MAGNETIZATION IN CRYSTALS — DILLON 387 



It turns out that the temperature dependence [5] of the total iron 

 magnetization is different from that of the total rare earth magnetiza- 

 tion. The rare earth magnetization falls off faster with increasing 

 temperature than the iron magnetization. Figure 1 shows how this 

 leads to a magnetization versus temperature curve which changes sign 

 with temperature. Let us say the moment is positive when it is 

 parallel to that of the Fe+ + + on tetrahedral sites. In the case of 

 yttrium iron garnet, the net moment is parallel to that of the three 

 tetrahedrally coordinated Fe+ + + per foniiula unit and thus positive 

 up to the Curie point, the temperature above which the magnetic 

 moment is zero. In gadolinium iron garnet, on the other hand, the 

 moment of the Gd+ + + ions overrides the net iron moment by a con- 

 siderable factor. Plowever, as temperature increases, the Gd+ + + 

 lattice moment falls off faster than that of the Fe+ + + lattice. At 

 about 290° K. they are equal and the net spontaneous magnetic mo- 

 ment is zero. Above this compensation point, the moment of the 

 Fe+ + + lattice predominates. The spontaneous moment, that which 

 would be observed in no applied magnetic field, finally decreases to 

 zero at the Curie point. The Curie points of the f errimagnetic garnets 

 are associated with the exchange coupling of the iron sublattices, and 

 consequently are very nearly the same for all the ferrimagnetic 

 garnets, about 560° K. 



CRYSTALS AND SAMPLES 



Plate 1 is intended to give some idea of the appearance of the 

 crystals with which we work. In bulk they are shiny and black. In 

 the crystals chosen for photographs, the faces are sharp, well de- 

 fined, and very nearly flat. Occasionally a variation of the cooling 

 cycle produces a run of crystals which have been etched in some degree 

 by the same melt from which they were grown. There is a tendency 

 for the crystals to be sound near the faces, but in many cases they 

 contain inclusions of the oxide mixture from which they were grown. 

 Sometimes the crystals are cracked. The shiny surfaces invariably 

 have many small etch pits, which can only be seen on examination with 

 a microscope. 



We might remark in passing that these crystals [6] are made by 

 heating a mixture of iron sesquioxide (FcaOs) , lead oxide (PbO) , and 

 the appropriate yttrimn or rare earth sesquioxide (M2O3) . This mix- 

 ture is heated to about 1,350° C, well above the melting point, then 

 cooled slowly. This is done with the mixture in a platinum crucible 

 with a furnace whose temperature can be programed very accurately. 



The growth of crystals involves many difficulties. There are many 

 variables to be controlled or to slip out of control. Large sound 

 crystals are associated with slow cooling rates, uniform cooling rates. 



