation \ertically with a magnetostrictive drive.^ The surface of the 

 disc is covered with a sIvht}^ of carborundum or diamond dust, and this 

 abrasive is made to cut a hole in the disc as the vibrating jig is slowly 

 lowered. With the hole cut in the proper orientation, the outer parts of 

 the disc are ground down to form the legs of the sample. Another jig of 

 the proper shape is used to hold the sample in position during this process. 



A hysteresis loop is taken as soon as the sample is cut. A relatively 

 good loop taken on our best sample is shown in Fig. 2. All stich loops 

 on these samples are taken on the Cioffi recording fluxmeter.'" This 

 loop is obviously not yet in the form which we finally need. In order to 

 scpiare the hysteresis loop, we anneal the sample for approximately an 

 hour at ()00°C in a magnetic field of 10 to 20 oersteds. The field is pro- 

 duced by running a current through a few turns of glass insulated wire 

 wound on the sample. After such a heat treatment the hysteresis loop 

 of this sample assumed the form shown in Fig. 3. 



Once the sample is prepared, the next problem is to observe the domain 

 pattern and find if any important deviations from the pattern shown 

 in Fig. 1 occur. The heat-treatment we give them corrodes the polished 

 surfaces of the sample, and of course the faces exposed when the sample 

 is cut fi'om the disc have not yet been polished. Consequently both the 

 major (110) faces of the sample and the outer faces of the legs are pol- 

 ished, and the sample is then etched in the same way as before. Usually 

 a hysteresis loop is again taken at this point as a check. If the sample 

 is good, it is not significantly different from the loop taken immediately 

 after heat-treatment. The sample is l)rought to a demagnetized con- 

 dition at this point so that the movable wall will be near the center of 

 the sample where it can be observed. This completes the process of 



