MAGNETIZATION IN CRYSTALS — DILLON 



401 



Figure 9. — An indication of the way in which the magnetization direction varies on passing 



through a 180° domain wall. 



thinner, and the other forcing it to be thicker. The actual thickness, 

 of course, depends on the size of the maximum anisotropy energy for 

 the crystal at that particular temperature and the size of the exchange 

 energy. For yttrium iron garnet at room temperature, we would 

 expect the wall to be about 7,000 A thick. This is equal to 550 times 

 the lattice constant, the edge of the cubic unit cell. 



If there are neighboring domains in which the magnetization is 

 directed along two different easy directions, a little consideration will 

 show that in these crystals we could have 180°, 110°, and 70° walls. 

 In crystals in which the polishing strain determines a single easy axis, 

 there are only 180° walls. The geometrical situation is considerably 

 more complicated in the general case, since it is necessary to take into 

 account the crystal plane in which the wall itself lies. 



Note in figure 9 that the wall is drawn so that the magnetic 

 moments are always parallel to the plane of the wall. If they were 

 not, there would be a large magnetostatic energy associated with the 

 wall in the same way as that of the single large domain in figure 8. 

 The stringency of this condition varies as the magnetization, and when 

 the magnetization approaches a compensation point it disappears 

 altogether. Thus in gadolinium iron garnet at 14° C, the angular 

 variation of, say, the octahedral iron magnetization is somewhat dif- 

 ferent from that in the yttrium iron garnet at the same temperature. 



The expected domain wall thickness in our samples is somewhat 



