Pyrrhotite (Fe(1-x)S) is one of the most commonly occurring metal sulfide minerals and is recognised in a variety of types of ore deposits. Since the principal nickel ore mineral, pentlandite, almost ubiquitously occurs with pyrrhotite, the understanding of the behaviour of pyrrhotite during flotation is of fundamental interest. For many nickel processing operations, pyrrhotite is rejected to the tailings in order to control circuit throughput and concentrate grade and thereby reduce excess sulfur dioxide smelter emissions. For the platinum group element processing operations however, pyrrhotite recovery is targeted due to its association with the platinum group elements and minerals. Therefore, the ability to be able to manipulate pyrrhotite flotation performance is of importance. It can be best achieved if the mineralogical characteristics of the pyrrhotite being processed are known and their relationship to flotation performance is understood. Pyrrhotite is known to naturally occur in different forms that have varying physical and chemical attributes. These different pyrrhotite forms are commonly known as magnetic (Fe7S8) and non-magnetic pyrrhotite (Fe9S10, Fe10S11, Fe11S12) and as a result of their varying properties are expected to show some difference in their reactivity towards oxidation and flotation performance. Yet the accounts in the literature are inconsistent as to which of the pyrrhotite types is more reactive. Similarly, there appears to be little agreement in the literature as to which of the pyrrhotite types is more floatable. It is probable that this lack of agreement arises from the fact that previous studies have not given due consideration to the effect of the mineralogy of the samples examined. The success of the discipline of process mineralogy as a whole however, has been to gain an understanding of how the mineralogy of an ore affects its processing properties. The objective of this process mineralogy study was to develop the relationship between pyrrhotite mineralogy and flotation performance based on a thorough characterisation of pyrrhotite from selected nickel and platinum group element ore deposits in terms of their crystallography, mineral association, mineral chemistry and mineral reactivity. This was achieved through the characterisation of the mineralogy and mineral reactivity of pyrrhotite samples obtained from the Sudbury ore in Canada, Phoenix ore in Botswana and the Merensky Reef and Nkomati ores in South Africa. Based on the linkage of these characteristics to flotation performance, an understanding of the relationship and mechanisms that cause pyrrhotite mineralogy to influence pyrrhotite flotation performance has been gained. Mineralogical characterisation of the pyrrhotite samples in this study was performed using ore petrography, x-ray diffraction and mineral chemistry analysis. On the basis of these results pyrrhotite samples were classified as: single phase magnetic 4C Fe7S8 pyrrhotite, single phase non-magnetic 5C Fe9S10 pyrrhotite; two phase magnetic 4C Fe7S8 pyrrhotite intergrown with non-magnetic 5C Fe9S10 pyrrhotite and as two phase non-magnetic 6C Fe11S12 pyrrhotite intergrown with 2C FeS troilite. Nickel was identified as the main trace element impurity in the pyrrhotite structure and the amount of solid solution nickel in the pyrrhotite structure was correlated with whether the pyrrhotite was magnetic or non-magnetic, and whether it coexisted with another pyrrhotite phase. All pyrrhotite samples investigated showed a strong association to pentlandite that occurred in both granular and flame pentlandite forms. These key features of pyrrhotite mineralogy were in turn shown to be controlled by the bulk composition and cooling history of the monosulfide solid solution (MSS) from which pyrrhotite is derived. The reactivity of the different pyrrhotite samples towards oxidation was determined using open circuit potential, cyclic voltammetry and oxygen uptake measurements at both pH 7 and 10. Non-magnetic Sudbury Copper Cliff North pyrrhotite was the most unreactive of the samples examined, whereas magnetic Sudbury Gertrude West pyrrhotite was the most reactive. The magnetic Sudbury Gertrude West pyrrhotite was so reactive that open circuit potential and oxygen uptake measurements showed it was already passivated and likely covered with hydrophilic ferric hydroxides. The magnetic Phoenix pyrrhotite was slightly less reactive than the magnetic Sudbury Gertrude West pyrrhotite. The reactivity of the Nkomati Massive Sulfide Body (MSB) mixed pyrrhotite was in between that of the non-magnetic Sudbury Copper Cliff North and magnetic Phoenix pyrrhotite, due to the combined contribution of intergrown magnetic and non-magnetic pyrrhotite to its reactivity. The flotation performance of the different pyrrhotite samples was investigated at both pH 7 and 10 using microflotation tests. A variety of different reagent conditions was also investigated that included the use of different chain length xanthate collectors (sodium isobutyl xanthate (SIBX), sodium normal propyl xanthate (SNPX)) and the use of copper activation. The collectorless flotation of the non-magnetic Sudbury Copper Cliff North pyrrhotite was the greatest of the samples investigated. Only with the addition of flotation reagents were differences in the floatability of the other pyrrhotite samples identified. Magnetic Phoenix pyrrhotite showed good flotation performance whereas the flotation performance of the magnetic Sudbury Gertrude and Gertrude West pyrrhotite was very poor. The Nkomati MSB mixed pyrrhotite only showed good flotation performance at pH 7. All pyrrhotite samples generally showed improved flotation performance with the use of the longer chain length SIBX collector than the shorter chain length SNPX, whereas the efficiency of copper activation was influenced by pyrrhotite mineralogy, pH and collector chain length. Differences in the flotation performance of the pyrrhotite samples investigated were linked to their reactivity towards oxidation. Although not directly measured, the formation of hydrophilic ferric hydroxides on pyrrhotite surfaces due to oxidation was inferred as the reason for the poor flotation performance of some of the pyrrhotite samples. Key features interpreted to influence both pyrrhotite reactivity and flotation performance were pyrrhotite crystallography, mineral chemistry and mineral association. It has been proposed that differences in the amount of vacancies in the pyrrhotite crystal structure influence the oxidation rate and similarly the greater proportion of ferric iron in the magnetic pyrrhotite structure was argued to account for its greater reactivity relative to non-magnetic pyrrhotite. Differences in the solid solution nickel content and trace oxygen in the pyrrhotite structure were also proposed as additional characteristics influencing pyrrhotite oxidation rate and flotation performance. Depending on the degree of association of pyrrhotite to pentlandite, its flotation performance could be affected by the liberation characteristics and flotation of composite particles containing abundant locked flame pentlandite, although this could be manipulated by changing the grind size. The presence of nickel ions derived from the flame pentlandite in these composite particles could also assist in the activation of pyrrhotite and further improvement of its flotation performance. Some guidelines are also presented as to which simple mineralogical and mineral reactivity measurements have been of the most use in developing the relationship between mineralogy and flotation performance. / Thesis (DPhil)--University of Pretoria, 2009. / Materials Science and Metallurgical Engineering / unrestricted
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:up/oai:repository.up.ac.za:2263/28230 |
| Date | 27 September 2009 |
| Creators | Becker, Megan |
| Contributors | Prof J de Villiers, Prof D Bradshaw, megan.becker@uct.ac.za |
| Source Sets | South African National ETD Portal |
| Detected Language | English |
| Type | Thesis |
| Rights | © 2009, University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. |
Page generated in 0.0023 seconds