82 



BELL SYSTEM TECHNICAL JOURNAL 



Thus magnetostriction and the magnetic effects of strain are re- 

 ciprocal properties, and result from the same kind of magnetic forces 

 between atoms as those that account for the variation in magnetic 

 properties in dififerent directions in a crystal. Just as the crystal 

 structure determines a direction of easy magnetization in a strain-free 

 cr>^stal, so the strain controls the direction of easy magnetization when 

 the strain is sufficiently great. Figure 18 shows how the domains are 



UNSTRAINED IRON 



IRON IN TENSION 



NICKEL IN TENSION 



CRYSTAL 

 AXES 



DIRECTION 

 OF TENSION 



Fig. 18 — Domains are oriented by crystal forces and by strain; H = 0. 



magnetized parallel to the crystal axes in unstrained iron, and how a 

 sufficiently large tension will orient the magnetization parallel to the 

 direction of tension in iron and at right angles to the direction of tension 

 in nickel. When the stress is as large as 10,000 to 30,000 pounds per 

 square inch, the strain effect begins to predominate over the crystal 

 effect and the direction of magnetization is determined mainly by the 

 strain. The calculations show also that in a material having positive 

 magnetostriction the magnetization is increased by tension. In a 

 qualitative way these considerations explain the increase in perme- 

 ability of 65 permalloy (having positive magnetostriction) and the 

 decrease in 85 permalloy (with negative magnetostriction). But so far 

 the theory is quite inadequate to predict the magnitude of the effect. 

 In addition to uniaxial homogeneous strains, such as those pro- 

 duced by stretching a wire in the direction of its length, random 

 (heterogeneous) strains are often found that vary in magnitude, sign, 

 and direction from point to point throughout a material. Such strains 

 are produced by cold working, phase transformations, and the like. In 

 such materials the direction of magnetization in a domain is deter- 

 mined by the local strain, and is more stable the larger the strain. So it 

 can be appreciated that it is harder to change the magnetization of a 

 material that is more severely hard worked. These internal strains are 

 the same ones that contribute to the hardness of a metal — hence the 

 parallelism between magnetic hardness and mechanical hardness, 

 \\hich is so well known. 



