Mz. Dang et al., Interplay of surface conditions, particle size, stoichiometry, cell parameters, and magnetism in synthetic hematite-like materials, HYPER INTER, 117(1-4), 1998, pp. 271-319
We have studied several synthetic hematite-like materials, produced via dif
ferent reactions using various hydrothermal conditions and various temperat
ures of annealing in air, by bulk elemental analysis, weight loss measureme
nts, scanning electron microscopy, powder X-ray diffraction, Mossbauer spec
troscopy, and SQUID magnetometry. We conclude that hematite-like materials
cannot be related to pure stoichiometric hematite via a single stoichiometr
ic or physical parameter and that at least two degrees of freedom are requi
red. This is most clearly seen when we introduce a plot of the cell paramet
er c versus the cell parameter a on which hematite-like materials do not fa
ll on a single line but occupy an entire region that is bounded by hydrohem
atite-hematite and protohematite-hematite lines. A Morin transition boundar
y on this c-a plot separates a region where Morin transitions occur from a
larger region where Morin transitions do not occur down to 4.2 K. Previous
claims that particle size is the dominant factor controlling the Morin tran
sition are understood in terms of correlations between stoichiometry and pa
rticle size that are produced at synthesis. Changing contents of incorporat
ed molecular water and structural hydroxyls with associated cation vacancie
s have different characteristic effects on the crystal structure and move t
he: sample coordinates in different directions on a c-a plot. It is also sh
own that an accessory sulphate content is adsorbed on the individual hemati
te crystallites and is not structurally incorporated. Mossbauer spectroscop
y is used, as usual, to identify and characterize the spin structure. In ad
dition, hyperfine field distributions from room temperature spectra, extrac
ted by a new method, give a sensitive measure of sample conditions but not
a unique one since several factors affect the extracted distributions in si
milar ways.