Iron Substitution in Soil a N D Synthetic Anatase

-Iron nI for Ti TM substitution in the structure of pedogenic and synthetic anatase of up to Fe/ (Ti+Fe) 0.1 mol/mol was indicated by an increase in unit cell size as measured by XRD line shifts. MiSssbauerand electron paramagnetic resonance spectra at both, 298 K and 4.2 K supported this by the presence of signals typical for octahedrally coordinated Fe I11 in a diamagnetic matrix. Charge compensation was achieved by structural OH, as indicated by FTIR bands at 3360 and 960 cm -~, which were absent in pure anatase and which disappeared on heating. The weight loss on heating amounted to ca. 0.5 mol H20/mol Fe. At 600~ structural Fe was ejected, the unit cell size decreased to that of pure anatase, and pseudobrookite, Fe2TiOs, was formed. Key Words--Anatase, Crystal chemistry, EPR, Fe substitution, FFIR, M6ssbauer, OH--O-substitution, Pseudobrookite, TA, TEM, XRD. I N T R O D U C T I O N Mean Ti concentrations in various soils range between 1.5 and 29.3 g/kg (Milnes and Fitzpatrick 1989). Ti tanium is found in fairly stable residual minerals such as ilmenite, titano-magnetite and rutile, but also in pedogenic minerals of which anatase has been most frequently identified. Pseudorutile and titano-maghemite are two others. Anatase is considered a weathering product of Ti-bearing, Fe-rich silicates (biotites, pyroboles) and sphene (Milnes and Fitzpatrick 1989). It can be assumed, therefore, that anatase in the weathering environment forms in the presence of an ample supply of Fe. If not sufficient to create Fe-Ti-phases, like pseudorutile, the question arises if Fe III may enter the tetragonal structure of anatase in measurable amounts. To the best of our knowledge, the literature holds no proof of Fe-for-Ti substitution in natural anatase except that Fe has been noted as a trace constituent (Deer et al 1962). Fitzpatrick et al (1978) synthesized Ti-Fe-mixed phases from coprecipitates at 70~ and obtained anatase as the sole phase between Ti / (Ti+Fe) ratios of 0.7 and 1.0. The Fe was no longer oxalate extractable, and the XRD peaks became broader as Fe content increased. On these grounds, structural incorporation of Fe was suggested. A slight shift of the 101 X-ray peak toward higher d-values was indeed visible, but was not noticed by the authors. A similar line shift was recently observed with anatases from deep saprolites of lateritic profiles on basalt of the Jos Plateau, central Nigeria (Zeese et al 1994). Based on these observations, the properties of both natural and synthetic anatases were studied using chemical and thermal analysis, X-ray diffraction, and MiSssbaner-, electron spin resonanceand FTIR-spectroscopy to obtain more conclusive evidence regarding Fe-for-Ti-substitution in this mineral. Copyright 9 1995, The Clay Minerals Society MATERIALS AND METHODS Four saprolite samples (II 1; II 2; II 5; IV 6a) formed from basalts in central Nigeria (Zeese et al 1994) and one (Inl5) from a biotite gneiss from south India (Schwertmann 1984) were examined. Pedogenic anatase concentrates were prepared from the kaoliniticgoethitic-hematitic saprolites by first removing the Fe oxides with DCB (Mehra and Jackson 1960). Kaolinite was subsequently dissolved using first hot 5 M NaOH and then cold 0.5 M HCl to remove sodalite and possible Fe oxides formed from kaolinite-Fe during NaOH treatment. For comparison, five euhedral, dipyramidal, non-pedogenic anatase crystals 1-3 m m in length from various museum collections were analyzed without any pretreatment. A series of synthetic anatases (series 71) was produced by precipitating mixed FeC13-TiC14 solutions with NH3 at pH 5.5 (Fitzpatrick et al 1978) and then treating the precipitates hydrothermally in a 200-ml teflon vessel with 100 ml of water at 180~ for 40 hours in an attempt to improve crystallinity. After treatment, the samples were washed with water and freeze-dried. The Fe/(Ti+ Fe) ratios of the original solutions were 0; 0.025; 0.05; 0.075; 0.10; 0.125 and 0.2 tool/tool, and the samples were designated as 71/0 to 71/20, accordingly. An additional sample for Mrssbauer spectroscopy was prepared with isotopically enriched 57Fe and a Fe/ (Ti+Fe) of 0.002 mol/mol. This sample was produced to test i f trace amounts of Fe in a TiO2 matrix are homogenously distributed or clusters of Fe oxides. To remove any coprecipitated We III oxide in the synthetic samples, 400 mg of each sample were treated twice with 20 ml of 6 M HC1 on a sand bath at ca. 100~ for 1 hour, washed twice with 50 ml of 6 M HC1 and freeze-dried. Total Fe and Ti were determined after dissolving ca. 30 mg of both natural and synthetic samples with 2.5