Characterisation and magnetic concentration of an iron ore tailings

The Quadrilátero Ferrífero mineral province is one of the most important sources of iron ore in Brazil. Nowadays this region produces 65% of the Brazilian iron ore. However, since the 1970s, there has been a continuous depletion of high Fe grade ores, obliging the mineral companies to concentrate low Fe grade ores in order to reach the quality demand of the steel mills. It is estimated that 400 kg of tailings is produced for each tonne of beneficiated iron ore. This means there is a very big volume of tailings dumped from industrial plants over the years. Owing to the inefficiency of the industrial operations, the tailings deposited in tailings dam have Fe grades similar to or higher than the Fe grades of some iron ores exploited nowadays (30–45 wt%). Therefore, the reprocessing of this material can recover millions of tonnes of discharged Fe and decrease the volume of the existent tailings, which can be interesting for the environmental and economic sustainability of this region. Herein, we present the characterisation of a tailings sample from a gravity concentration circuit of a mine located in northwest of Quadrilátero Ferrífero with the objective of proposing a concentration route for this material. The main characteristics of this tailings sample were: d80 = 4 mm; grades of 30.3 wt% Fe and 55.4 wt%SiO2, while the identified minerals by XDR, optical microscopy and thermal analyses were quartz, hematite (martitic, lamellar, granular, sinuous, specularitic), goethite (alveolar, amphibolitic, botryoidal), magnetite, kaolinite (5.3 wt%) and apatite (0.7 wt%). Liberation of quartz for size fraction – 0.150 mm was of 80%. Bench magnetic concentration of this material increased the Fe grade up to 23 wt% and decreased the SiO2 grade up to 39.4 wt% in the obtained concentrates, which is satisfactory for the rougher concentration step. Keywords: Iron ore, Tailings, Particle size, Mineralogy, Magnetic concentration

Abstract licence: CC BY-NC

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.