An investigation into the flotation response of sperrylite (PtAs2) by comparative evaluation of crystal structure and bonding atoms

Doctoral Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title
The Bushveld Complex in South Africa is the biggest platinum group elements reserve in the world. One of the major platinum group mineral (PGM) components in this ore is sperrylite (PtAs2), which has been found to be slow floating compared to the other PGMs and often reports to the tailings stream. Sperrylite comprises about ~21% of the platinum groups minerals (PGMs) in the Platreef ore hence, improving its recovery will be of great economic value to the South African PGM industry. The flotation of PGM ores aims to recover at least five minerals using the same reagent suite. These minerals have different physical and chemical properties and only a few studies so far have investigated the floatability of individual PGM minerals. The reasons for the poor floatability of sperrylite are not yet clear although some studies suggest that poor collector adsorption or poor collector choice might be the major cause. Molecular modelling computations have indicated that the platinum atoms on sperrylite working surfaces are not well exposed, posing some steric hindrances to the approaching collector ligands. In addition, adsorption studies, using molecular modelling computations, indicated that the OH⁻ ions had higher binding energy compared to SEX, suggesting that increasing pH may have a detrimental effect on collector adsorption. This study sought to establish the reasons why xanthate collectors, which are used as the main collector used in industry in the flotation of PGM ores, appear not to promote the flotation of sperrylite. Also, novel collectors were suggested based on the donor-acceptor theory with the aim of improving the floatability of sperrylite. It was hypothesized that increasing the electron density around the collector coordinating atoms will result in stronger bonds with sperrylite. The suggested novel collectors were screened using molecular modelling computations carried out by researchers at the University of Limpopo. Collectors that gave higher binding energies were then synthesized at BGRIMM Technology, in China. Tests were carried out to compare the floatability and related phenomena of sperrylite, pyrite (FeS2), arsenopyrite (FeAsS) and cooperite (PtS). These minerals were chosen due to their having either similar crystal structures and/or binding ligands (S, As) on the mineral surface and to ascertain how their floatability compares to that of sperrylite. Pyrite was selected as a cheap proxy for sperrylite since they have the same isometric crystal structure. Also, the effect of its S bonding ligand as opposed to As in sperrylite was investigated. Arsenopyrite was also selected to investigate the effect of a monoclinic crystal structure and the effect of the As and S ligands. Lastly, cooperite was also selected as a standard sulphide PGM mineral with a tetragonal crystal structure that is known to have a good flotation response. These minerals were included in this study in order to investigate how their floatability differs from that of sperrylite based on their crystal structures and bonding atoms. In addition, experiments were carried out to determine the energy and extent of collector adsorption on the minerals and electrochemical interactions between the minerals and standard collectors. XPS and TOF-SIMS analyses were also carried out to determine the surface products formed on sperrylite and arsenopyrite in the presence of synthetic plant water with and without collectors. Novel collectors were tested to establish the molar energy of interaction with sperrylite using isothermal titration calorimetry and this was compared to those determined by molecular modelling computations. The results indicated that the crystal structure and the binding ligands played a key role in the floatability of the selected minerals. Minerals with the isometric crystal structure had poor collectorless recoveries, viz. FeS2 (12.3%) < PtAs2 (14.5%) < FeAsS (85.4%) < PtS (94.7%) and different collector-induced flotation recoveries. However, the recovery of FeS2 (94%) improved to a greater extent when PNBX was used as a collector compared to PtAs2 (26%). Sperrylite and pyrite had poor natural floatabilities under both alkaline and acidic conditions, indicating that their floatability is strongly dependent on their interaction with collectors. The poor recovery of this relatively hydrophilic mineral is due to poor collector adsorption of the collectors at pH 9 which is the usual pH in operating plants. Rest potential tests revealed that sperrylite had the least extent of interaction with the standard collectors compared to the other minerals. Furthermore, sperrylite rest potentials remained unchanged with a change in pH, indicating its resistance to surface alteration. The minerals rest potentials at pH 9.5 followed the order FeS2 (282) > PtS (254) > PtAs2 (233) > FeAsS (177) mV, while there was a great increase for all the minerals except sperrylite at pH 3.5, with the order PtS (523) > FeS2 (407) > FeAsS (333) > PtAs2 (243) mV. Surface alteration is key in the flotation of different minerals as it either determines the natural floatability of the minerals or aids or hinders collector adsorption. Furthermore, XPS results indicated that the collectors were not chemisorbed onto sperrylite, since the chemisorption process requires concurrent oxidation of the collector and the mineral surface. It was further revealed that metal oxides were formed on the sperrylite surface when conditioned in plant water at pH 9. Plant water ions were also found to be detrimental to the recovery of sperrylite. The recovery of sperrylite with PNBX in deionised water (51%) was double that with synthetic plant water at pH 9 (26%), even though the adsorption tests which were carried out in deionised water indicated that only 5.7% of the dosed PNBX was adsorbed onto the mineral surface. Sperrylite might be also very sensitive to any pulp chemical changes. This was shown by the reduced recoveries after adding copper sulphate and after raising the pulp Eh using sodium hypochlorite. Also, the higher recoveries in distilled water can be a pointer that surface contamination could be one of the reasons for the poor recoveries of sperrylite. This study has also indicated that the reactivity of the arsenic bonding atom was dependent on the metal bonded to it (Pt or Fe). The reactivity of arsenopyrite when conditioned in synthetic plant water resulted in the formation of a hydrophobic surface layer that led to its high natural floatability and poor interaction with collectors. In the case of an unreactive surface of sperrylite, the arsenic bonding atom, which can also act as an electron acceptor could be the major reason for the poor collector adsorption at pH 9. This leads to an increase in the number of bonding sites that the sulphur ligands are attracted to (Pt, As) as opposed to pyrite (Fe), thereby reducing the overall collector binding energy with the mineral. Molecular modelling computations, in collaboration with the University of Limpopo, also showed that the sulphur ligand had the ability of bonding with both platinum and arsenic on the sperrylite mineral surface. Nevertheless, pH seemed to play a role in the poor collector adsorption and flotation recoveries of sperrylite at pH 9, however, the moderate recoveries at pH 4, in synthetic plant water, were a sign that this is not the major cause for the poor floatability of sperrylite. Moreover, recovery using PNBX remained unchanged at pH 4 despite the indication by the adsorption test results that 93% of the collector was adsorbed on the mineral surface under acidic conditions. In addition, the isothermal titration calorimetry tests indicated that the binding energy of the hydroxide ions in a high pH solution with sperrylite was lower than that with PNBX and SNDBDTC collectors. This indicates that collectors have the ability to displace the hydroxide ions which could be preferentially adsorbed on the mineral surface at high pH during the conditioning period. This is in contradiction with the molecular modelling computation findings. Ultimately this work has shown that sperrylite is a relatively hydrophilic mineral under both acidic and alkaline conditions and requires the action of collector reagents to effectively recover it. This study also corroborates the suggestion that poor collector adsorption is the major cause for the poor floatability of sperrylite. It is also important to note that all the adsorption tests were carried out in deionised water and that all the adsorbed collectors were only physisorbed onto the mineral. Chemisorption of collectors on minerals is generally key to effectively recover the valuable minerals and a degree of oxidation is also a requirement on the mineral to catalyse the oxidation of collectors. However, this study has shown that sperrylite is resistant to surface alteration, making it difficult to recover it. Even though the mineral has more binding sites as opposed to pyrite, the mineral has proved to be sensitive to any form of contamination in the form of ions present in plant water or any activator ions, such as copper ions, that could be found in the pulp. Hence, the hypothesis which proposed that a ligand with a greater inter-atomic distance between the two chelating atoms compared to that of typical thiol collectors will have a better crystal structure compatibility with PtAs2 as well as the hypothesis which proposed that linking atoms that are less electronegative (S, N), or electron-donating would be superior options were not sustained in the case of sperrylite. However, the hypotheses were sustained in the case of pyrite which is fairly reactive and also consists of a sulphur bonding ligand. Since the collectors were adsorbed to a greater extent under acidic conditions, it is recommended that collectors that have higher binding energies with sperrylite, which are stable under acidic conditions should be designed to improve the recoveries of sperrylite. Moreover, sperrylite was found to be relatively inert under conditions of increasing pH, it is suggested that a complete mapping of floatability as a function Eh and pH be made to determine floatability domains, as is common for base metal sulphide minerals.