Influence of mineralogy and ore texture on pentlandite flotation at the Nkomati nickel mine, South Africa G. Mishra ⇑ , K.S. Viljoen, H. Mouri Department of Geology, University of Johannesburg, P.O. Box 524, Auckland Park 2006, South Africa article info Article history: Available online 3 May 2013 Keywords: Sulphide ore Froth flotation Mineral processing Mineral liberation analyser Extractive metallurgy abstract The influence of ore mineralogy and ore texture on flotation response was studied for 29 samples from the main mineralised zone at Pit 3 of the Nkomati Ni mine, through laboratory scale flotation testing, lab- oratory assay, and mineral liberation analyser examination of the ore and the concentrates. The individ- ual sample flotation responses vary widely in terms of Ni grade, and cumulative Ni recovery. It is demonstrated that this is a complex function of ore mineralogy and ore texture. Chalcopyrite is the first sulphide to float, followed by pentlandite and finally pyrrhotite, in ore samples with dominant chalcopy- rite, or where pentlandite, pyrrhotite and chalcopyrite occur in equal abundance. However in samples with a high ratio of pyrrhotite to pentlandite and chalcopyrite, pyrrhotite floats earlier than expected, reports to concentrate over the entire flotation period, and depress and extend the flotation of pentland- ite over the flotation interval with no clear peak of Ni recovery during flotation. Primary silicates (e.g. olivine and pyroxene) and alteration-related minerals (talc, tremolite and chlorite) are naturally floating, and hence affect the flotation of pentlandite in a similar manner to that of pyrrhotite. The most problem- atic ore at Nkomati in terms of Ni recovery is characterised by fine disseminated and fine bleb- or net- texture sulphides, contain abundant olivine, pyroxene, amphibole, talc and tremolite, and include abun- dant metamorphism-related country rock xenoliths (with calc-silicate minerals such as diopside and tremolite). Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction Froth flotation is a mature technology that has been practiced successfully for more than a century (Yalcin and Kelebek, 2011). It is, however, still not fully understood and remains relatively inefficient (Shean and Cilliers, 2011). Several tools are available for the study of industrial flotation, with laboratory scale flotation testing playing a central role (Evans et al., 2011; Runge, 2010). Curves of grade vs recovery, as well as element recovery as a function of time, are commonly used to rep- resent and analyse flotation performance (Neethling and Cilliers, 2008, 2012; Shean and Cilliers, 2011). Associated mineralogical information was obtained in the past through element balance cal- culations, optical microscopy, and X-ray diffraction. All of these are hampered by poor sensitivity, i.e. minerals present at concentra- tions <5 wt.% are typically not considered, while some minerals are often indirectly monitored through, e.g. MgO assays. The increasing availability of automated mineralogy systems in the last decade (Evans et al., 2011; Fandrich et al., 2007) now allow for very detailed mineralogical assessment of flotation feed, concentrates, and tailings. However few detailed studies (using these modern methods) of the relationship between flotation and the mineralogy of the ore, as well as the flotation sequence of the various minerals reporting to concentrate, are available in the international pub- lished literature. Variations in recovery are to a large extent caused by mineral- ogical and textural changes in the ore (Adams, 2007; Hay, 2010; Hay and Roy, 2010; Lotter, 2011; Malysiak et al., 2002; Mulaba- Bafubiandi and Medupe, 2007; Senior et al., 1994; references therein), resulting in, e.g. the dilution of the bulk concentrate by minerals with positive flotation characteristics (Lotter, 2011). For instance, concentrate dilution by naturally floating silicates such as orthopyroxene, talc and biotite is observed in the processing of mafic ore deposits (Becker et al., 2009; Evans et al., 2011; Lotter et al., 2008; Mani et al., 1997). Furthermore, some sulphide miner- als are known to be fast floating, (e.g. pyrite and chalcopyrite; Becker et al., 2010; Pearse, 2005; references therein) and may require a depressant strategy to prevent concentrate dilution dur- ing the recovery of pentlandite from Ni ores. A depressant strategy may also be necessary to limit the emission of sulphur dioxide from smelters through the rejection of pyrrhotite at the flotation stage (Agar, 1991; Allison and O’Connor, 2011; Kelebek et al., 1996; Miller et al., 2005; Pearse, 2005; Senior et al., 1994, 1995). 0892-6875/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.mineng.2013.04.009 ⇑ Corresponding author. Tel.: +27 829693052. E-mail address: gargigeo@gmail.com (G. Mishra). Minerals Engineering 54 (2013) 63–78 Contents lists available at SciVerse ScienceDirect Minerals Engineering journal homepage: www.elsevier.com/locate/mineng