Magnetization Ripple and Arctic Foxes

A model of a polycrystalline magnetic film is considered in which the film is made up of transverse strips of (N) grains, with the magnetization M everywhere parallel within a strip but varying in direction from strip to strip. It is shown that the effective crystalline anisotropy in each strip can...

Full description

Bibliographic Details
Published in:Journal of Applied Physics
Main Authors: Callen, H. B., Coren, R. L., Doyle, W. D.
Format: Article in Journal/Newspaper
Language:English
Published: AIP Publishing 1965
Subjects:
Online Access:http://dx.doi.org/10.1063/1.1714104
https://pubs.aip.org/aip/jap/article-pdf/36/3/1064/10551487/1064_1_online.pdf
Description
Summary:A model of a polycrystalline magnetic film is considered in which the film is made up of transverse strips of (N) grains, with the magnetization M everywhere parallel within a strip but varying in direction from strip to strip. It is shown that the effective crystalline anisotropy in each strip can be represented by a transverse field Ht which varies randomly from strip to strip, where Ht (rms) = K/[2M(2N)½] (K is the crystalline anisotropy constant). It is recalled that random number sequences appear to have a periodicity of about 3.5 units, a phenomenon now recognized as having led to the assignment of a periodicity in the population of arctic foxes. A similar apparent periodicity of the transverse anisotropy field leads to an apparent periodicity of the magnetization ripple. Exchange forces, however, modify the amplitude and wavelength λ of the magnetization ripple when λ<λ0, where λ0=2π (A/Ku)½ is the ``relaxation length'' (A is the exchange constant, Ku is the induced anisotropy constant). When ripple is observed in the electron microscope, an additional instrumental factor must be taken into account. A simple treatment is presented which shows that for grain sizes a<λ0, the angular amplitude of the magnetization ripple θ(rms)≃(πa/λ0)½× {K/[Ku(2N)½]}∼1° and the apparent wavelength λ≃4.6a. Both of these values are in agreement with experiment. When a∼λ0/2π the approximation fails and the linearity between λ and a is destroyed, as observed.