Mechanochemical treatment of high silica bauxite with lime

Minerals Engineering 15 (2002) 211–214 www.elsevier.com/locate/mine Mechanochemical treatment of high silica bauxite with lime P.G. McCormick a a,* , T. Picaro b, P.A.I. Smith a Department of Mechanical and Materials Engineering, Special Research Centre for Advanced Mineral and Materials Processing, University of Western Australia, Nedlands, WA 6907, Australia b Queensland Alumina Ltd., Parsons Point, Gladstone, Qld 4680, Australia Received 1 November 2001; accepted 7 November 2001 Abstract The mechanochemical processing of bauxite with lime has been investigated. The milling of bauxite and lime mixtures in an attritor mill resulted in the formation of an iron rich hydrogarnet phase in bauxite/lime slurries containing greater than 9% CaO. Up to 90% of the quartz contained in the bauxite was found to have been removed by the hydrogarnet reaction. The hydrogarnet compound was stable during the high temperature alumina extraction step, resulting in a 30% drop in caustic soda consumption. Alumina extraction was not influenced by the mechanochemical processing. Ó 2002 Elsevier Science Ltd. All rights reserved. Keywords: Non-ferrous metallic ores; Grinding; Leaching; Extractive metallurgy 1. Introduction The Bayer process accounts for over 90% of the world’s commercial production of alumina. The process involves digestion of bauxite in a caustic soda solution at temperatures ranging from 100 to 250 °C, depending on the form of alumina in the bauxite, e.g. gibbsitic, boehmitic or diasporic. The digestion of bauxite not only dissolves the majority of the alumina in the bauxite but also dissolves silica. There are two main forms of silica which occur in bauxite: (i) reactive silica, mainly as kaolinite ðAl2 O3  2SiO2  2H2 OÞ and (ii) quartz. Kaolinite is readily attacked by caustic liquor to form a sodalite-type desilication product (DSP) with the general formula 3ðNa2 O  Al2 O3  2SiO2  0–2H2 OÞ  2NaX (Whittington, 1996), where X is an anionic species such 2 as OH , Cl , CO2 3 , SO4 . The majority of the DSP is discarded with the red mud, resulting in the loss of caustic soda from the process. Quartz is not readily attacked by caustic liquors at low temperatures, but is increasingly attacked during the high temperature digestion conditions required for boehmite and diaspore dissolution. In this case there is a loss of alumina as well as caustic soda in forming DSP. Each tonne of silica that dissolves from bauxite during digestion consumes approximately 1.2 tonnes of * Corresponding author. Tel.: +61-9-3803122; fax: +61-9-3801024. E-mail address: pgm@mech.uwa.edu.au (P.G. McCormick). soda in forming DSP. It is generally considered uneconomic to treat bauxites containing greater than about 5% reactive silica by the conventional Bayer process because of excessive soda losses (O’Connor, 1988; Grubbs, 1987). Baksa et al. (1986) estimated that soda required for processing high silica bauxites accounts for as much as 20% of alumina production costs. The addition of lime during pre-desilication or digestion can minimise soda losses by forming hydrogarnet (calcium–aluminium–hydrosilicate) or low-soda calcium–cancrinite in which Ca2þ replaces some of the Naþ in the DSP (Baksa et al., 1986). Under standard Bayer conditions hydrogarnets with the formula Ca3 Al2 ðSiO4 Þn ðOHÞ124n , (where n ffi 0:6) are formed (Whittington, 1996). However, hydrogarnet formation results in unwanted consumption of alumina. Each tonne of lime added to the process will consume approximately 0.6 tonnes of Al2 O3 . The production of high iron and silica hydrogarnets has been proposed as a means of reducing both caustic soda and alumina losses (Zoldi et al., 1987). This normally requires higher temperatures and pressures than is used in most Bayer digestion plants, around 260–340 °C, to produce iron substituted hydrogarnets of the general formula Ca3 AlFeðSiO4 ÞðOHÞ8 . Such compounds have a higher silica content than the standard Bayer process hydrogarnet as well as less alumina which has been partly replaced by iron. The amount of iron replacement varies depending on the reaction conditions. 0892-6875/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 8 9 2 - 6 8 7 5 ( 0 1 ) 0 0 2 0 7 - 2 212 P.G. McCormick et al. / Minerals Engineering 15 (2002) 211–214 It is well known that the reactivity of minerals can be improved by ultrafine grinding – a process known as ‘mechanical activation’ (Juhasz and Opoczky, 1990). Mechanochemical reactions during ball milling have been the subject of numerous studies (McCormick, 1995). Chemical reactions have been shown to occur at low temperatures when the reactants are milled together in a ball mill. During milling, the microstructure and chemistry of the powders change due to the occurrence of solid state reactions between the particles. Mechanical milling provides a number of favourable factors to allow these solid state reactions to occur. Milling reduces the particle size which increases the surface area available for reaction. The particles are repeatedly welded together then fractured, thus dynamically maintaining fresh surfaces for reactions to take place. Pawlek et al. (1992) have proposed the mechanochemical processing of bauxite with NaOH as an alternative to the caustic digestion phase of the Bayer process. This paper reports the results of an initial study of the mechanochemical processing of bauxite/lime slurries. It is shown that for sufficiently high lime to bauxite ratios mechanical milling results in the removal of silica from bauxite via the formation of an iron substituted hydrogarnet phase. The hydrogarnet phase is found to be stable under the Bayer digestion conditions, resulting in reduced soda consumption. 2. Experimental procedure The materials used in this study were a high silica bauxite and reagent grade CaO. The composition of the bauxite is given in Table 1. Milling was carried out in a sealed, 1 l capacity horizontal attrition mill. The ball charge of 3 kg of 6 mm hardened steel balls was stirred by a rotor operating at 600 rpm. Four different ratios of bauxite and CaO were milled corresponding to the CaO compositions shown in Table 2. All samples were milled wet, using a bauxite solids concentration of approximately 500 g/l. The bauxite was pre-ground in a ring grinder for 15 s before being slurried with distilled water and added to the mill. The CaO was not pre-mixed with the bauxite, but added to the mill just before milling. All samples were milled for 1 h. Following milling, the slurry was recovered by adding additional water to the mill and slowly milling for a Table 1 Composition of bauxite wt% Al2 O3 Fe2 O3 SiO2 TiO2 H2 O, trace elements 55.1 11.3 5.7 2.6 Balance Table 2 CaO content of samples Sample wt% CaO QA1 QA2 QA3 QA4 4.0 5.9 9.1 14.3 further 2–5 min. The balls and slurry were tipped into a sieve and the mill and balls were further washed with distilled water. The slurry was then filtered to recover the milled bauxite residue. After drying the samples were examined using X-ray diffraction (XRD) to determine qualitatively the various fractions present, and X-ray fluorescence (XRF) to provide a standard chemical analysis which was used to calculate the weight percent of metal oxides in the bauxite. The amount of unreacted quartz in the milled bauxite was measured by a standard quartz analysis method involving acid digestion. The milled bauxite samples were digested at 250 °C for 10 min using the Bayer digestion liquor. After digestion the bombs were quickly cooled and the liquor and red mud separated by centrifuging. Caustic soda and alumina levels in the liquor were measured thermometrically using a Technicon Auto-analyser. Liquor impurities were measured by XRF. The solids (red mud) underwent a double ammonia wash, using 10% w/w ammonia solution, before being analysed by XRF to measure the amount of metal oxides in the mud, and XRD to identify the various compounds present. 3. Results and discussion X-ray diffraction patterns of unmilled bauxite and milled samples QA2 and QA3, corresponding to lime contents of 5.9% and 9.1% CaO, respectively, are shown in Fig. 1. In addition to major peaks associated with the gibbsite and boehmite phases, minor peaks corresponding to hematite, goethite, kaolinite and anatase were also evident on the diffraction pattern. Diffraction peaks associated with quartz could not be unambiguously identified on the diffraction pattern. Comparison of the diffraction pattern of sample QA2 with the unmilled bauxite sample showed that peak broadening and a significant reduction of peak intensities occurred in the milled sample. Apart from, a minor calcite peak, no new phases were detected in QA2 after milling. The diffraction pattern for sample QA1 (4% CaO) was similar to that of QA2. Examination of the diffraction pattern of sample QA3 after milling revealed a set of new peaks, which were indexed to the iron substituted hydrogarnet phase, Ca3 AlFeðSiO4 ÞðOHÞ8 . As shown in Fig. 1 (curve c), P.G. McCormick et al. / Minerals Engineering 15 (2002) 211–214 213 gibbsite and hydrogarnet were the major phases present after milling. Diffraction peaks associated with hematite and goethite were still present at lower relative intensities, however, the peak associated with kaolinite was no longer evident. A similar pattern was exhibited by sample QA4. Measurements of the effect of CaO concentration on the amount of quartz remaining in the milled samples is shown in Fig. 2. Increasing the lime concentration decreased the amount of quartz remaining in the sample to 10% of the original value. It is noted that, based on the stoichiometry of the hydrogarnet phase and the composition of the bauxite, the theoretical conversion of the silica contained in the bauxite (whether as quartz or in kaolinite) should increase from 26% to 100% on increasing the CaO concentration from 4% to 14%. It was not possible to measure the kaolinite content in the samples after milling to determine the amount of kaolinite converted to hydrogarnet. However, the reduction in the relative intensity of the kaolinite peak in sample QA2 and its removal in QA3 and QA4 is indicative of conversion to hydrogarnet samples. As shown in Fig. 1, the main CaCO3 peak was present in the milled samples, indicative of an incomplete reaction. The effect of CaO concentration on soda consumption during digestion is shown in Fig. 3. The soda/silica ratio (moles soda/moles silica in red mud) represents the amount of caustic soda consumed per amount of silica in the red mud. The maximum value of the ratio as determined from the formula for DSP is approximately 0.67. Soda/silica ratios below this figure indicate a reduction in soda consumption. As shown in Fig. 3 the soda/silica decreased significantly with increasing lime charge. To quantify the effect of milling on soda consumption, the soda/silica ratios in Fig. 3 have been compared with soda/silica ratios of unmilled bauxite digested using similar concentrations of CaO. Since different bauxites and liquors were used in the two tests, the soda consumption figures for each set of data are normalised to a base case where each bauxite sample was digested with no lime. Based on these results soda consumption is estimated to have decreased by up to 30% due to mechanochemical treatment. It was also found that the milling had no significant effect on the extraction of alumina. Alumina extraction levels of 92.1–96.6% were obtained in the milled samples as compared to 95.6% in the unmilled bauxite. The present measurements show that the mechanochemical treatment of bauxite with lime has the potential to convert the silica components in bauxite into a non-reactive form, prior to the introduction of caustic soda solution for digestion. Theoretically if all the silica Fig. 2. Effect of milling with CaO on quartz content in bauxite. Fig. 3. Effect of milling with CaO on soda consumption during bauxite digestion. Fig. 1. X-ray diffraction patterns: (a) as received bauxite; (b) QA2; (c) QA3. 214 P.G. McCormick et al. / Minerals Engineering 15 (2002) 211–214 were converted to hydrogarnet this would eliminate soda consumption during bauxite digestion. Increasing the degree of iron substitution in the hydrogarnet compound formed during milling would reduce alumina losses asso-ciated with lime addition in the Bayer process. Further research work is required to optimise the process. Another possible benefit of the mechanochemical treatment is the destruction of organic contaminants (Hall et al., 1996) in the bauxite which cause problems in precipitation. 4. Conclusions The mechanochemical treatment of a bauxite/water slurry with lime resulted in the formation of an iron substituted hydrogarnet for CaO concentrations exceeding 9 wt%. Analysis of the quartz content of the milled bauxite revealed a decrease of up to 90% due to the quartz silica reacting with lime. Digestion tests showed that the iron substituted hydrogarnet was stable under the conditions of digestion and resulted in a decrease in soda consumption. Alumina extraction was not affected by the milling process. The present study shows that the mechanochemical treatment of bauxite with lime has the potential to convert the total silica content in bauxite into a nonreactive form which theoretically would eliminate present soda consumption problems. Further work is required to fully understand the reaction chemistry and optimise milling and other process variables. Acknowledgements The authors would like to thank the management of Queensland Alumina Ltd. for their permission to publish this paper. The assistance of Darwin Del Aguila in carrying out the bauxite digestion tests is gratefully acknowledged. References Baksa, G., Vallo, F., Sitkei, F., Zoldi, J., Solymar, K., 1986. Complex causticisation: an effective means for the reduction of NaOH losses in an alumina plant. Light Met. TMS, Warrendale, 75–80. Grubbs, D.K., 1987. Reduction of fixed soda losses in the Bayer process by low temperature processing of high silica bauxites. Light Met. TMS, Warrendale, 19–25. Hall, A.K., Harrowfield, J.M., Hart, R.J., McCormick, P.G., 1996. Mechanochemical reaction of DDT with calcium oxide. Envir. Sci. Tech. 30 (12), 3401–3407. Juhasz, A.Z., Opoczky, L., 1990. Mechanical Activation of Minerals by Grinding. Ellis Horwood, Chichester, UK. McCormick, P.G., 1995. Application of mechanical alloying to chemical refining. Mater. Trans. JIM 36, 161–169. O’Connor, D.J., 1988. Alumina extraction from non-bauxitic materials. Aluminium-Verlag GmbH, D€ usseldorf. Pawlek, F., Kheiri, M.J., Kammel, R., 1992. The leaching of bauxite during mechano-chemical treatment. Light Met. TMS, Warrendale, 91–95. Whittington, B.I., 1996. Quantification and characterisation of hydrogarnet and cancrinite present in desilication product (DSP) by powder X-ray diffraction. In: Fourth International Alumina Quality Workshop, Darwin, pp. 414–424. Zoldi, J., Solymar, K., Zambo, J., Jonas, K., 1987. Iron hydrogarnets in the Bayer process. Light Met. TMS, Warrendale, 105–111.