The N.N.W.-S.S.E. striking Rustenburg Fault zone, in the western Transvaal Basin, South Africa, has been mapped, in order to unravel its tectonic history. Thickness differences in the Daspoort Formation of the Pretoria Group on opposite sides of the Fault suggest that the Fault was active during Pretoria Group sedimentation, with normal faulting producing localised second-order basins on the down-thrown side of the Fault. In post-Pretoria Group times, but before the intrusion of the Bushveld Complex at -2050 Ma, the area surrounding the Fault zone underwent two compressive events. The first was directed N.E.S. W., producing S.E.-N.W. trending folds, and the second was directed N.W.-S.E., producing N.E.-S.W. trending folds. The second set of folds refolded the first set to form typical transitional Type 1-Type 2 interference folding, and this compression ultimately caused reactivation of the Rustenburg Fault, so that dextral strike-slip movement displaced the Pretoria Group sediments by up to 10.6 km. The subsequent intrusion of the Bushveld Complex into the adjacent strata intensely recrystallised, and often assimilated, the strata along the Fault zone. The fault rocks within the Fault zone were also recrystallised, destroying any pre-existing tectonic fabric. Locally, the Fault zone has been assimilated by the Bushveld Complex. After the intrusion of the Bushveld Complex, little movement has occurred along the Fault, especially where the Fault passes under areas occupied by the Bushveld Complex. It is thought that the crystallisation of the Bushveld Complex has rheologically strengthened the neighbouring strata, preventing them from being refaulted. This model presented above is at variance with previous assumptions that continuous regional extension during Pretoria Group sedimentation culminated in the intrusion of the Bushveld Complex. / Dissertation (MSc)--University of Pretoria, 1997. / gm2015 / Mining Engineering / MSc / Unrestricted

The 2.06 to 2.054 Ga Bushveld Igneous Complex intruded into the sedimentary rocks of the Transvaal Supergroup and generated an extensive contact metamorphic aureole mainly developed in the upper Pretoria group. The studied samples represent the Silverton Daspoort and Timeball Hill formations and are divisible into garnet bearing hornfels (DY918, DY954 and DY956) and garnet-free staurolite-bearing metapelites (DY916, DY982 and DY987). The garnet-bearing hornfelses marks the garnet zone within the aureole and the garnet formation is controlled by different reactions forming from 490 to 630 0C. On the other hand, the garnet free staurolite-bearing Fe-Al rich metapelites define the staurolite zone restricted to the Timeball Hill formation. The recorded P-T conditions in G0 and G1 garnets of the DY954 hornfels imply that the two garnets formed under different conditions indicating two stages of metamorphism. However, the Lu-Hf isotope systematics of these garnets records a 2061 Ma age for all garnet porphyroblasts in both the DY918 and DY954 hornfelses, which support co-genetic garnet growth regardless of their stratigraphic positions. Therefore, the 2061 Ma garnet age denote the emplacement age of the Lower Zone and Critical Zone magmas which was synchronous with the extrusion of the Rooiberg Group volcanics. The fact that all analysed garnets do not record the 2059 – 2054 intrusion of the Main Zone and Upper Zone magmas probably means that the crystallisation temperatures of the later magma pulse was not significant enough to shift the Lu-Hf isotopic signatures. Euhedral staurolites are widespread within the Fe-Al rich metapelites with grain sizes of up 4mm; texturally the majority of them have been altered or overgrown by biotite and chloritoid. The alteration or of these staurolite porphyroblasts is due to isobaric cooling during uplift, and the St-Bt assemblage represent the peak equilibrium conditions and marks the upper stability limit of the Chl-Ctd assemblage. / Dissertation (MSc)--University of Pretoria, 2013. / gm2014 / Geology / unrestricted

Thesis (MSc)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: It is widely accepted that sulphide is the carrier and concentrator of PGEs during magmatic mineralization episodes in the Merensky Reef (MR). PGE concentration peaks and sulphide volume percent peaks are very closely correlated. Koegelenberg, (2011), showed in an experimental investigation that sulphide movement through a cumulate silicate and cumulate oxide pile behave in such a way that sulphide melt gets trapped in chromitite layers. When looking at the compositional distribution of sulphide within the MR it is noted that not only does the sulphide volume percent varies with MR stratigraphy but also the sulphide composition. Sulphide composition is more Cu-rich in the chromitite layers and more Fe and Ni dominated in the hanging wall to the chromitite layers. Until now the more Cu-rich assemblage of the chromitite layers are accepted to be of a sulphide melt composition compared to the Fe and Ni dominated Monosulphide Solid Solution or MSS composition in the hanging wall. In this study we used an experimental approach with a sulphide starting composition thought to exist as the parental sulphide composition of the MR to investigate the phase relations with changing temperature. It is found that the sulphide composition in the chromitite layers represent a sulphide melt composition at 1000 ± 50ºC. At 1000ºC, 50% of the sulphide system would exist as a melt. This Cu-rich melt would have segregated from the MSS and be trapped in the chromitite layer. Also at 1000ºC the partitioning of the Pt would have induced a secondary enrichment step of the Pt concentration in melt through the partitioning of Pt between a sulphide melt and a sulphide solid phase. The experimental evidence in this study points towards a possible source for the parental sulphide magma to the MR, which could have been a slightly Cu enriched mantle sulphide composition. Also, the secondary enrichment of Pt through sulphide melt fractionation at 1000ºC plays an important role in the shaping of the ore body. / AFRIKAANSE OPSOMMING: Dit word wydliks aanvaar dat die sulfied fraksie van die Merensky Rif (MR) die draer en die konsentrasie agent is vir Platinum Groep Elemente (PGE`s) gedurende mineralisasie episodes. PGE konsentrasie en sulfied volume persentasie is op `n hoogtepunt by gelyke stratigrafiese posisies in the MR. Koegelenberg, (2011), het deur middel van eksperimente bewys dat `n sulfied smelt deur `n voorafbestaande kumulaat laag kan beweeg en dat veranderende fisiese eienskappe tussen sulfied smelt en silikaat kristal en sulfied smelt en chromiet kristal, die sulfied smelt sal opsuig en verhoud om verder deur te suipel. Dit is egter oplettend dat nie net die sulfied volume persentasie varieer as `n funksie van die MR stratigrafie nie, maar ook die sulfied samestelling. Die meer Cu-ryke sulfied samestelling in die chromiet lae word aanvaar as `n sulfied smelt fraksie en die meer Fe en Ni dominerende sulfied samestelling in die oorhangende wandgesteentes verteenwoordig die Monosulfied Vaste Oplossing (MVO) soliede fase. In hierdie studie maak ons gebruik van eksperimentele petrologie tesame met `n begin samestelling verteenwoordigend van die oorsprong sulfied samestelling van die MR, om die fase verwantskappe van hierdie spesifieke samestelling te ondersoek. Dit word gevind dat die fraksionering tydens die vorming van die MR plaasgevind het by ongeveer 1000 ±50 C. By hierdie temperatuur is 50% van die sisteem teenwoordig as `n smelt fase. Hierdie Cu-verykte smelt was daartoe instaat om deur die silikaat laag te suipel, geskei te raak van die Fe en Ni dominerende MVO en vasgevang te word in die chromiet lae. Hierdie fraksionering van die sulfied smelt het ook `n sekondêre effek gehad op die verspreiding van Pt tussen sulfied smelt en sulfied soliede fases. Hierdie eksperimentele bewyse dui eerstens op die moontlikheid van `n sulfied smelt in die MR wat sy oorsprong vanuit `n effense Cu-verykte mantel bron kan hê, en tweedens op die belangrikheid van `n sekondêre proses vir Pt re-distribusie tydens die vorming van die MR.

Student Number : 0201438A - PhD thesis - School of Geosciences - Faculty of Science / This project focuses on the Upper Critical and Lower Main Zones in the eastern Bushveld Complex, South Africa. Lithological and stratigraphic information show that there are distinct differences at this level between the eastern and western limbs of the complex. Geochemical studies are centred on the Merensky and Bastard Cyclic Units in which the platiniferous Merensky Reef occurs. A major geochemical hiatus occurs in the Bushveld Complex at the level of the platiniferous Merensky Reef, close to the Critical/Main Zone boundary. The origin of this hiatus and its relation to mineralisation has not been fully resolved. Geochemical parameters are investigated that allow minerals in the Merensky and Bastard Cyclic Units to be classified as originating from either Critical or Main Zone magmas. Modelling of element ratios (Ni/Y, Cr/Ni, Cr/Co, Y/Co, Cr/V, Co/V and Cr/MgO) demonstrates the varying reliability of using ratios as geochemcial tools to constrain magma influxes within a chamber. However, it is shown that the Cr/MgO ratio is effective in determining real differences across the Critical/Main Zone boundary that are independent of lithology. In addition, initial Sr isotope ratios for plagioclase are significantly different in Critical and Main Zone rocks. Geochemical data through the Merensky and Bastard Cyclic Units indicate that orthopyroxene that originated from magma with composition like that of the Critical Zone magma sometimes occurs together with plagioclase that originated from Main Zone magma. In detail, in the pyroxenite at the base of the Merensky Unit, both plagioclase and orthopyroxene display Critical Zone signatures, but in the overlying part of the Merensky Cyclic Unit, plagioclase increasingly shows a Main Zone signature, whereas orthopyroxene continues to display a Critical Zone signature. Similarly, in the Bastard pyroxenite, Sr isotopes and absolute Sr in plagioclase display a range of values from Main Zone to Critical Zone, but orthopyroxene consistently displays Critical Zone affinity. These observations of mineral disequilibrium clearly show that the two major minerals in the Merensky and Bastard Cyclic Units were formed from two different, but coexisting, magmas. A model that accounts for this disequilibrium is proposed here. It invokes the influx of Main Zone magma at the level of the base of the Merensky unit that dispalced the Critical Zone magma upward, but the two magmas did not mix. The latter continued to crystallise orthopyroxene which sank through the Main Zone influx, due to its density contrast. These crystals collected on the crystal pile to form the Merensky pyroxenite. The Main Zone magma, into which the cumulus Critical Zone orthopyroxene accumulated, crystallised interstitial plagioclase that had a Main Zone Sr isotopic ratio. Whole-rock, major element geochemical data show that a variable proportion of the plagioclase in both the Merensky and Bastard pyroxenites is cumulus. It is inferred to have accumulated with orthopyroxene and has a Critical Zone initial Sr isotope ratio. Thus the two pyroxenites now yield a mixed Sr isotopic signature of Critical Zone cumulus and Main Zone intercumulus and possibly cumulus plagioclase that varies along strike. Above the two pyroxenites, the Sr signature of the norites and anorthosites of both cyclic units is dominated by cumulus plagioclase from the Main Zone magma. It is concluded that the variations in initial Sr isotope ratios do not result from mixing of magmas, but result from accumulation of orthopyroxene and plagioclase from a higher, isotopically distinct layer of magma into an underlying layer. The Merensky and Bastard Cyclic Units therefore display features of Critical or Main Zone magma characteristics depending upon which chemical parameter is considered. These cycles are therefore classified as a Transitional Unit.

Evidence of mineral disequilibrium is presented for the Merensky Reef at Winnaarshoek in the eastern Bushveld Complex. Petrographic disequilibrium textures, disequilibrium in orthopyroxene, plagioclase and clinopyroxene mineral compositions as well as disequilibrium in Sm-Nd isotopic compositions of whole rock samples and coexisting plagioclase and orthopyroxene are presented. Disequilibrium textures presented include clinopyroxene exsolution lamellae in orthopyroxene; resorbed plagioclase in orthopyroxene or relict plagioclase; various inclusions such as orthopyroxene, plagioclase or clinopyroxene in larger oikocrysts of clinopyroxene or orthopyroxene; discontinuous rims of clinopyroxene surrounding orthopyroxene; resorbed orthopyroxene in clinopyroxene; and corona textures associated with olivine. These textures were used to derive a possible mineral crystallization sequence. At least two sequences of crystallization took place, both of which crystallized plagioclase first. One sequence then crystallized olivine which was then consumed to produce orthopyroxene which crystallized prior to late clinopyroxene. The other sequence indicates orthopyroxene crystallization after plagioclase crystallization, followed by crystallization of clinopyroxene. These sequences indicate at least two magmas were responsible for the genesis of the Merensky Reef and its hanging wall and footwall units. Compositionally, disequilibrium is evident in the range of compositions found in coexisting orthopyroxene, plagioclase and clinopyroxene with stratigraphic height, with particular reference to the change in mineral composition in each of the hanging wall, Reef and footwall units. Orthopyroxene compositions range in Mg numbers between 74.6 and 82.9 (77.4) in the hanging wall, 78.5 and 87.0 (avg. 81.1) in the Reef, and 77.9 and 84.1 (avg. 81.3) in the footwall. Plagioclase compositions range in An content between An64.9 and An82.3 (avg. An75.1) in the hanging wall, An56.8 to An70.8 (avg. An62.7) in the Reef, and An54.2 to An86.3 (avg. An73.2) in the footwall. In terms of Sm-Nd isotopic compositions, disequilibrium is evident between both whole rock samples and coexisting plagioclase and orthopyroxenes. Bulk rock Sm-Nd isotopic compositions show a range in ԐNd values between ԐNd (2.06 Ga) = -4.8 to -6.4 in the hangingwall, ԐNd (2.06 Ga) = -6.3 to -8.5 in the Reef, and ԐNd (2.06 Ga) = -4.5 to -6.3 in the footwall. Similar ԐNd values are present in the hanging wall and footwall units, with a clear “spike” in the Merensky Reef. ԐNd values in plagioclase are between ԐNd (2.06 Ga) = -5.8 and -7.8, while orthopyroxene isotopic Sm-Nd values are between ԐNd (2.06 Ga = -7.1 and -9.1. The mineral disequilibrium features presented within this study help elucidate the crystallization sequence of the magma as well as to constrain the contamination of the magma upon ascension and emplacement of the Merensky Reef. The results of this study favour a model where a mantle plume resulted in the ascent of a new magma which was contaminated by the assimilation of old, lower crust. Contamination took place prior to the possible lateral emplacement of the Merensky reef as a density current. 5-10% contamination of depleted mantle or a B2-“like” source by Archaean TTGs is modeled to achieve the contamination “spike” of ԐNd = -8.5 in the Merensky Reef.

Volatiles are a fundamental component of the Magmatic-Hydrothermal model of platinum group element (PGE) ore deposition for PGE deposits in layered mafic intrusions such as Bushveld and Stillwater. Volatiles have the potential to complex with PGEs in silicate melts and hydrothermal fluids, increasing PGE solubility; in order to assess the models of PGE ore deposition reliable estimates on the solubilities in the various magmatic phases must be known. However, experimental studies on the solubility and partitioning behaviour of PGEs in mafic magmatic-hydrothermal systems under relevant conditions are sparse, and the data that do exist produce conflicting results and new or adapted experimental methods must be applied to investigate these systems. Experimental results are presented here, investigating the effect of volatiles (i.e. H2O, Cl and CO2) on Pt and Ir solubility in a haplobasaltic melt and fluid-melt partitioning of Pt between an aqueous fluid and a haplobasaltic melt under magmatic conditions using a sealed-capsule technique. Also included are the details of the development of a novel experimental technique to observe fluid-melt partitioning in mafic systems and application of the method to the fluid-melt partition of Pt. Solubility experiments were conducted to assess the effect of volatiles on Pt and Ir solubility in a haplobasaltic melt of dry diopside-anorthite eutectic composition at 1523K and 0.2GPa. Synthetic glass powder of an anhydrous, 1-atm eutectic, diopside-anorthite (An42-Di58) haplobasalt composition was sealed in a platinum or platinum-iridium alloy capsule and was allowed to equilibrate with the noble metal capsule and a source of volatiles (i.e. H2O, H2O-Cl or H2O-CO2) at experimental conditions. All experiments were run in an internally-heated pressure vessel equipped with a rapid quench device, with oxygen fugacity controlled by the water activity and intrinsic hydrogen fugacity of the autoclave (MnO-Mn3O4). The resultant crystal- and bubble-free run product glasses were analyzed using a combination of laser ablation ICP-MS and bulk solution isotope-dilution ICP-MS to determine equilibrium solubilities of Pt and Ir and investigate the formation and contribution of micronuggets to overall bulk determined concentrations. In water-bearing experiments, it was determined that water content did not have an intrinsic effect on Pt or Ir solubility for water contents between 0.9 wt. % and 4.4 wt. % (saturation). Water content controlled the oxygen fugacity of the experiment and the resulting variations in oxygen fugacity, and the corresponding solubilities of Pt and Ir, indicate that over geologically relevant conditions both Pt and Ir are dissolved primarily in the 2+ valence state. Pt data suggest minor influence of Pt4+ at higher oxygen fugacities; however, there is no evidence of higher valence states for Ir. The ability of the sealed capsule technique to produce micronugget-free run product glasses in water-only experiments, allowed the solubility of Pt to be determined in hydrous haplobasalt at lower oxygen fugacities (and concentrations) then was previously observed. Pt and Ir solubility can be represented as a function of oxygen fugacity (bars) by the following equations: [Pt](ppb)= 1389(fO-sub-2)+7531(fO-sub-2)^(1/2) [Ir](ppb)=17140(fO-sub-2)^(1/2) In Cl-bearing experiments, experimental products from short run duration (<96hrs) experiments contained numerous micronuggets, preventing accurate determination of platinum and iridium solubility. Longer run duration experiments showed decreasing amounts of micronuggets, allowing accurate determination of solubility; results indicate that under the conditions studied chlorine has no discernable effect on Pt solubility in the silicate melt from 0.6 to 2.75 wt. % Cl (saturation). Over the same conditions, a systematic increase in Ir solubility is found with increasing Cl content; however, the observed increase is within the analytical variation/error and is therefore not conclusive. If there is an effect of Cl on PGE solubility the effect is minor resulting in increased Ir solubilities of 60% at chlorine saturation. However, the abundance of micronuggets in short run duration experiments, which decreases in abundance with time and increases with Cl-content, offers compelling evidence that Cl-bearing fluids have the capacity to transport significant amounts of Pt and Ir under magmatic conditions. It is suggested that platinum and iridium dissolved within the Cl-bearing fluid are left behind as the fluid dissolves into the melt during the heating stages of the experiment, leaving small amounts of Pt and Ir along the former particle boundaries. With increasing run duration, the metal migrates back to the capsule walls decreasing the amount of micronuggets contained within the glass. Estimates based on this model, using mass-balance calculations on the excess amount of Pt and Ir in the run product glasses (i.e. above equilibrium solubility) in short duration experiments, indicate estimated Pt and Ir concentrations in the Cl-bearing fluid ranging from tens to a few hundred ppm, versus ppb levels in the melt. Respective apparent (equilibrium has not been established) partition coefficients (D,fluid-melt) of 1x10^3 to 4x10^3 and 300-1100 were determined for Pt and Ir in Cl-bearing fluids; suggesting that Cl-bearing fluids can be highly efficient at enriching and transporting PGE in mafic magmatic-hydrothermal ore-forming systems. Platinum solubility was also determined as a function of CO2 content in a hydrous haplobasalt at controlled oxygen fugacity. Using the same sealed capsule techniques and melt composition as for H2O and Cl, a hydrous haplobasaltic melt was allowed to equilibrate with the platinum capsule and a CO2-source (CaCO3 or silver oxalate) at 1523 K and 0.2 GPa. Experiments were conducted with a water content of approximately 1 wt. %, fixing the log oxygen fugacity (bars) between -5.3 and -6.1 (log NNO = -6.95 @ 1573 K and 0.2 GPa). Carbon dioxide contents in the run product glasses ranged from 800-2500 ppm; and over these conditions, CO2 was found to have a negligible effect on Pt solubility in the silicate melt. Analogous to the Cl-bearing experiments, bulk concentrations of Pt in CO2-bearing experiments increased with increasing CO2 content due to micronugget formation. Apparent Pt concentrations in the H2O-CO2 fluid phase, prior to fluid dissolution, were calculated to be 1.6 to 42 ppm, resulting in apparent partition coefficients(D,fluid-melt) of 1.5 x 10^2 to 4.2 x 10^3, increasing with increasing mol CO2:H2O up to approximately 0.15, after which increasing CO2 content does not further increase partitioning. As well, a novel technique was developed and applied to assess the partitioning of Pt between an aqueous fluid and a hydrous diopside-anorthite melt under magmatic conditions. Building upon the sealed-capsule technique utilized for solubility studies, a method was developed by adding a seed crystal to the capsule along with a silicate melt and fluid. By generating conditions favourable to crystal growth, and growing the crystal from the fluid, it is possible to entrap fluid inclusions in the growing crystal, allowing direct sampling of the fluid phase at the conditions of the experiment. Using a diopside seed crystal with the diopside-anorthite eutectic melt, it was possible to control diopside crystallization by controlling the temperature, thus allowing control of the crystallization and fluid inclusion entrapment conditions. Subsequent laser ablation ICP-MS analysis of the fluid inclusions allowed fluid–melt partition coefficients of Pt to be determined. Synthetic glass powder of an anhydrous, 1-atm eutectic, diopside-anorthite (An42¬Di58) haplobasalt composition (with ppm levels of Ba, Cs, Sr and Rb added as internal standards), water and a diopside seed crystal were sealed in a platinum capsule and were allowed to equilibrate at experimental conditions. Water was added in amounts to maintain a free fluid phase throughout the experiment, and the diopside crystal was separated from the melt. All experiments were run in an internally heated pressure vessel equipped with a rapid-quench device, with oxygen fugacity controlled by the water activity and intrinsic hydrogen fugacity of the autoclave (MnO-Mn3O4). Experiments were allowed to equilibrate (6-48 hrs) at experimental conditions (i.e. 1498K, 0.2 GPa, fluid+melt+diopside stable) before temperature was dropped (i.e. to 1483K) to induce crystallization. Crystals were allowed to grow for a period of 18-61 hours, prior to rapid isobaric quenching to 293K at the conclusion of the experiment. Experimental run products were a crystal- and bubble-free glass and the diopside seed crystal with a fluid-inclusion-bearing overgrowth. Analysis of fluid inclusions provides initial solubility estimates of Pt in a H2O fluid phase at 1488 K and 0.2 GPa at or near ppm levels and fluid melt partition coefficients ranging from 2 – 48. This indicates substantial metal enrichment in the fluid phase in the absence of major ligands such as carbonate or chlorine. The results of this study indicate that the volatiles studied (i.e. H2O, CO2, and Cl) do not have a significant effect on Pt and Ir solubility in a haplobasaltic melt at magmatic conditions. These results suggest that complexing of Pt and Ir by OH, Cl, and carbonate species in a haplobasaltic melt is insignificant and the presence of these volatiles will not result in significantly increased PGE contents over their dry counterparts, as has been suggested. Preliminary evidence of minor Cl-complexing of Ir is presented; however, resulting in only a slight increase (<100%) in Ir solubility at Cl-saturation. Significant partitioning of Pt and Ir into a fluid phase at magmatic conditions has been demonstrated; with estimates of fluid-haplobasaltic melt partition coefficients increasing from 1x10^1 for pure water to up to an apparent 4x10^3 with the addition of Cl or CO2 to the system. This result indicates complexing of Pt and Ir with OH< HxCOy≤ Cl. Using these estimates, Cl- or CO2-bearing magmatic fluids can be highly efficient at enriching and transporting platinum group elements (PGEs) in mafic magmatic-hydrothermal ore-forming systems.

Volatiles are a fundamental component of the Magmatic-Hydrothermal model of platinum group element (PGE) ore deposition for PGE deposits in layered mafic intrusions such as Bushveld and Stillwater. Volatiles have the potential to complex with PGEs in silicate melts and hydrothermal fluids, increasing PGE solubility; in order to assess the models of PGE ore deposition reliable estimates on the solubilities in the various magmatic phases must be known. However, experimental studies on the solubility and partitioning behaviour of PGEs in mafic magmatic-hydrothermal systems under relevant conditions are sparse, and the data that do exist produce conflicting results and new or adapted experimental methods must be applied to investigate these systems. Experimental results are presented here, investigating the effect of volatiles (i.e. H2O, Cl and CO2) on Pt and Ir solubility in a haplobasaltic melt and fluid-melt partitioning of Pt between an aqueous fluid and a haplobasaltic melt under magmatic conditions using a sealed-capsule technique. Also included are the details of the development of a novel experimental technique to observe fluid-melt partitioning in mafic systems and application of the method to the fluid-melt partition of Pt. Solubility experiments were conducted to assess the effect of volatiles on Pt and Ir solubility in a haplobasaltic melt of dry diopside-anorthite eutectic composition at 1523K and 0.2GPa. Synthetic glass powder of an anhydrous, 1-atm eutectic, diopside-anorthite (An42-Di58) haplobasalt composition was sealed in a platinum or platinum-iridium alloy capsule and was allowed to equilibrate with the noble metal capsule and a source of volatiles (i.e. H2O, H2O-Cl or H2O-CO2) at experimental conditions. All experiments were run in an internally-heated pressure vessel equipped with a rapid quench device, with oxygen fugacity controlled by the water activity and intrinsic hydrogen fugacity of the autoclave (MnO-Mn3O4). The resultant crystal- and bubble-free run product glasses were analyzed using a combination of laser ablation ICP-MS and bulk solution isotope-dilution ICP-MS to determine equilibrium solubilities of Pt and Ir and investigate the formation and contribution of micronuggets to overall bulk determined concentrations. In water-bearing experiments, it was determined that water content did not have an intrinsic effect on Pt or Ir solubility for water contents between 0.9 wt. % and 4.4 wt. % (saturation). Water content controlled the oxygen fugacity of the experiment and the resulting variations in oxygen fugacity, and the corresponding solubilities of Pt and Ir, indicate that over geologically relevant conditions both Pt and Ir are dissolved primarily in the 2+ valence state. Pt data suggest minor influence of Pt4+ at higher oxygen fugacities; however, there is no evidence of higher valence states for Ir. The ability of the sealed capsule technique to produce micronugget-free run product glasses in water-only experiments, allowed the solubility of Pt to be determined in hydrous haplobasalt at lower oxygen fugacities (and concentrations) then was previously observed. Pt and Ir solubility can be represented as a function of oxygen fugacity (bars) by the following equations: [Pt](ppb)= 1389(fO-sub-2)+7531(fO-sub-2)^(1/2) [Ir](ppb)=17140(fO-sub-2)^(1/2) In Cl-bearing experiments, experimental products from short run duration (<96hrs) experiments contained numerous micronuggets, preventing accurate determination of platinum and iridium solubility. Longer run duration experiments showed decreasing amounts of micronuggets, allowing accurate determination of solubility; results indicate that under the conditions studied chlorine has no discernable effect on Pt solubility in the silicate melt from 0.6 to 2.75 wt. % Cl (saturation). Over the same conditions, a systematic increase in Ir solubility is found with increasing Cl content; however, the observed increase is within the analytical variation/error and is therefore not conclusive. If there is an effect of Cl on PGE solubility the effect is minor resulting in increased Ir solubilities of 60% at chlorine saturation. However, the abundance of micronuggets in short run duration experiments, which decreases in abundance with time and increases with Cl-content, offers compelling evidence that Cl-bearing fluids have the capacity to transport significant amounts of Pt and Ir under magmatic conditions. It is suggested that platinum and iridium dissolved within the Cl-bearing fluid are left behind as the fluid dissolves into the melt during the heating stages of the experiment, leaving small amounts of Pt and Ir along the former particle boundaries. With increasing run duration, the metal migrates back to the capsule walls decreasing the amount of micronuggets contained within the glass. Estimates based on this model, using mass-balance calculations on the excess amount of Pt and Ir in the run product glasses (i.e. above equilibrium solubility) in short duration experiments, indicate estimated Pt and Ir concentrations in the Cl-bearing fluid ranging from tens to a few hundred ppm, versus ppb levels in the melt. Respective apparent (equilibrium has not been established) partition coefficients (D,fluid-melt) of 1x10^3 to 4x10^3 and 300-1100 were determined for Pt and Ir in Cl-bearing fluids; suggesting that Cl-bearing fluids can be highly efficient at enriching and transporting PGE in mafic magmatic-hydrothermal ore-forming systems. Platinum solubility was also determined as a function of CO2 content in a hydrous haplobasalt at controlled oxygen fugacity. Using the same sealed capsule techniques and melt composition as for H2O and Cl, a hydrous haplobasaltic melt was allowed to equilibrate with the platinum capsule and a CO2-source (CaCO3 or silver oxalate) at 1523 K and 0.2 GPa. Experiments were conducted with a water content of approximately 1 wt. %, fixing the log oxygen fugacity (bars) between -5.3 and -6.1 (log NNO = -6.95 @ 1573 K and 0.2 GPa). Carbon dioxide contents in the run product glasses ranged from 800-2500 ppm; and over these conditions, CO2 was found to have a negligible effect on Pt solubility in the silicate melt. Analogous to the Cl-bearing experiments, bulk concentrations of Pt in CO2-bearing experiments increased with increasing CO2 content due to micronugget formation. Apparent Pt concentrations in the H2O-CO2 fluid phase, prior to fluid dissolution, were calculated to be 1.6 to 42 ppm, resulting in apparent partition coefficients(D,fluid-melt) of 1.5 x 10^2 to 4.2 x 10^3, increasing with increasing mol CO2:H2O up to approximately 0.15, after which increasing CO2 content does not further increase partitioning. As well, a novel technique was developed and applied to assess the partitioning of Pt between an aqueous fluid and a hydrous diopside-anorthite melt under magmatic conditions. Building upon the sealed-capsule technique utilized for solubility studies, a method was developed by adding a seed crystal to the capsule along with a silicate melt and fluid. By generating conditions favourable to crystal growth, and growing the crystal from the fluid, it is possible to entrap fluid inclusions in the growing crystal, allowing direct sampling of the fluid phase at the conditions of the experiment. Using a diopside seed crystal with the diopside-anorthite eutectic melt, it was possible to control diopside crystallization by controlling the temperature, thus allowing control of the crystallization and fluid inclusion entrapment conditions. Subsequent laser ablation ICP-MS analysis of the fluid inclusions allowed fluid–melt partition coefficients of Pt to be determined. Synthetic glass powder of an anhydrous, 1-atm eutectic, diopside-anorthite (An42¬Di58) haplobasalt composition (with ppm levels of Ba, Cs, Sr and Rb added as internal standards), water and a diopside seed crystal were sealed in a platinum capsule and were allowed to equilibrate at experimental conditions. Water was added in amounts to maintain a free fluid phase throughout the experiment, and the diopside crystal was separated from the melt. All experiments were run in an internally heated pressure vessel equipped with a rapid-quench device, with oxygen fugacity controlled by the water activity and intrinsic hydrogen fugacity of the autoclave (MnO-Mn3O4). Experiments were allowed to equilibrate (6-48 hrs) at experimental conditions (i.e. 1498K, 0.2 GPa, fluid+melt+diopside stable) before temperature was dropped (i.e. to 1483K) to induce crystallization. Crystals were allowed to grow for a period of 18-61 hours, prior to rapid isobaric quenching to 293K at the conclusion of the experiment. Experimental run products were a crystal- and bubble-free glass and the diopside seed crystal with a fluid-inclusion-bearing overgrowth. Analysis of fluid inclusions provides initial solubility estimates of Pt in a H2O fluid phase at 1488 K and 0.2 GPa at or near ppm levels and fluid melt partition coefficients ranging from 2 – 48. This indicates substantial metal enrichment in the fluid phase in the absence of major ligands such as carbonate or chlorine. The results of this study indicate that the volatiles studied (i.e. H2O, CO2, and Cl) do not have a significant effect on Pt and Ir solubility in a haplobasaltic melt at magmatic conditions. These results suggest that complexing of Pt and Ir by OH, Cl, and carbonate species in a haplobasaltic melt is insignificant and the presence of these volatiles will not result in significantly increased PGE contents over their dry counterparts, as has been suggested. Preliminary evidence of minor Cl-complexing of Ir is presented; however, resulting in only a slight increase (<100%) in Ir solubility at Cl-saturation. Significant partitioning of Pt and Ir into a fluid phase at magmatic conditions has been demonstrated; with estimates of fluid-haplobasaltic melt partition coefficients increasing from 1x10^1 for pure water to up to an apparent 4x10^3 with the addition of Cl or CO2 to the system. This result indicates complexing of Pt and Ir with OH< HxCOy≤ Cl. Using these estimates, Cl- or CO2-bearing magmatic fluids can be highly efficient at enriching and transporting platinum group elements (PGEs) in mafic magmatic-hydrothermal ore-forming systems.

Magister Scientiae - MSc (Earth Science) / This study area is situated within the Eastern Bushveld complex at Dwarsrivier chrome mine, which is approximately 30 km from Steelpoort and 60km from Lydenburg in the Mpumalanga province. The primary aim of the project is to identify the petrological and geochemical characteristics that can be used to distinguish the various rock types of feldspathic pyroxenites, chromitites, anorthosites and chromitite pyroxenites and determine whether the various rock types are from the MG2 package and MG3 package were formed from a single or multiple magma pulses. The geochemical and mineralogical variation studies were carried out using cores from borehole DWR74 and DWR172 located on the farm Dwarsrivier 372 KT. Using the combination of various multivariate statistical techniques (factor, cluster and discriminant analysis) multi element diagrams and trace element ratios, the outcome of the study demonstrated that each of the four rock types can be sub-divided into two groups.

M. Tech. (Department of Metallurgical Engineering, Faculty of Engineering Technology), Vaal University of Technology. / Lonmin mining company located in the Bushveld Complex of South Africa is one of the main platinum group elements (PGEs) producers in the world. Its core operations are made up of 11 shafts and inclines. There are resources of 181 million troy ounces of 3PGE + Au, and there are reserves of 32 million ounces of 3PGE + Au. One of the ore type produced at Lonmin is UG2 ore which is dominated by the high presence of chromite. The UG2 ore is also associated with PGE assemblages divided into sulphides and non-sulphides, and it is beneficiated through the froth flotation technique. Froth flotation is a physico-chemical process that is used for separation of desired valuable minerals from the gangue minerals by utilising the difference in surface properties. The process has been achieving lower recoveries with P4 (shaft name) UG2 ore compared to Eastern Platinum Limited (EPL) UG2 ore when using similar reagents suite, this leads to loss of valuable minerals to the tailings, both ores were from Lonmin. The first step was to conduct the mineralogical analysis conducted using Scanning electron microscopy- energy dispersive spectroscopy (SEM-EDS), X-ray diffraction (XRD), and optical microscopy to study the mineral composition of the two ores, and to identify any differences between them (two ores) considering that EPL UG2 ore is a blend of P1, P2, P3 (shaft names) and P4 UG2 ores while P4 UG2 ore is not blended with any other ores. The mineralogical results showed the presence of chromite, plagioclase, enstatite and sulphide minerals. The PGEs could not be detectable by any of the techniques used due to their small size and rarity. However, X-ray diffraction detected differences in concentrations of minor gangue constituents such as talc, muscovite, chlorite and actinolite and these results suggest that reagent consuming gangue mineralogy may have contributed to the differences in PGE recoveries by flotation. Batch flotation tests were also conducted. The existing reagent suite consisted of CuSO4 as an activator, Sodium n-propyl xanthate (SPNX) as a collector, carboxymethyl cellulose CMC as a depressant, and Senfroth 200 as a frother, and this was a single collector system. Therefore it was imperative to conduct flotation n investigation on alternative collector blends in order to improve the recovery of P4 UG2 ore. SNPX was used as the primary collector and it was blended with the following co-collectors: alkyl dithiocarbamates (DTC), two formulations of S-alkyl-N-butyl thionocarbamates (ABTC C1 & ABTC C2), and two formulations O-isopropyl-N-ethyl thionocarbamate (IPETC 30 & IPETC 31), one co-collector at a time. The first test incorporated the SNPX at dosage of 150 g/t without a blend and this dosage was selected based on the current optimum practice used at Lonmin and to use as a benchmark for the project. Trying to maintain the same dosage of 150 g/t of collectors, SNPX + co-collector were blended at two different dosages of 100 g/t + 50 g/t, and was also due to the fact that the co-collectors were highly concentrated and small dosages were expected to perform very well with SNPX. Lastly, the SNPX + co-collector at dosages 100 g/t + 125 g/t, here the dosage of co-collector was very high compared to 50 g/t and this was to check the effect of high dosages of highly concentrated collectors on the performance of the ore. The flotation results showed that the use of 50 g/t of co-collectors yielded optimum PGEs + Au recoveries and grades, while the dosage of 125 g/t decreased recoveries and grades. The high dosage quantities of collectors do not necessarily mean they will yield improved recoveries and grades. Different chain structures can be used to alter the behaviour of a collector, and these may increase or decrease their capabilities to cause higher recoveries. By using a collector with a longer hydrocarbon chain the flotation limit may be extended without loss of selectivity, consequently bringing about greater water repulsion, instead of increasing the concentration of a shorter chain collector. At 100 g/t of SNPX and 50 g/t of co-collector i.e. SNPX + IPETC 30 yielded improved 3PGE + Au recovery of 85.7 % at 3PGE + Au grades of 60.14 g/t, compared to the unblended SNPX (150 g/t) which yielded 3PGE + Au recovery of 81.1 % but insignificantly higher grade of 60.53 g/t. On the other hand, SNPX + IPETC C1 blend yielded low 3PGE + Au recoveries compared to SNPX + IPETC 30 and SNPX + IPETC 31 blends, but it achieved the highest grade of 76.1 g/t. Evidently, this proves that the relationship between recovery and grade is a trade-off. The results have also shown the synergic effects, especially for SNPX blended with IPETC 30, and SNPX blended with IPETC 31 at dosage of 100 g/t (SNPX) and 50 g/t (IPETCs). It can be concluded that the different interaction obtainable from the thionocarbamate (ROCSNHR), effectively complement that from the xanthate ion (ROCS2–) to achieve more collector interaction at surface sites otherwise interactable for xanthate only. Therefore the collector blends rendered the mineral of interest hydrophobic and as a result the minerals were recovered to the concentrate. On the other hand, too much of collectors may not be beneficial. At the dosage of 100 g/t of SNPX and 125 g/t of collectors, SNPX + DTC attained lower recoveries compared to SNPX, SNPX + IPETC 30, SNPX + IPETC 31, however the grade was higher than achieved SNPX + IPETC 30, SNPX + IPETC 31 and SNPX + ABTC C1. Nevertheless, comparing these results to the dosage of 50 g/t of the co-collectors, the 125 g/t did not perform well at all. The dosage of 125 g/t of co-collectors lead to loss of collecting power and selectivity, especially for SNPX + IPETC 30, SNPX + IPETC 31, and SNPX + ABTC C1 blends. It is therefore wise to conduct an optimisation test to determine the correct dosing rate. In addition, the chromite entrainment was below the smelter limit and is very beneficial since chromite is detrimental to the furnace. Therefore, it is concluded that the blends of SNPX with IPETC 30 and IPETC 31 at a dosage 50 g/t have shown satisfying recoveries and the FeCr2O4 recovery is less than 1 % meaning there will not be any smelter penalties for FeCr2O4 content. Therefore, these are the recommended collector blends. It is recommended that further mineralogical study of the ores be conducted so that it may provide deeper insight into the causes of low recoveries under SNPX only. The system of blended collectors and its optimisation would be beneficial and can be practiced. The Chemisorption studies between the minerals and co-collectors used will provide more specific insight and details into the actual interaction synergy that gave the improved recoveries.

Thesis (M. Sc.(Geography)) -- University of Limpopo, 2019 / Global production of Platinum Group Metals (PGMs) is dominated by South Africa due to its large economic resources base in the Bushveld Igneous Complex (BIC). PGMs are used in a wide range of high technology applications worldwide including medicinal, industrial and commercial purposes, and its contribution to the Gross Domestic Product (GDP) and creating jobs for many. In an area where mining activities dominate, there are likely to be problems that need effective environmental management approaches, which can be facilitated through legislations. Marula Platinum Mine (MPM) is located in Limpopo province BIC which has the second largest number of mining productivity in South Africa. Environmental legislations have been put in place by the South African government in order to avoid or minimise the footprints caused by PGM mining. This study looked at environmental compliance with solid waste management practices by Marula Platinum Mine (MPM) as guided by Mineral and Petroleum and Resource Development Act (MPRDA) and National Environmental Management Act (NEMA) as well as the environmental impacts of MPM in the surrounding communities. Both primary (questionnaires, field observations and key informant interviews) and secondary (NEMA, MPRDA, journals, reports, pamphlets, internet and books) data was used to address the objectives of the study. Descriptive method and Statistical Package for Social Sciences (SPSS) version 25 were used for the analysis of data. The key research results revealed that MPM was compliant with 65% and 21% partially compliant with solid waste management practices. Only 14% of information on solid waste management practices could not be accessed because MPM is still operational. MPM had also had negative footprints on the surrounding villages such as dust generation and cracks on walls and floors on houses of community members, strikes and increase in the usage of substance abuse. Recommendations of the study are that MPM should address challenges that hinder environmental compliance so as to be 100% compliant with MPRDA and NEMA regulations. MPM should also provide other mitigation measures for blasting of explosives to reduce dust generation and problems of cracks on houses of surrounding village members.