The domain structure of an antiferromagnetic superlattice is studied. Synchrotron Mössbauer and polarized neutron reflectometric maps show micrometer-size primary domain formation as the external field decreases from saturation to remanence. A secondary domain state consisting mainly of at least 1 order of magnitude larger domains is created when a small field along the layer magnetizations induces a bulk-spin-flop transition. The domain-size distribution is reproducibly dependent on the magnetic prehistory. The condition for domain coarsening is shown to be the equilibrium of the external field energy with the anisotropy energy.

No abstract

No abstract

The theory of both transmission and grazing incidence Mössbauer spectroscopy is re-analyzed. Starting with the nuclear susceptibility tensor a common concise first order perturbation formulation is given by introducing the forward scattering amplitude into an anisotropic optical scheme. Formulae of Blume and Kistner as well as those of Andreeva are re-derived for the forward scattering and grazing incidence geometries, respectively. Limitations of several previously intuitively introduced approximations are pointed out. The grazing incidence integral propagation matrices are written in a form built up from 2 × 2 matrix exponentials which is particularly suitable for numerical calculations and practical fitting of both energy domain (conventional source experiment) and time domain (synchrotron radiation experiment) Mössbauer spectra.

Using the general approach of Lax for multiple scattering of waves a 2 × 2 covariant expression for the reflectivity of polarized slow neutrons of a magnetic layer structure of arbitrary complexity is given including polarization effects of the external magnetic field. The present formalism is identical to the earlier published one for the (nuclear) resonant x-ray (Mössbauer) reflectivity and properly takes the effect of the external magnetic field of arbitrary direction on the neutron beam into account. The form of the reflectivity matrix allows for an efficient numerical calculation.