The formation of corundum and aluminous hematite by the thermal dehydroxylation of aluminous goethite

Abstract

Dehydroxylation of synthetic and natural goethites with a range of AIsubstitution from 0-28 mole~o was investigated with a view to predicting the behaviour of soil goethites heated by bush fires. Hematites formed at temperatures < 500~ retain the initial A1-content of the precursor goethite up to a maximum of 28 mole% A1. Loss of A1 from the hematite structure occurred at 700~ for synthetic hematites with levels of substitution > 18 mole% A1, but no crystalline alumina phase was present. Crystallization of corundum occurred for synthetic Al-goethites with levels of substitution > 18 mole~o when heated at 900~ Formation of corundum reduces the maximum level of Al-substitution in hematites to ~ 12 mole~o. The exsolved corundum occurs as aggregated, platy crystals, 40-45 nm in diameter, containing a maximum of 7 mole% Fe. Al-substituted maghemite (7 mole% A1) formed from high AIgoethites heated at 900~ Although some corundum in soils may be produced by heating of aluminous goethite by fires, the absence of corundum in natural Al-goethites calcined at 900~ and the very high temperature (900~ at which corundum formed from synthetic goethite suggest that other sources of AI may be required as precursors for the corundum formed by heating of soils.Corundum is a common accessory mineral of silica-undersaturated, high-grade igneous and metamorphic rocks and thermally metamorphosed sediments (Deer et al., 1962), but it would not be expected to crystallize during low-temperature metamorphism. Corundum can be formed in the laboratory by the high temperature (~ 1200~ calcination of the aluminium tri-hydroxides (Rooksby, 1961), and by the topotactic transformation of diaspore (ct-AIOOH) at 450-580~ (Brown, 1980

Comparative mineralogical study of thermally-altered coal-dump waste, natural rocks and the products of laboratory heating experiments

“…Hematite has been noted as a main component in a mixture containing maghemite at 680°C (Pan et al, 2000); this corresponds well with our observations on sample NRPi-C1. Though the Wells et al (1989) report that Al-doped maghemite may crystallize at 900°C concerns post-goethite maghemite, it too seems to correspond well with our observations concerning maghemite as a hightemperature species (samples NRPi-C1 and 60). As the same authors found that hematite is produced from goethite at 350°C, it might be expected that hematite form in pyrite-rich sample 46 at temperatures as low as 400°C, especially in the light of pyrite decomposition starting at 350°C (Querol et al, 1994).…”

Properties and Acid Dissolution of Metal-Substituted Hematites

“…Unit-cell dimensions of Al-rich hematites decreased with increasing A1 substitution [a = 0.5029 -1.42 • 10 3 (mole % A1), r 2 = 0.990***; c = 1.3746 -2.97 • 10 3 (mole % AI), r 2 = 0.81"*, Table 1, where ** represents a 95% confidence level, and *** indicates a 99% confidence level], which is consistent with incorporation of the smaller A13 § ion ( Table 2). Contraction of unitcell parameters is less than that predicted by the Vegard relationship, which is consistent with reported data for synthetic Al-substituted hematites (Schwertmann et al, 1979;DeGrave et al, 1988;Wells et al, 1989). This difference is related partly to the presence of structural defects and to incorporation of structural H20 (Table 1), with OH replacing 02-ions (Stanjek and Schwertmann, 1992).…”

Dehydroxylation of Aluminous Goethite: Unit Cell Dimensions, Crystal Size and Surface Area

“…However, the value of slope for the a dimension was similar to that for Al-hematite synthesized via Al-ferrihydrite precursor (-1.5 x 10 -3) and for Al-hematite formed by the dehydroxylation of Al-goethite (-1.6 x 10 -3) (DeGrave et al 1982). Wells et al (1989) obtained a slope value of -5.7 x 10 -3 for the c dimension of hematite formed by the dehydroxylation of synthetic goethite at 350~ which was larger than the value of -3.5 x 10 -3 at 260-270~ for this study but both values were less than the slope value of the Vegard line derived from JCPDS data (-7.6 x 10-3). Schwertmann et al (1979) and DeGrave et al (1982) found no systematic trend and reported a scattering of values for the slope for the c dimension, while Stanjek and Schwertmann (1992) reported that the regression lines for the c dimension versus mole % A1 varied irregularly as affected by synthesis temperature.…”