A spin-relaxation model was used to interpret the $^{57}\mathrm{Fe}$ M\ossbauer spectra of the ferric oxide spin glass ${\mathrm{BaSn}}_{\mathit{x}}$${\mathrm{Ti}}_{2\mathrm{\ensuremath{-}}\mathit{x}}$${\mathrm{Fe}}_{4}$${\mathrm{O}}_{11}$, with x=0.0, 0.4, 0.8, 1.2, 1.6, and 2.0. This interpretation is suggested as an alternative to distributions of static hyperfine nuclear fields, which require many free parameters to account for the temperature dependence of the spectra, and do not have an obvious physical interpretation. The spin-relaxation analysis used here reduces to the number of free parameters to essentially two: a spin-freezing molecular field and an electronic spin-flip frequency. The results verify the existence of a double magnetic transition in the sample with x=0.0, and the suppression of the upper transition in the samples with x=0.4 and 0.8. The results also suggest the presence of a double magnetic transition in the samples with xg1.0 which has not been observed using other measurements, and may be related to a known difference in structure between the x=0.0 and 2.0 samples. The least-squares fits obtained rival the results of distributions of static fields while requiring fewer free parameters that have direct physical significance. It is suggested that many analyses of spin-glass materials in terms of static-hyperfine-field distributions should be reconsidered.
Read full abstract