Inferring Grain Size Distributions from Magnetic Hysteresis in M-type Hexaferrites

TL;DR

This paper introduces a stochastic-dynamic model to infer grain size distributions from magnetic hysteresis data in M-type hexaferrites, providing a non-imaging microstructural characterization method validated by experiments.

Contribution

It presents a novel inverse modeling approach combining stochastic grain growth and magnetic relations to estimate microstructural parameters from hysteresis loops.

Findings

Successful inference of grain size distributions from magnetic data

Validation on synthesized strontium hexaferrite shows interpretable microstructural trajectories

The method offers an alternative to imaging-based microstructural analysis

Abstract

We develop a stochastic-dynamic framework to infer latent grain size distribution from magnetic hysteresis data in M-type hexaferrite materials, offering an alternative to imaging-based characterization. A stochastic nucleation-growth process yields a Modified Lognormal Power-law grain size distribution. This is combined with Brown's relation to obtain a coercivity probability distribution, which is embedded within a dynamic magnetization model. A key feature is the joint estimation of microstructural parameters, including the critical grain radius, through inverse optimization of full hysteresis loops. Experimental validation on hydrothermally synthesized strontium hexaferrite subjected to nitrogen treatment and recalcination reveals interpretable trajectories of nucleation, growth, and structural memory encoded in the magnetic response.