On the Theory of Spin Waves in Ferromagnetic Media

TLDR

Spin‑wave theory explains the Bloch T^{3/2} law for saturation magnetization in ferromagnetic insulators and metals and relates it to Bloch interdomain wall energy. A field‑theoretical treatment of ferromagnets as continuous media with spin‑density vector fields yields long‑wavelength spin‑wave properties, including anisotropy and magnetic force effects, and reproduces the Holstein‑Primakoff field dependence of saturation magnetization. The study shows that spin‑wave theory is more generally valid than the Heitler‑London‑Heisenberg model, applies to both insulators and metals, and, when extended to include missing spin‑wave states, predicts T^{3/2} magnetization reversal and linear‑T specific heat, while also providing insights into exchange‑Ising relations, exchange integral limits, and spin waves in rhombic lattices.

Abstract

The theory of spin waves, leading to the Bloch ${T}^{\frac{3}{2}}$ law for the temperature variation of saturation magnetization, is discussed for ferromagnetic insulators and metals, with emphasis on its relation to the theory of the energy of the Bloch interdomain wall. The analysis indicates that spin-wave theory is of more general validity than the Heitler-London-Heisenberg model from which it was originally derived. Many properties of spin waves of long wavelength can be derived without specialized assumptions, by a field-theoretical treatment of the ferromagnetic material as a continuous medium in which the densities of the three components of spin are regarded as amplitudes of a quantized vector field. As applications, the effects of anisotropy energy and magnetic forces are calculated; and it is shown that the Holstein-Primakoff result for the field dependence of the saturation magnetization can be derived in an elementary manner. An examination of the conditions for validity of the field theory indicates that it should be valid for insulators, and probably also for metals, independently of any simplifying assumptions. The connection with the itinerant electron model of a metal is discussed; it appears that this model is incomplete in that it omits certain spin wave states which can be proved to exist, and that when these are included, it will yield both a magnetization reversal proportional to ${T}^{\frac{3}{2}}$ and a specific heat proportional to $T$. Incidental results include some insight into the relation between the exchange and Ising models for a two-dimensional lattice, an upper limit to the effective exchange integral, and a treatment of spin waves in rhombic lattices.