Formation of Monoclinic Fe2TiO5 in the Fe2O3(s)–TiO2(s)– Cl2(g) System

Abstract

In the Fe2O3 ±TiO2 system, different compounds of general formula Fe2Tin-2O2ny1, with n ˆ 3, 4 and 5 are possible. Pseudobrookite (Fe2TiO5) (n ˆ 3), which is thermodynamically stable above 600 8C [1], is the only phase reported to be formed by heating Fe2O3 ±TiO2 mixtures in air for 72 h at 1350 8C [1, 2]. Pseudobrookite has an orthorhombic structure and it was ®rst described by Pauling [3] from X-ray diffraction (XRD) patterns of a natural single crystal. No other compounds with the general formula Fe2Tiny2O2ny1 form by reaction between Fe2O3 and TiO2. For instance, the Fe2Ti2O7 compound (n ˆ 4) has been reported as a metastable intermediate product formed after heating ilmenite (FeTiO3)±pseudorutile (Fe2Ti3O9) mixtures in oxygen at temperatures between 700 and 1000 8C [4]. The compound Fe2Ti2O7 decomposes to Fe2O3 and TiO2 below 800 8C, and to pseudobrookite and TiO2 above 800 8C [4]. Pseudorutile (n ˆ 5) crystallizes in a disordered structure of hexagonal symmetry [5, 6] and exists in nature as an intermediate alteration product of ilmenite in which all the iron has been oxidized to the trivalent state [5]. In the course of further investigations [7±11], the formation of a new polymorph of Fe2TiO5 has been reported. It was synthesized by the reaction of Fe2O3 with TiO2 in the presence of alkaline-earth oxides using the high-temperature solution (HTS) method [7±10]. Its structure was monoclinic, as determined by XRD analysis and MoEssbauer spectroscopy [7, 9]. In the HTS method described in [7] Fe2O3 ±TiO2 ± MO mixtures (M ˆ Ca, Ba or Sr) were mixed with an appropriate amount of ux (PbO±2V2O5), melted at 1350 8C followed by rapid cooling at a rate between 7 and 10 8C miny1. At slow cooling rates, too few or no crystals of monoclinic Fe2TiO5 were observed. In [7±9] it was suggested that either the presence of impurities or the high cooling rate may be stabilizing factors of the monoclinic phase even though the in uence of alkaline-earth oxides was not understood [7]. However, the monoclinic phase Fe2TiO5 was thermodynamically metastable [7, 11] and transformed irreversibly to pseudobrookite, the orthorhombic polymorph, as observed at about 1000 8C [7, 9]. Recently [12] we have reported the formation of pseudobrookite from heating of Fe2O3 ± TiO2 mixtures in a chlorine atmosphere at 850 and 950 8C. In the present work the additional formation of monoclinic Fe2TiO5 from reaction between Fe2O3 and TiO2 in the presence of chlorine gas is reported. Samples were characterized by powder XRD, scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDXS). The starting materials, characterized elsewhere [13, 14], were Fe2O3 haematite (Spex Industries, Inc., USA) and TiO2 (Mallinckrolt Chemical Work) powders with Brunauer±Emmett±Teller areas of 3.0 m gy1 [13] and 9.4 m gy1 [14], respectively. The Fe2O3 powder had a distribution of particles from about 200 to less than 10 im, which in turn were agglomerates composed of very small grains with a mean size of about 0.2 im [13]. Titanium dioxide powder, which initially contained 95% anatase±5% rutile [14], was formed by particles ranging from less than 1 im to about 60 im, composed of small grains which had a mean size of 0.2 im [14]. The procedure was as follows: the powder mixtures were mechanically prepared at the molar ratio 1:1. This mixture was placed in a quartz spherical capsule and outgassed in vacuum (1:3 3 10y5 MPa) for 1 h at room temperature. Chlorine gas (99.8% purity; Indupa, Argentina) was injected into the capsule, which was then sealed to obtain a chlorine pressure of 0.1 MPa at the working temperature. A sample, henceforth called sample 1, was heated for 24 h at 750 8C. In order to compare the effect of chlorine on the oxide mixture reactivity with the same thermal treatment in air, the Fe2O3 ± TiO2 mixture (sample 2) was heated in air in an open quartz crucible for 24 h at 750 8C. After sample 1 had been heated in chlorine, it was divided into two portions, identi®ed as samples 3 and 4, which were heated in the open air for 6 h at 750 8C and 950 8C, respectively. After thermal treatments all the samples were quenched in air, and XRD analysis (Philips Electronic Instruments PW 1710) using Cu Ka Ni-®ltered radiation was performed to identify the crystalline phases. Then the samples were observed by SEM (Philips Electronic Instruments SEM 515) to characterize their morphologies. XRD analysis on sample 2 indicated that after treatment in air no new crystalline phases were formed. In addition, from the SEM observations it was also possible to conclude that microstructural changes did not take place. On the other hand, when the Fe2O3 ±TiO2 mixture was heated in chlorine (sample 1), signi®cant microstructural changes were observed in some particles. Also, several particles