MOTION OF IXniVIDlAL DOMAIN WALLS 1033 



Sample 4- The stationaiy walls at the corners, although patchy, were 

 seen. There was some extraneous domain wall structure on the major 

 (110) faces, but nothing which looked at all like the main wall. 



On the outside (112) faces of the legs, however, the wall was traced 

 almost all the way around. Only short sections were impossible to trace. 

 The wall curved as usual, and there was some extraneous domain wall 

 structure on these faces, but substantiallj'^ the whole of the ideal pattern 

 shown in Fig. 1 was seen on this sample. The hysteresis loops shown in 

 Figs. 2 and 3 and the domain pattern pictures shown in Figs. 4 and 5 

 were taken on this sample. 



Not only was the domain pattern on sample 4 the best and most com- 

 plete, but the data taken on this sample was much the most reproducible. 

 We shall therefore report the data taken on this sample in detail, and 

 simply refer to the results on other samples as a check and to indicate 

 the sort of variations which occurred from sample to sample. 



Measurements 



Our procedure in making the measurements is as follows. The sample 

 is wound with a primary and a secondary winding. A square pulse of 

 positive voltage is applied to the primary winding in series with a re- 

 sistor which is large enough to keep the pulse rise time short. The rise 

 time must be short compared to the time required for the field produced 

 by the pulse to reverse the magnetization of the sample. On the other 

 hand, since the pulse is applied for the purpose of reversing the magneti- 

 zation of the sample, the length of the pulse must be at least comparable 

 with the time required for the reversal to occur; if possible, it should be 

 longer than this. The reversal time, of course, is the time required for 

 the mobile domain wall to move from one side of the sample to the other 

 under the field produced by the applied pulse. A second pulse, of nega- 

 tive voltage, is applied to the primary during each cycle of the pulser 

 in order to bring the wall back to its original position so that the phe- 

 nomenon may be observed repetitively. 



By sjaichronizing an oscilloscope sweep with the pulser, the signal 

 induced in the secondary winding is observed while the applied pulse 

 is on the primary. Since this signal is proportional to the velocity of the 

 wall, it is constant to a first approximation during the application of a 

 constant field. Irregularities in the crystal may cause the velocity of the 

 wall to vary somewhat as it moves across the sample, however, and this 

 will cause the signal to vary too. In this case, the observer reads the 

 average value. Fig. 6 shows an example of the signal induced in the 

 secondary winding as seen on an oscilloscope. Sample 4 was used to 

 obtain this picture. 



