M.Sc., Faculty of Science, University of the Witwatersrand, 1992

M.Sc. (Geology) / The Witwatersrand Supergroup is host to a number of auriferous and uraniferous conglomeratic reefs, which have been extensively exploited along the Witwatersrand Basin margins. The current study investigates the Vaal Reef, in the Klerksdorp gold field with particular focus on conducting a geometallurgical characterization of the ore which may ultimately enhance the recovery of gold and uranium and our understanding of how the ore responds to processing. Six samples were collected from AngloGold Ashanti’s Moab Khotsong mine and prepared for a chemical and mineralogical deportment study. These samples were milled and crushed down to 80% passing -75μm and processed for head chemistry assays, grading analysis as well as heavy liquid separation analysis as part of the chemical deportment. The samples were also submitted for gold cyanide, acid uranium and diagnostic leach tests.....

A dissertation submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of requirements for the degree of Master of Science. 23 September, 2014, Johannesburg. / The Witwatersrand Basin has been mined for the last 125 years and is still one of the world’s largest producers of gold and has produced over 50 000 tonnes. However, uranium has also been mined as a by-product of gold from the Witwatersrand reefs, and over 150 000 tonnes have been produced. Over the past decades, the origin of this world class gold and uranium deposit has been debated and still remains controversial. Three main hypotheses were developed, and these are the placer, modified placer and hydrothermal models. In this study, the aims are: to evaluate how many generations of uranium-bearing minerals are in the Vaal Reef samples analysed from Great Noligwa, Moab Khotsong and Kopanang mines and to determine which model among the placer, modified place and hydrothermal best supports the emplacement of the uranium-bearing minerals in the reef. The Vaal Reef occurs in the lower parts of the Strathmore Formation of the Johannesburg Subgroup in the Central Rand Group of the Klerksdorp Goldfield in the Witwatersrand Supergroup. The Vaal Reef is split into three facies, namely the C, B and A Facies; the C and A Facies are the most economic facies at the three mines. The C Facies is well developed at Kopanang mine and the A Facies is well developed at both Moab Khotsong and Great Noligwa mines. Geochemical analyses revealed that the C Facies is enriched in uranium, carbon, sulphur and aluminium; this is due to the presence of uraninite, carbonaceous matter, pyrite and sheet silicate minerals, respectively. The A Facies, however, is more enriched in gold and quartz content, although high uranium, carbon and sulphur concentrations are found, they do not exceed the C Facies concentrations. Mineralogical investigations showed that uraninite, brannerite and uraniferous leucoxene are the uranium-bearing minerals present in the Vaal Reef samples. Uraninite is the main mineral and occurs firstly with detrital minerals such as pyrite, zircon and chromite in the quartz matrices; the second occurrence of uraninite is with the carbonaceous matter. Brannerite and uraniferous leucoxene are suggested to be formed from the breakdown of detrital uraninite grains interacting with Ti-rich minerals such as rutile. Unlike uraninite, brannerite and uraniferous leucoxene occur mainly in the C Facies matrix and occur as patchy or irregular-shaped minerals. The uraninite grains associated with the detrital minerals are mainly round in shape with sizes up to ~150 to 200 μm. This association with the detrital minerals suggests that uraninite was deposited together with the detrital minerals at the same time and that they were in hydraulic equilibrium with one another. Therefore, uraninite is also detrital in origin. The second generation of uraninite grains in the carbonaceous matter mainly show replacement and breakdown of uraninite by the latter, in many observations, uraninite grains are penetrated by the carbonaceous matter through cracks and are further fragmented into smaller grains. The sizes of these fragmented grains vary between 5 – 80 μm and have angular shapes, suggesting that they were first rounded and later broken down and replaced by the carbonaceous matter. A four-staged paragenetic sequence of the Vaal Reef samples was developed, and more importantly the paragenesis showed that the carbonaceous matter post-dates the deposition of uraninite. The three-dimensional microfocus X-Ray computed tomography (3D μXCT) was applied to the Vaal Reef samples and the main objectives were to visualise and analyse the uranium-bearing minerals in the Vaal Reef samples for their sizes, shapes and distribution with respect to other mineral components in the samples in 3D. The technique is currently unable to distinguish individual minerals from one another, especially when minerals have similar grey values as a result of close attenuation coefficients, mineral compositions and density. Mineral groups were identified following this similarity, include quartz and sheet silicates as one mineral group, all sulphides as another group and uranium-bearing minerals with gold as a third mineral group. The analysed uraninite with gold mineral group in the matrix, exhibited grains up to 200 μm in size which were round in shape, as observed in 2D mineralogical techniques. These observations support mineralogical observations acquired by conventional mineralogical techniques suggesting that 3D μXCT can be used to complement other mineralogical techniques in obtaining 3D information on minerals. However, 3D μXCT has limitations such as spatial resolution, partial volume effect and overlapping of mineral grey values. It is therefore, suggested that the technique not be used as an independent tool for mineral characterisation, but rather in support of the existing mineralogical techniques. The source area of the uraninite in the Vaal Reef of the Klerksdorp Goldfield is suggested to have been the hydrothermally altered Archaean basement granite bodies of the Witwatersrand Basin hinterland, from the Hartebeesfontein Dome northwest of the goldfield in particular. High UO2/ThO2 ratios, as determined by electron microprobe analyses (EMPA), support the notion that the uraninite grains are not a product of hydrothermal fluids, and furthermore high Pb contents showing that the uraninite grains are older than the age of the Witwatersrand deposition. In conclusion, the emplacement of uranium-bearing minerals in the Vaal Reef samples analysed in this study is best explained by the modified placer model.

No description available.

Thesis (Ph.D. (Physics)) --University of Limpopo, 2013 / The intention of the current study is to investigate structure, ion transport and reactivity of various forms of the cobalt pentlandite, Co9S8, at different temperatures using atomistic simulation methods with the support of electronic structure calculations. The first interatomic potentials of Co9S8 were derived with input data as structure and elastic properties from experiment and electronic structure calculations respectively. The potentials were validated by running energy minimization and molecular dynamics calculations. Structure, elastic properties and phonon spectra were well reproduced, together with the complex high temperature transformations and melting of Co9S8 as deduced from crystal structure, radial distribution functions, density profiles and diffusion coefficients. Amongst the high symmetry surfaces {111}, {101} and {101} atomistic surface energy calculations proposed the {111} surface of Co9S8 as the most stable in agreement with experimental morphologies, and water adsorption energies on the such surfaces which mostly agreed with those from electronic structure calculations. The structural and ion transport variations with temperature were investigated and predicted surface melting at lower temperatures than the bulk. The effects of hydration on the surfaces at low and high temperatures were also studied. The structural and ion transport properties of Co9S8 nanoparticles of varying sizes, covered by high symmetry surfaces {111}, {101} and {100} were predicted using molecular dynamics method based on our derived interatomic potentials. The structural and ion transport properties of Co9S8 nanoparticles of varying sizes, covered by high symmetry surfaces {111}, {101} and {100} were predicted using molecular dynamics method based on our derived interatomic potentials. Generally for {111}, {101} nanoparticles, high temperature transitions were abrupt for smaller nanoparticles and these tended to disintegrate and form voids. However, for larger nanoparticles the transitions were more gradual. Transitions in the {100} bound nanoparticles were less dramatic for all sizes and the formation of voids was reduced at high temperatures. Generally, the melting temperatures of different sizes of nanoparticles increases with the particle size hence approach the bulk limit. The interaction of nanoparticles with water was investigated. / Anglo Platinum, National Research Foundation (South Africa), and The Royal Society (UK)

No description available.

Production of platinum and associated metals is a major source of revenue for South Africa. Significant losses occur in the concentrating stage (10 to 15 per cent) and this research is focused on optimising platinum flotation. Research begins by conducting laboratory batch flotation tests. However, subsequent pilot-plant tests often produce different results. It is believed these differences arise from the artificial nature of laboratory techniques. This project was focused on improving flotation techniques in the laboratory. The largest source of platinum in South Africa is the UG2 reef and two samples of this ore were used for testing: ‘good’ and ‘bad’ ore. These had different characteristics with regards to the recovery of PGMs and the presence of talc. The latter is an unwanted floatable mineral, which must be depressed to prevent excessive recovery. The conventional laboratory test procedure makes use of batch tests in various sizes of flotation cells. The procedure was made more realistic, by using four stages of flotation, rather than just two, to mimic a typical platinum flotation plant. The use of four stages made it possible to separate the fast-floating and slow-floating stages and to control froth conditions accordingly. Attention was also given to the fact that in laboratory tests, water is often added to the ‘cleaner’ stage of flotation, to make up the level. Experiments showed that this dilution, which does not take place in practice, had a significant impact on overall efficiency. A method of measuring frother concentration was developed and used to determine the realistic level of frother in cleaning tests. Tests at these levels of frother concentration showed that significant improvements could be made to plant performance, by making use of a thickener to reduce the frother concentration in the cleaning stages. The improved test procedure was used on both good and bad ores, and the effect of regrinding was also tested. A combined solids recovery of 2 % over both cleaners was targeted for all test work. At this recovery, the regrinding of the bad ore increased the PGM recovery from 67 to 76 per cent at the cost of an additional 8 g/t depressant. An investigation of the effect of frother concentration in the cleaning stage, using good ore, demonstrated that that rejection of chromite could be improved significantly by reducing frother concentration. The tests mimicked the use of a thickener to separate some of the water with a high concentration of frother. Tests conducted on the good ore showed that use of two thickeners, as opposed to none, reduced the Cr2O3 content of the final concentrate from 4.2 to 3.2 per cent for the equivalent concentrate mass and PGM recovery. The depressant requirement was also reduced from 67 to 55 g/t. These tests provided insight on how to improve performance on a platinum flotation plant, particularly when floating the bad ore.

The Okiep Copper District situated in the north-western Cape Province, covers some 3 000 km and is the oldest mining area in the Republic of South Africa. The O'okiep Copper Company Limited commenced production in 1940 with a proven ore reserve of 9 million tons at 2,45 % cu. Production since 1940 and present ore reserves total some 93 million tonnes at 1,08 % Cu. The rocks comprising the Okiep Copper District are of Proterozoic age and have been subdivided into a meta-volcanosedimentary succession, intruded by various sub-horizontally emplaced granitoid intrusions. The various intrusions occurred at different stages relative to the main structural and metamorphic events. The copper deposits are confined to basic rocks which are the youngest major group of intrusives in the District. They occur as swarms of generally easterly-trending, steep northdipping, irregular dyke-like bodies consisting of diorite, anorthosite and norite. The dominant silicate constituents are andesite ranging to labradorite, hypersthene, biotite and phlogopite. Copper sulphides are preferentially associated with the more basic varieties. The copper sulphides are mainly chalcopyrite, bornite and subsidiary chalcocite. The copper content of the basic rocks is erratic ranging over small distances from a mere trace to several percent. Emplacement of the cupriferous basic rocks is predisposed to a large extent by enigmatic structural features locally referred to as steep structures. The most common manifestation of steep structure deformation is typically a narrow antiformal linear feature along which continuity of the country rocks has been interrupted by piercement folding and shearing. In places, pipelike bodies of megabreccia occur along steep structures, and also act as hosts to the basic rock. Areas of steep structure are thus prime exploration targets, due to their close spatial association with the cupriferous basic rocks. Exploration techiques employed in the Okiep Copper District in~ elude regional and detailed geological mapping, geophysical surveys utilizing magnetic, gravimetric and electrical methods, as well as limited application of soil and stream-sediment geochemistry. Final evaluation is by surface and underground diamond drilling. Exploration has to date discovered 18 new mines with individual ore reserves ranging from 200 000 to 37 000 000 tonnes. All are underground operations, and the sub-level open stoping method of mining is standard.

D.Phil. Geology) / This study presents detailed petrographic, mineral-chemical and geochemical characteristics ofmafic intrusions from three iron oremining areas - Thabazimbi, Sishen and Hotazel - in southern Africa In addition, as themafic intrusions at the Thabazimbi, Sishen and Hotazel mines occur in close spatial association with iron and manganese ore, this study addresses the aspect of whether these intrusions have a bearing on the localization of these ores. Precise geochronologic data of these previously undated mafic dykes and sills is presented to classify them into a regional context. particularly in considering whether these dykes and sills are part of known Large Igneous Provinces (LIPs) in sonthem Africa. The Thabazimbi dykes are coarse grained dolerites while the sills are diabases. The dykes are younger than the sills. Composition wise, sills are dominantly basaltic andesites, while the dykes are dominantly hasaltic. Different to the sills, the dykes are characterized byrestriction of olivine, higher HFSE and LREE as well as less prominent negative Bu" anomaly. Geochemical and isotope chemical characteristics of the Thabazimhi dykes and sills are explained in terms of a combined partial melting, followed by fractional crystallization and crustal contamination with differentiation model. with the dykes showing greater crustal assimilation. The petrogenetic characteristics of the Sishen dolerite dykes in many ways resemble both the Colombia River Basalts and the typical Umkondo dolerites, and point to significant crustal contamination, typical of continental tholeiites. Geochemical characteristics of the Sishen dolerites is acconnted by the partial melting followed by fractional crystallization and crustal contamination, but unlike the Thabazimbi dykes and sills crustal assimilation is significant. At Hotazel, the petrographyand geochemistry of 'bostonites' bas been used to define their true composition while at the same time highlighting the presence of a -2 - 3 m thick iron ore unit associated with banded iron formation and manganese ore. Geochemically, the Hotazel 'bostonites' are "basaltic andesites' while textnra1ly, the Hotazel <bostonites' are essentially diabases. Regarding the iron ores that occurs in the Hotazel Formation at the base of the Pre-Mapedi nnconformity, they are composed of various forms of hematite with variable minor chlorite. quartz and carbonates. Iron ore genesis is attributed to supergene processes similar to those that have affected Sishentype iron ore below the Gamagara unconformity on the Maremane Dome. A U-Pb 2046.6±3.4 Ma age has been obtained for the crystallization of a Thabazimbi sill by dating titanite. This is interpreted to indicate that while the mafic/ultramafic component ofthe Bushveld Complex may have been emplaced over a short period, there was later magmatic activity at 2046 Ma, represented by the Thabazimbi sill and other late Bushveld Complex related intrusions such as the Uitkomst Complex at 2044 Ma. The 1044.3±7.5 Ma age obtained for the Sishen dolerites has resulted into two major interpretations. Firstly. the age is interpreted to be the minimum age for Umkondo LIP. The age shows that while large sections of the Umkondo LIP may have been emplaced within a short time interval, emplacement of end members ofthis LIP was in progress at least - 50 Ma later. Secondly. the age of these dolerites shows that they were emplaced subsequent to major iron ore genesis at Sishen. The new age. coupled with the limited extent ofthese intrusions shows that they played a no part in the origin ofthe bulk ofthe iron ore at the mine.

The giant Palaeoproterozoic manganese deposits of the Kalahari manganese field (KMF), Northern Cape Province, South Mrica, have been a world renowned resource of manganese ore for many decades. In recent years, the mineralogical composition, geochemistry and genesis of these deposits have been the objects of many geological investigations, yet their origin remains contentious up to the present day. A characteristic feature of the Kalahari deposits is the intimate association of manganese ore and iron-formation of the Superior-type, in the form of three discrete sedimentary cycles constituting the Hotazel Formation. This striking lithological association is an almost unique feature on a global scale. From that point of view, the present study is effectively the first attempt to shed light on the origin and post-depositional history of the Hotazel succession, using as prime focus the petrographic and geochemical characteristics ofthe host iron-formation. Petrographic and whole-rock geochemical information of iron-formation from the southern parts of the KMF, suggests that the Hotazel iron-formation is almost identical to other iron-formations of the world of similar age and petrological character. The rock exhibits essentially no high-grade metamorphic or low-temperature alteration effects. Mineralogically, it contains abundant chert, magnetite, subordinate amounts of silicate minerals (greenalite, minnesotaite, stilpnomelane) and appreciable concentrations of carbonate constituents in the form of coexisting calcite and ankerite. Such mineralogical composition is indicative of processes occurring in a diagenetic" to burial (up to very low-greenschist facies) metamorphic environment. Bulk-rock geochemical data point towards a simple composition with Si02, total Fe-oxide and CaO being the chief major oxide components. Whole-rock rare-earth element data suggest that the iron-formation precipitated from a water column with chemical signatures comparable to modern, shallow oceanic seawater. The virtual absence of positive Eu anomalies is a feature that compares well with similar data from Neoproterozoic, glaciogenic iron-formations of the Rapitan type, and suggests but only a dilute hydrothermal signal, poten!ially derived from distal submarine volcanic activity. Carbon and oxygen isotope data from iron-formation and Mn-bearing carbonates as well as overlying ferriferous limestone of the Mooidraai Formation, compare well with the literature. The former exhibit variable depletion relative to seawater in terms of both BC and 180, while the latter have signatures comparable to normal marine bicarbonate. Isotopic variations appear to be related to fluctuations in the amount of co-precipitated marine carbonate, in conjunction with processes of coupled organic matter oxidation - FelMn reduction in the diagenetic environment. Oxygen isotope data from quartz-magnetite-calcite triplets suggest that crystallisation took place under open-system conditions, with magnetite being the most susceptible phase in terms of fluid-rock isotopic exchange. Data also suggest that the calcite-magnetite pair may constitute a more reliable geothermometer than the quartz-magnetite one, mainly due to the interlinked diagenetic histories between calcite and magnetite. Iron-formation from the northern parts of the KMF can by categorised into three main classes, namely pristine, altered and oxidised. Pristine iron-formation is identical to the one seen in the southernmost parts of the field. Altered iron-formation corresponds to a carbonate-free derivative of intense oxidation and leaching processes at the expense ofpristine iron-formation, and contains almost exclusively binary quartz-hematite mixtures. The rock appears to have lost essentially its entire pre-existing carbonate-related components (i.e., Ca, Mg, Sr, most Mn and Ba) and displays residual enrichments in elements such as Cr, Th, V, Ni and Pb, which would have behaved as immobile constituents during low-temperature alteration. The low temperature origin of altered iron-formation is supported by oxygen isotope data from quartz-hematite pairs which indicate that isotopically light hematite would have derived from oxidation of magneftte and other ferroussilicate compounds in the presence of a low-temperature meteoric fluid, while quartz would have remained isotopically unchanged. Occasional occurrences of acmite-hematite assemblages suggest localised metasomatic processes related to the action ofNaCI-rich fluids at the expense of altered iron-formation. The conditions of acmite genesis are very poorly constrained due to the very broad stability limits of the mineral in environments ranging from magmatic to surface-related. Oxidised iron-formation constitutes a distinct rock-type and shares common attributes with both the pristine and the altered iron-formation. The rock contains hematite as an important constituent while the amount of magnetite is substantially reduced. With regard to carbonate nlinerals, calcite contents are clearly very low or absent, having being replaced in most instances by a single, Mgenriched, dolomite/ankerite:type species. Oxidised iron-formation contains somewhat higher amounts of iron and reduced amounts of Sr and Ba relative to pristine iron-formation, whereas enrichments in elements such as Ni, Th, Pb, Cr, and V are seen, similar to altered iron-formation. Oxidised iron-formation appears to have originated from processes of dissolution-mobilisationreprecipitation of solutes derived primarily from leaching that produced altered iron-formation. It is proposed that the Hotazel iron-formation and associated manganese deposits were formed as a result of episodic sea-level fluctuations in a stratified depositional environment that gradually evolved into a shallow carbonate platform. A critical parameter in the development of manganese sediment may include regional climatic patterns related to a glacial event (Makganyene diamictite) prior to deposition of the Hotazel strata. This suggestion draws parallels with processes that are believed to have led to the formation of worldwide iron-formations and associated manganese deposits subsequent to Neoproterozoic episodes of glaciation. Submarine volcanism related to the underlying Ongeluk lavas appears to have had very little (if any) metallogenic significance, while evidence for a sudden rise in the oxygen contents of the atmosphere and ambient waters is lacking. With regard to later alteration processes, combination of geological and geochemical data point towards the potential influence of surface weathering prior to deposition of rocks of the unconformably overlying Olifantshoek Supergroup, possibly coupled with fault- and/or thrustcontrolled fluid-flow and leaching of the Hotazel succession during post-Olifantshoek times.