1050 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1954 



.4 



ments. The Ki at room temperature is —3.8 X 10 . The room tempera- 

 ture value of X domain wall given in Table I and the values at various 

 temperatures given in Table II are obtained from (13) using these data. 

 An independent value of X may be obtained from the ferromagnetic 

 resonance line width. The relation between the observed Hne width 

 2 AH and X has been given elsewhere. Sample shape enters this relation, 

 but not in a critical way, and we therefore ignore it except as it affects 

 the value of the dc magnetic field at resonance, Hres • The relation is: 



X = AHyMs/H,,, . (37) 



22 



From Bickford's data on line width and (37) we find X = 9 X 10 for 

 Fe304 at room temperature. From the ferromagnetic resonance data 

 reported elsewhere on (NiO)o.7o(FeO)o.25Fe203 , the material used in 

 the present research, we find X = 10 X 10' at room temperature. Table I 

 compares the room temperature values of X obtained in the two ways on 

 the two materials. 



The differences between the domain wall experiments and the fer- 

 romagnetic resonance experiments lead to quite different behavior of 

 X in the two cases at low temperatures. These differences can be under- 

 stood in terms of the frequency dependence of X as given by an extension 

 of the relaxation theory given in the third part of the theoretical dis- 

 cussion. A discussion of these relationships must await the detailed 

 report on the ferromagnetic resonance results which is now in prepara- 

 tion, where such an extension will be given. As Table I shows, the room 

 temperature values obtained in the two ways are of the same order of 

 magnitude. 



Let us now turn to a discussion of the relation between (31) and (33) 

 and the data shown in Fig. 8. Qualitatively, of course, the 1/t factor in 

 v/Ho which (31) reveals, taken together with (32) explains most satis- 

 factorily the sharp increase in viscous damping of the domain wall at 

 low temperatures. Furthermore, it seems tjuite possible, although the 

 author has not investigated it, that the higher order terms in (2-1) account 

 for the nonlinearities in Figs. 9 and 10. 



It should be mentioned that an increase in relaxation time at low 

 temperatures which is consistent with (32) has been deduced by Bloem- 

 bergen and Wang and Healy from ferromagnetic resonance data 

 taken by them. 



Quantitatively, we have inadeciuate data for a satisfactory compari- 

 son between Fig. 8 and (31), and the assumptions of the theory should 

 perhaps be investigated further before any such comparison is taken 



