Clustering in iron-doped magnesium oxide
The distribution of iron in single crystals of magnesium oxide has been investigated using electron paramagnetic resonance (EPR), ferrimagnetic resonance (FMR) and reflection electron diffraction (RED). The total iron content of the samples was in the range 100 - 13000 ppm by weight, the crystals being examined in the as-received state, after solution treatment, and following various aging treatments. The solution treatment involved the samples being held at 1400ºC for 24 hours in an oxygen atmosphere, and then quenched to room temperature. The samples were aged in oxygen at temperatures in the range 600 - 800ºC for various lengths of time. The EPR investigations were carried out at 9 GHz, over the temperature range from 4 - 300 K. Comparison of the integrated intensity of the EPR spectral lines with those of a standard suggested that in both the as-received and solution treated crystals very little (in some cases <0.1%) of the iron contributes to the Fe(^3+) isolated ion cubic site spectrum. Analysis of the experimental linewidths and shapes lends support to this suggestion. The experimental linewidths are in all cases broader than is expected on the basis of dipolar broadening theory, and it is suggested that this broadening is partly due to interactions with the undetected fraction of the iron. An unusual broad (1,3 kG wide) line which shows a complex structure at liquid helium temperatures was detected in two of the samples following solution treatment. Its appearance was accompanied by a complete absence of fine structure lines from the spectra in which it appeared. It is probable that the disappearance of the fine structure and the appearance of this broad line are related, but the relationship is not at present clear. Following aging treatments ferrimagnetic resonance was detected at 9 GHz, in the temperature range 4 - 500 K. This resonance arises out of the precipitation of magnesioferrite from the host lattice. In most of the samples which showed evidence of precipitation two quite different FMR lines were found - an isotropic line which appeared after short aging times, and an anisotropic line which apparently replaces the isotropic one as aging progresses. Analysis of the measured anpsotropy field of the particles precipitated at 800ºC indicated that the rate of growth of the volume of the precipitates is linear with time, after about one hour's aging. The chemical formula for these particular precipitates was determined to be Mg(_x)Fe(_3-x)O(_4-(x-1)/2')) where x = 1.29, the fraction of Mg ions on tetrahedral sites being taken to be 0.30. Analysis of the magnetic characteristics of the precipitates generally gave agreement with what published data are available. In addition the measured anisotropy field H(^sp)(_ɑ) in all samples was found to obey the empirical relationship where D = -0.045 + .004 G(^0.5) K(^-1), T is the temperature and C is a sample dependent constant. The widths of the FMR lines obtained from the high concentration samples aged at 800ºC were independent of sample, and after aging for one hour decreased monotonically with continued aging. This change in width is possibly a consequence of 'voids' of host material in the initial precipitates filling with aging. These linewidths were also remarkable in that they showed a marked decrease with increased recording temperature. Since this behaviour is quite at variance with the behaviour of bulk magnesioferrite it is probably a consequence of the superparamagnetic nature of the particles. Reflection electron diffraction studies of the etched surfaces of the crystals showed only one pattern - the standard spinel pattern - when ferrimagnetic precipitates were present, regardless of the type of FMR spectral line produced by the precipitates. This spinel had a lattice parameter almost exactly twice that of the host. Two other RED patterns were recorded from some of the samples - one possibly due to the calcium-stabilised zirconia discussed by Venables, and the other tentatively assigned to an aluminium spinel.