Thesis (MSc (Earth Sciences))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: The Bushveld Complex of South Africa contains three of the most important platinum deposits in the world namely the Merensky Reef, the Upper Group Two (UG2) chromitite reef and the Platreef. These three ore bodies are principally beneficiated by froth flotation. During the beneficiation of chromite hosted PGE’s by froth flotation, chromite represents the principal gangue mineral. This is particularly true for the UG2 main seam. An excess of more than 3% in mass of chromite in the PGM concentrate is known to result in significant problems in the downstream processing and extraction of PGEs. The variability in texture and composition of chromite due to its primary crystallization and subsequent modification by the development of potholes or through IRUP intrusions are thought to influence the flotation behaviour of the UG2 main seam chromitite ore. This study conducted at Waterval Mine investigated the role of mineralogical characteristics of chromites on the flotation performance of three different environments for the UG2 main seam: (1) “normal” UG2 main seam; (2) UG2 main seam affected by pothole formation; and (3) UG2 main seam affected by IRUP intrusion. This was achieved through an extensive petrographic investigation of the chromites from each environment, to individually characterise their primary textures. This was followed by compositional characterisation of the chromite from each environment. Finally the flotation performance of the ore from each environment was investigated, using small scale batch flotation experiments, to establish any linkage between the textures, the composition and the flotation performance of the chromite from different environments. In this study it was found that the UG2 normal reef and the UG2 reef affected by pothole formation are both principally characterised by primary mineralogical features comprising mainly fine‐grained chromite as the cumulate phase and orthopyroxene and plagioclase as intercumulate phases. These two reef types were also found to be identical in the composition of the chromites present. In addition, in both of these almost unaltered reef types it was found that chromite showed small recoveries by flotation. On the other hand, it was found that the UG2 affected by IRUP intrusion was affected by post‐magmatic alteration that had overprinted primary textures and compositional features. This resulted in the replacement of primary minerals by secondary alteration assemblages. Orthopyroxene was iii replaced by serpentine, chlorite, amphibole and talc, while plagioclase is replaced by sericitic alteration. Furthermore, this alteration also resulted in modification of the chromite compositions. The compositional change in the chromites from the IRUP reef type resulted in Fe and Ti enrichment of chromite with increasing magnetic properties, and Cr, Al and Mg depletion. The alteration also resulted in the coarsening of chromite in the IRUP affected main seam reef particularly at the bottom and the top of the main seam. These compositional and textural modifications, principally the post‐magmatic alteration of intercumulate orthopyroxene, resulted in a greater recovery of chromite by flotation in the concentrate from the IRUP affected ore compared to the two other two ore types where there was small amount of chromite recovered. The characterisation of the recovered chromite revealed that the principal reason for chromite flotation was caused by the mineral association of chromite with hydrophobic Si, Mg, Fe rich phases, principally altered orthopyroxene and associated serpentine, chlorite, amphibole and talc. This investigation showed that the difference in mineralogical and flotation performances of chromite from the different UG2 main seam reef types was caused by the postcrystallisation alteration of cumulate and intercumulate phases due to the emplacement of IRUPs. Although IRUP affected UG2 main seam ore is not currently processed, it could be processed much more rapidly than the other two types of UG2 main seam ores because of its softer character resulting in shorter milling times. This is most likely a function of the presence of alteration phases and the presence of coarser chromite grains, as well as already brecciated chromite grains. Savings associated with the shorter milling time of this ore type are perhaps offset by the cost of the higher dosages of depressant required to suppress the floatable chromite in this ore type. However, given the energy cost of longer milling times, the cost of the depressant is likely to be insignificant. Moreover, the processing of the UG2 main seam ore affected by IRUP intrusion would also require a different approach to extraction of the ore to keep it separate from the normal reef ore.

Thesis (MSc (Earth Sciences))--University of Stellenbosch, 2011. / ENGLISH ABSTRACT: The Central Zone (CZ) of the Damara belt in central Namibia is underlain by voluminous Pan-African granites and is host to numerous pegmatite occurrences, some of which have economic importance and have been mined extensively. This study discusses the occurrence, geometry, relative timing and emplacement mechanisms for the Usakos pegmatite field, located between the towns of Karibib and Usakos and within the core of the regional-scale Kranzberg syncline. Lithological mapping of the Kuiseb Formation in the core of the Kranzberg syncline identified four litho-units that form an up to 800 m thick succession of metaturbidites describing an overall coarsening upward trend. This coarsening upwards trend suggests sedimentation of the formation’s upper parts may have occurred during crustal convergence and basin closure between the Kalahari and Congo Cratons, rather than during continued spreading as previously thought. The Kranzberg syncline is a regional-scale NW verging, NE-SW trending, strongly non-cylindrical structure that consists of a moderately SE dipping, normal NW limb and a steep- to overturned SE limb. First- and lower-order folds show relatively consistent E - SE plunges at moderate angles and stretching lineations and boudinage of competent layers point to a fold-parallel stretch during folding. Folding is associated with a moderate- to steep SE dipping transecting foliation that shows a consistent anticlockwise rotation with respect to the axial plane of the fold. The transecting cleavage and a component of non-coaxial shear along the overturned limb suggest that folding was accomopanied by a dextral component of shear thought to be related to the SW-directed extrusion of the adjacent Usakos dome during regional NW-SE directed shortening. It is further suggested that the Kranzberg syncline evolved within the overall regional pattern of regional dome and syncline structures in the sCZ, and not as a forced fold in response to the formation of neighbouring dome structures. Based on cross-cutting relationships and deformation, four main generations of bedding-concordant sills and bedding-discordant pegmatite dykes were identified. Along the normal limb, shallowly-dipping sills dominate, highlighting the significance of bedding anisotropies for sheet propagation. Along the overturned limb, interconnected dyke and sill geometries co-exist. Here, pegmatite emplacement appears to have been influenced by (1) the regional strain, (2) differing wall rock rheologies; (3) the orientation of pre-existing anisotropies; and (4) driving melt pressures. Dykes within the Usakos pegmatite field formed within dilational sites, at high angles to the regional stretch, whereas sills formed at high angles to the regional shortening strain and in contractional sites. Where driving pressures for melt ascent were high enough, an interconnectivity of dykes and sills and subsequent melt transfer from contractional into dilational sites is developed. Where melt pressures dropped below a critical value pegmatites were arrested, thus preserving the ascent pathways of the melts. These complex intersecting melt pathways are developed throughout the Kranzberg syncline. This suggests the existence of fairly stable melt networks in the continental crust. This geometrical complexity also accounts for the stockwork-like structures observed in pegmatite fields. / AFRIKAANSE OPSOMMING: Die Sentrale Sone (CZ) van die Damara gordel in sentrale Namibië is onder lê deur volumineuse Pan-Afrikaanse graniete en speel gasheer vir talle pegmatiet voorkomste, waarvan party van ekonomiese belang is en is ekstensief ontgin. Hierdie studie bespreek die voorkoms, geometrie, relatiewe tydsberekening en inplasing meganismes vir die Usakos pegmatiet gebied, wat tussen die stede van Karibib en Usakos en wat binne die kern van die regionale-skaal Kranzberg sinklien geleë is. Litologiese kartering van die Kuiseb Formasie in die kern van die Kranzberg sinklien het vier lito-eenhede geidentifiseer. Hierdie eenhede, wat saam tot 'n 800 m dik opeenvolging van metaturbidiete vorm, beskryf ‘n algemene opwaartse vergrowwing neiging. Hierdie tendens dui aan dat sedimentasie van die Formasie se boonste dele tydens die aardkorst konvergensie en kom sluiting tussen die Kalahari en die Kongo kratons voorgekom het, eerder as in 'n oseaanvloerverbreiding omgewing soos voorheen gedink was. Die Kranzberg sinklien is 'n regionale-skaal struktuur met ‘n NW vergensie, ‘n NOSW koersing, wat sterk nie-silindries is en wat uit 'n matige SO helling, normale NW flank en 'n steil-tot omgeslaande SO flank bestaan. Eerste-en laer-orde plooie vertoon relatief konsekwent matige O - SO duikings en strek lineasies en boudinage van kompetent lae dui 'n plooi parallel strek tydens plooiing aan. Plooiing is geassosieer met 'n ongeveer aksiale planêre, matig- tot steil SO helling foliasie wat omstandig waargeneem word om 'n konsekwente antikloksgewyse rotasie met betrekking tot die aksiale vlak van die plooi te hê. Hierdie antikloksgewyse rotasie is ‘n bewyse vir 'n komponent van nie-koaksiale regse skuifskeur deur die omgekeerde flank en dui dit ook aan dat 'n regse komponent van skeer gedurend of na plooiing plaasgevind het. Daar is gedink dat die regse komponent van skeur in verband met die laterale, SW-gerig extrusie van die aangrensende Usakos koepel gedurende plaaslike NW-SO verkorting ontwikkel het. Dit is verder voorgestel dat die Kranzberg sinklien binne die totale patroon van plaaslike koepel en sinklien strukture in die sCZ geontwikkel het, en nie as 'n gedwonge plooi in reaksie op die formasie van die naburige koepel strukture (bv. Usakos koepel). Gebaseer op kruis-sny verhoudings en deformasie, was vier generasies van gelaagdheid-konkordant plate en gelaagdheid-diskordant pegmatiet dyke geïdentifiseer. In die normale flank, vlak-helling plate oorheers, wat die belangerikheid van die laagvlak-anisotropiese op plaat voortplanting beklemtoon. In die steil, omgekeerde flank, bestaan onderlinge verbinde dyk en plaat geometrië gelyktydig. Hier is pegmatiet inplasing blykbaar beïnvloed deur (1) die regionale span; (2) verskillende wandgesteentes reologië; (3) die oriëntasies van anisotropie (ie. gelaagdheid ); en (4) smeltsel druk. Dyke in die Usakos pegmatiet gebied het binne dilatasionele liggings, teen hoë hoeke aan die regionale strek gevorm, terwyl plate teen hoë hoeke aan die plaaslike verkorting span en in kontraksionele liggings gevorm het. Waar smeltsel druk hoog genoeg was, is 'n onderlinge verbinding van dyke en plate, en die daaropvolgende smeltsel oordrag van kontraksionele liggings na dilatasionele liggings behou. In teenstelling, waar smeltsel druk onder 'n kritieke waarde geval het, word die pegmatiete geblokeer, en dus kan die behoude smeltsel styging paaie waargeneem word. Hierdie snyende smeltsel geometrië, in beide kontraksionele en dilatasionele liggings dui aan dat redelik stabiele smeltsel netwerke in die kontinentale kors kan bestaan en verder kan en verklaar die algemene stokwerk-agtige strukture wat in pegmatiet velde van ander mid-korstige omgewings waargeneem word.

Thesis (MSc)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: The Cretaceous Ambohiby Complex is an alkaline ring complex located in the central part of Madagascar and covers a mountainous area of approximately 225km2. The complex intrudes into Precambrian basement gneisses and consists of the following rock types in a chronological order; gabbros, monzonite, alkali-syenite, micro-granite and granites. Both mafic and felsic rocks are dominated by sodic mineralogies. Pyroxenes are generally aegirine, aegirine-augite, and hedenbergite and commonly occur in granites, micro-granites, syenites and monzonite. In gabbros and mafic dykes, augite is the more common composition. Amphiboles are represented by bluish to brownish-green varieties with arfvedsonite to eckermannite compositions in granites, and magnesia-arfvedsonite compositions in micro-granites. Ferro-edenite is present in some alkali-syenites and monzonite. Feldspars are usually single phase and are therefore hypersolvus. In granites, micro-granites and alkali-syenites, path and string perthite is very common. Graphic intergrowth of quartz and alkali feldspars is also common in granites and some alkalisyenites. Major elements variation diagrams plotted against SiO2 indicate that the mafic and felsic rocks of the Ambohiby Complex were formed by processes similar to those of Fractional crystallization. Chondrite normalised mafic rocks have slightly positive Eu anomalies while the felsic rocks have negative Eu anomalies, indicating fractionation of plagioclase feldspars. The Chondrite normalised gabbroic rocks shared similar trends of heavy rare earth with Chondrite normalised Marion Hotspot data. This suggests that the basaltic parent magma for the Ambohiby Complex, possibly related to the Marion hotspot plume. The Fractional crystallization model with an inclusion of olivine in the mineral assemblage seems to fit very well with the actual Ambohiby felsic end member rocks (i.e. granites). It is therefore clear that differentiation mainly occurred by fractional crystallization but variable initial Sr and Nd values indicate the magmas assimilated crustal material during emplacement. The Rb-Sr geochronology gave an age of 90±2.4 Ma for the intrusion of the Ambohiby Complex, which confirms that the Ambohiby Complex is associated with the Gondwana break-up. In addition the Marion Hotspot plume is believed to have been located in the southern tip of the island at around 90 Ma ago.

Thesis (MSc)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Five areas proximal to the world class Namakwa Sands heavy mineral deposit have been studied and include the farms Houtkraal Remainder Portion 2, Houtkraal Remainder, Geelwal Karoo, Graauwduinen and Rietfontein. These are locally referred to as the satellite deposits and are sub-economic occurrences. The primary objective of the study was to quantify the mineralogy and mineral chemistry, determine the provenance of the heavy mineral suite and draw a comparison between the satellite deposits and the Namakwa Sands deposit from an exploratory point of view. Methodology used to achieve the above objectives included optical microscopy, Scanning Electron Microscope (SEM), Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS), Quantitative Evaluation of Minerals by Scanning Electron Microscopy (QEMSCAN), X- Ray Fluorescence (XRF) and Zr-geothermometry of rutile. The five satellite areas contain the same heavy mineral suite but mineral proportions differ. The total heavy mineral population (THM) are diverse and consist of ilmenite and its alteration products (hydrated ilmenite, pseudorutile and leucoxene), magnetite, hematite, spinel, rutile, tourmaline, pyroxene, amphibole, garnet, aluminosilicates, staurolite, corundum, epidote, zircon, monazite and sphene. Ilmenite and garnet are the two most dominant heavy minerals present. The valuable heavy minerals (VHM) suite consists of ilmenite, zircon, rutile and leucoxene. The mineralogy of the satellite areas and chemistry of the ore minerals (rutile, zircon, ilmenite and leucoxene) are similar to the Namakwa Sands deposit. The whole spectrum of ilmenite alteration products (hydrated ilmenite, pseudorutile, and leucoxene) is present and allowed the quantitative use of the alteration index. The indices is low (22-24%) and indicates that the surficial deposits have formed under arid to semi-arid climatological conditions which preserved the pristine character of most of the minerals. This allowed reliable provenance studies using the characteristics of most of the heavy mineral suite, which showed that the minerals were derived from a diversity of source rocks. These included mainly medium to high-grade metamorphites and felsic intrusives of the underlying Mesoproterozoic Namaqualand Metamorphic Complex and a minor contribution from the Neoproterozoic Gariep Supergroup. This relationship indicates a limited transport distance from source to depositional basin. Mineral ratios in particular the THM-VHM relationship showed that the deposits located close to the shoreline such as Geelwal Karoo, Graauwduinen and Rietfontein have a relatively low proportion of valuable heavy minerals whereas those more inland such as Houtkraal Remainder Portion 2 and Houtkraal Remainder are close to unity. Heavy mineral concentration as such is low in the satellite areas and the mechanism to increase the concentration is clearly not only a function of distance from the present shoreline but is also topographically controlled. Steep sided linear depressions channelled the unconsolidated sediments and heavy minerals were upgraded into economic concentrations by aeolian processes. The quality of the valuable heavy minerals in the satellite areas however is similar to those of the adjacent Namakwa Sands deposit. This study has demonstrated that Houtkraal Remainder is the northeasterly continuation of the red aeolian sand (RAS) associated East Mine orebody and offers the best exploration potential. / AFRIKAANSE OPSOMMING: Vyf areas proksimaal aan die Namakwa Sands swaar mineral afsetting is bestudeer en sluit in Houtkraal Remainder, Houtkraal Remainder Portion 2, Geelwal Karoo, Graauwduinen en Rietfontein. Hierdie areas word plaaslik na verwys as satelliet afsettings en is subekonomies. Die hoofdoel van hierdie studie was om die mineralogie en mineral chemie te kwantifiseer, die oorsprong van die swaar mineraal suite te bepaal asook n vergelyking te tref tussen die satelliet areas en die Namakwa Sands afsetting vanuit n verkennende eksplorasie oogpunt. Optiese mikroskopie, SEM, LA-ICP-MS, QEMSCAN, XRF en die Zr-geotermometer van rutiel is gebruik om bostaande doele te bereik. Die vyf satellite areas bestaan uit dieselfde swaar minerale maar mineral proporsies verskil. Die totale swaar mineraal populasie is divers en bestaan uit ilmeniet en ilmeniet se veranderingsprodukte (gehidreerde ilmeniet, pseudorutiel en leukokseen), magnetiet, hematiet, spinel, rutiel, toermalyn, pirokseen, amfibool, granaat, aluminiumsilikate, stauroliet, korund, epidoot, sirkoon, monasiet and sfeen. Ilmenite en granaat is die twee mees dominante swaar minerale teenwoordig. Die waardevolle swaar mineraal populasie bestaan uit ilmeniet, rutiel, sirkoon en leukokseen. Die mineralogie van die satelliet areas en die chemie van die erts minerale (rutiel, sirkoon, ilmeniet en leukokseen) is dieselfde as die van die Namakwa Sands afsetting. Die hele spektrum ilmeniet veranderingsprodukte is teenwoordig en het die kwantitatiewe gebruik van die alterasie indeks toegelaat. Die alterasie indekse is laag (22-24%) en dui aan dat die oppervak afsettings gevorm het tydens droë tot semi droë toestande wat die eertydse karakter van meeste minerale bewaar het. Deurdat die karakter van meeste minerale behoue gebly het, kon provenans studies toegepas word op die swaar mineraal suite. Provenans studies het aangedui dat die swaar minerale afkomstig is van n verskeidenheid van bron gesteentes. Dit sluit in medium tot hoë graad metamorfe gesteentes en felsiese intrusies van die Mesoproterosoïese Namakwaland Metamorfiese Kompleks met n geringe bydrae van die Neo Proterosoïese Gariep Supergroep. Hierdie verhouding dui n beperkte vervoer afstand aan vanaf die bron tot by die afsettings omgewing. Mineraal vehoudings spesifiek die totale swaar mineraal-waardevolle swaar mineraal verhoudings dui aan dat afsettings na aan die kus soos Geelwal Karoo, Graauwduinen en Rietfontien n lae inhoud van waardevolle swaar minerale het teenoor afsettings soos Houtkraal Remainder Portion 2 en Houtkraal Remainder wat meer land in is met verhoudings na aan eenheid. Swaar mineral konsentrasie is laag in die satellite areas en die meganisme verantwoordelik vir die toename in konsentrasie is nie net n funksie van die afstand van die bestaande kuslyn nie maar word ook deur topografie beheer. Steil sydige lineêre depressies kanaliseer die ongekonsolideerde sediment en swaar minerale en word opgradeer tot ekonomiese konsentrasies deur wind prossese. Die kwaliteit van die waardevolle swaar minerale in die satelliet areas is egter dieselfde as die van die aangrensende Namakwa Sands afsetting. Hierdie studie het gewys dat Houtkraal Remainder is die noordelike voortsetting van die Rooi Aeoliese Sand geassosieerde Oos Myn ertsliggam en bied die beste eksplorasie potensiaal.

Thesis (MSc)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: To date, the only volcanic rocks described from the Cape Granite Suite are ignimbrites that crop out along the western margin of Langebaan lagoon, to the south of Saldahna. These ignimbrites, with an age of 515 ± 3 Ma, represent the youngest rocks within the Suite. This study aims to investigate the petrogenesis of fine grained granitic rocks exposed to the north of Langebaan lagoon, as well as to reassess the classification of these rocks as a sub-volcanic quartz porphyritic intrusive. These rocks exhibit a dark grey to brown micro- to crypto-crystalline matrix containing prominent feldspar and quartz phenocrysts that are often embayed and broken. Phenocrysts of biotite and orthopyroxene (Fs50-70) can be identified microscopically, as can replacement of both phases by chlorite. Ilmenite commonly occurs in close association with the orthopyroxene phenocrysts. A second generation of poikiloblastic biotite overgrows the matrix and is clearly formed through sub-solidus reaction. The feldspar phenocrysts are commonly microcline microperthite. Contact exposures can be observed in the study area which indicates that the finer grained rock intruded older coarser grained granites. The predominance of broken phenocrysts as well as the presence of fiamme present in outcrop in rocks with a microcrystalline matrix is indicative of a volcanic origin, suggesting that these rocks be classified as ignimbrite as opposed to quartz porphyry. The foregoing observations and features could be interpreted to represent a welded ignimbrite deposit which is underlain by coarse grained granite. The finer grained ignimbrite would then represent a subsequent intrusion by a later pulse of similar magma along the contact with the coarser grained granite. The ignimbrites are silicic with SiO2 ranging between 69 and 76 wt. %; they are mildly peraluminous with values for ASI (ASI = mol. Al2O3/ (CaO+Na2O+K2O)) ranging from 1.02 to 1.09; and ASI is negatively correlated with Mg + Fe (hereafter maficity). Tight to very tight inter-element correlations exist for several major elements as well as trace elements when plotted against maficity. The following R² values apply: Al =0.94; Ca = 0.98; Si = 0.97; Ti = 1.00; Na =0.90; Zr =0.95; La = 0.87. These elements are all positively correlated with maficity, except for Si which is negatively correlated. Orthopyroxene and ilmenite represent early formed, high temperature minerals in the magma. In the biotite-poor rocks, ilmenite represents the main reservoir of titanium whilst orthopyroxene represents the main MgO and FeO reservoir. The exceptionally tight Ti: maficity correlation requires that both these minerals always be present at the same molecular ratio in the magma, despite the significant range in maficity portrayed by the rocks and despite the fact that these minerals have different size-density relationships. This exceptionally tight correlation can be readily interpreted to reflect entrainment of a peritectic assemblage consisting of ilmenite and orthopyroxene. The Al, Ca and Na correlations require the entrainment of peritectic plagioclase. The decreasing trend for ASI requires the entrainment of peritectic clionopyroxene. A near perfect match with the concentrations of these elements in the ignimbrites is produced by modelling entrainment of a peritectic assemblage consisting of plagioclase, ilmenite, orthopyroxene and clionopyroxene in stoichiometric proportions dictated by the melting reaction. A peritectic assemblage formed by these phases’ points to the partial melting of a source undergoing coupled biotite and hornblende fluid-absent melting, with hornblende being subordinate. The opx- and ilmenite-rich micro-domains in the rocks represent zones in the magma rich in original peritectic orthopyroxene and ilmenite. In contrast, the peritectic plagioclase demanded by the chemistry of the rocks has melted during ascent due to overheating and decreasing water solubility in the melt. The phenocrystic potassium feldspar observed in the rocks crystallised after significant cooling and the physical behaviour of these crystals does not shape the chemistry of the magma. K contents of the ignimbrites are however not well replicated by this modelling, which predicts a significant K decrease due to dilution. K in the rocks is not correlated with maficity. This may reflect the fact that the K behaviour represents two slightly different source protoliths with differing K contents. / AFRIKAANSE OPSOMMING: Tot op hede is die enigste vulkaniese rots van die Kaapse Granietgroep wat al beskryf is ignimbriete wat teen die westelike grens van die Langebaan-lagune, na die suide van Saldanha, aan die oppervlak kom. Hierdie ignimbriete, met ’n ouderdom van 515 ± 3 Ma, verteenwoordig die jongste gesteentes in die Groep. Die doel van hierdie studie was om die petrogenese van fynkorrelrige granietrots wat na die noorde van die Langebaan-lagune blootgestel is, te ondersoek, en ook die klassifikasie van hierdie rotse as ’n subvulkaniese kwartsporfier- intrusiewe gesteentes te assesseer. Hierdie gesteentes toon ’n donker grys tot bruin mikro- tot kriptokristalvormige matriks wat prominente veldspaat en kwartsfenokriste bevat wat dikwels bogtig en gebreek is. Fenokriste van biotiet en ortopirokseen (Fs50-70) asook vervanging van albei fases deur chloriet kan mikroskopies geïdentifiseer word. Ilmeniet kom dikwels naby ortopirokseenfenokriste voor. ’n Tweede generasie poikiloblastiese biotiet groei die matriks toe en is duidelik deur subsolidusreaksie gevorm. Die veldspaatfenokriste is gewoonlik mikroklien-mikropertiet. Kontakblootstelling kan in die studiegebied waargeneem word, wat aantoon dat die fynkorrelrige rots ander grofkorrelrige granietgesteentes intrudeer het. Die oorheersing van gebroke fenokriste asook die teenwoordigheid van ‘fiamme’ strukture in die rotsdagsoom met ’n mikrokristalvormige matriks dui op vulkaniese oorsprong, wat aan die hand doen dat hierdie gesteentes as ignimbriete eerder as kwartsporfier geklassifiseer kan word. Hierdie waarnemings en eienskappe kan geïnterpreteer word as verteenwoordigend van ’n gelaste ignimbriet-afsetting wat deur grofkorrelrige graniet onderlê word. Die fynkorrelrige ignimbriet stel dan ’n daaropvolgende intrusie voor deur ’n latere puls van soortgelyke magma teen die kontak met die grofkorrelrige graniet af. Die ignimbriete is silisies met SiO2 wat wissel tussen 69 en 76 wt. %; hulle is matig peralumineus met waardes vir ASI (ASI = mol. Al2O3/ (CaO+Na2O+K2O)) wat wissel van 1.02 tot 1.09; en ASI is negatief gekorreleer met Mg + Fe (hierná mafiese komponent). Nou tot baie nou inter-elementkorrelasies bestaan vir verskeie groot elemente asook spoorelemente wanneer dit teen die mafiese komponent gestip word. Die volgende R²-waardes is van toepassing: Al =0.94; Ca = 0.98; Si = 0.97; Ti = 1.00; Na =0.90; Zr =0.95; La = 0.87. Hierdie elemente is almal positief met die mafiese komponent gekorreleer, buiten Si, wat negatief gekorreleer is. Ortopirokseen en ilmeniet verteenwoordig vroeg gevormde, hoëtemperatuur-minerale in die magma. In die biotiet-arme rotse stel ilmeniet die hoofreservoir van titaan voor, terwyl ortopirokseen die vernaamste MgO- en FeO-reservoir voorstel. Die buitengewoon nou Ti: mafiese-korrelasie vereis dat albei hierdie minerale altyd in dieselfde molekulêre verhouding in die magma teenwoordig moet wees, ondanks die beduidende omvang van die mafiese komponent wat deur die gesteentes getoon word en ondanks die feit dat hierdie minerale verskillende grootte–digtheidsverhoudings het. Hierdie buitengewoon nou korrelasie kan geredelik geïnterpreteer word om meesleping van ’n peritektiese groep te weerspieël wat uit ilmeniet en ortopirokseen bestaan. Die Al-, Ca- en Na-korrelasies vereis die meesleping van peritektiese plagioklaas. Die verminderende neiging tot ASI vereis die meesleping van peritektiese klionopirokseen. ’n Byna perfekte passing met die konsentrasies van hierdie elemente in die ignimbriete word voortgebring deur die modellering van meesleping van ’n peritektiese groep bestaande uit plagioklaas, ilmeniet, ortopirokseen en klionopirokseen in stoïgiometriese verhoudings wat deur die smeltreaksie bepaal word. ’n Peritektiese groep wat deur hierdie fases gevorm word, dui op die gedeeltelike smelting van ’n bron wat gekoppelde biotiet- en horingblende- vloeistofafwesige smelting ondergaan, met horingblende wat ondergeskik is. Die ortopirokseen- en ilmeniet-ryke mikrodomeins in die gesteentes verteenwoordig sones in die magma wat ryk is aan oorspronklike peritektiese ortopirokseen en ilmeniet. Hierteenoor het die peritektiese plagioklaas wat deur die chemie van die gesteentes vereis word tydens styging gesmelt weens oorverhitting en dalende wateroplosbaarheid in die smeltsel. Die fenokristiese kaliumveldspaat wat in die rotse waargeneem is wat ná aanmerklike afkoeling gekristalliseer het en die fisiese gedrag van hierdie kristalle vorm nie die chemie van die magma nie. Die K-inhoud van die ignimbriete word egter nie goed deur hierdie modellering gerepliseer nie, wat ’n aanmerklike K-afname weens verdunning voorspel. K in die rotse is nie met mafiese komponente gekorreleer nie. Dit kan die feit weerspieël dat die K-gedrag twee effens verskillende bronprotoliete met verskillende K-inhoud voorstel.

Thesis (MSc)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: Rhythmic magmatic layering in granites is an intriguing feature that has been reported from plutons with contrasting chemical compositions from a wide range of tectonic settings. Layered granites are rare and occur in association with volumetrically dominant non-layered rocks having similar composition. Understanding the origin of such layering in granites, particularly from well exposed outcrops, may give crucial insights on the physical-chemical conditions and processes (such as fractional crystallization, size and composition of individual magma batches, efficiency of mixing between batches etc.) within magma chambers during their construction. In the Cape Granite Suite, rhythmic layering is exceptionally well preserved in outcrops of the S-type granodioritic to monzogranitic Peninsula pluton along the coastline of the small town of Llandudno. At Llandudno, the granodioritic facies of the Peninsula pluton is exposed as a coarse-grained, cordierite and K-feldspar phenocrystic granite (referred to as Llandudno granodiorite in this study). Layering within the Llandudno granodiorite occurs within several lens-shaped bodies, of which the largest, with a thickness of about 5 m. The studied rhythmic sequence comprises 50 layers with thicknesses ranging between 5 and 50 cm. Each layer is typically characterised by a sharp lower contact overlain by a biotite-rich portion containing approximately 50 vol. % biotite. This mafic lower portion, grades upwards into a strongly leucocratic upper portion enriched in quartz, plagioclase and K-feldspar. The layered rocks host a substantially lower abundance of K-feldspar megacrysts and cordierite than the surrounding Llandudno granodiorite. In particular, cordierite crystals are five times less abundant in the layered zone than the un-layered Llandudno granodiorite. The K-Feldspars megacrysts that do occur within the layering are generally concentrated in the mafic part of some of the layers and are commonly oriented parallel to the layering. The K-feldspar megacrysts are significantly smaller than those in Llandudno granodiorite, but were found to be similar in composition. The whole rock compositions of the layered rocks are peraluminous with A/CNK >1.4. They have lower Mg#s (51 to 58) than the Llandudno granodiorite (51 to 65). The Na2O/CaO ratio of mafic and leucocratic portions is higher than in the Peninsula pluton. The mafic portions show an enrichment of trace and rare earth elements relative to the leucocratic portions and Peninsula pluton. The difference in bulk rock composition as well as trace and rare earth elements composition shows that the layered rocks were not formed by magmas produced by differentiation of the Llandudno granodiorite. The differences in biotite composition in basal sections of adjacent layers suggest that each layer represents a separate magma pulse, with the mafic portions of the layers largely representing an accumulation of the crystals in the magma batch at the time of injection and that these crystals mostly consist of orthopyroxene and biotite. This is in agreement with the findings based on the whole rock chemistry of the layers compared with the Llandudno granodiorite. The inclusion of K-feldspar megacrysts from the host granite into the layering, as well as the exceptional state of preservation of the layering, suggests that the layering formed relatively late in the crystallization sequence of the granite. This, in combination with the evidence for subtle differences in the chemistry of the magmas that formed separate layers, suggests that the layers represent a frozen feeder zone in the Peninsula pluton recording small successive pulses of magma addition. Frequent occurrence of the less-preserved layered biotite schlieren in the pluton may represent the equivalents of the layering that have become partially digested and texturally equilibrated with the host magma. / AFRIKAANSE OPSOMMING: Ritmiese magmatiese gelaagdheid in graniet is ’n interessante kenmerk van plutone met teenstellende chemiese samestellings in ’n wye verskeidenheid tektoniese omgewings. Gelaagde graniet is seldsaam en kom saam met volumetries dominante, niegelaagde gesteentes met soortgelyke samestellings voor. ’n Begrip van die oorsprong van hierdie gelaagdheid in graniet, veral in goed sigbare dagsome, kan belangrike insig verleen in die fisies-chemiese omstandighede en prosesse (soos fraksionele kristallisasie, die grootte en samestelling van individuele volumes magma, die doeltreffendheid van volume-vermenging, ensovoorts) in magmakamers gedurende die vorming daarvan. In die Kaapse granietgroep het ritmiese gelaagdheid besonder goed behoue gebly in dagsome van die S-tipe granodioritiese tot monsogranitiese Skiereilandse plutoon langs die kuslyn van die voorstad Llandudno. Die granodioritiese fasies van die Skiereilandse plutoon by Llandudno manifesteer in die vorm van ’n grofkorrelrige, kordiëritiese en K-veldspatiese fenokris-graniet (wat in hierdie studie Llandudno-granodioriet genoem word). Gelaagdheid in die Llandudno-granodioriet kom voor in verskeie lensvormige massas, waarvan die grootste sowat 5 m dik is. Die bestudeerde ritmiese opeenvolging bestaan uit 50 lae met diktes van tussen 5 cm en 50 cm. Elke laag word gekenmerk deur ’n skerp onderste kontakvlak wat bedek is met ’n biotiet-ryke gedeelte wat uit sowat 50 vol % biotiet bestaan. Hierdie mafiese onderste gedeelte gradeer opwaarts in ’n sterk leukokratiese boonste gedeelte wat ryk is aan kwarts, plagioklaas en K-veldspaat. Die gelaagde gesteentes bevat beduidend minder K-veldspatiese megakriste en kordiëriet as die omliggende Llandudno-granodioriet. Die kordiëritiese kristalle in besonder kom vyf keer minder in die gelaagde sone as in die niegelaagde Llandudno-granodioriet voor. Die K-veldspatiese megakriste wat wél in die gelaagdheid voorkom, is oor die algemeen in die mafiese gedeelte van sommige lae gekonsentreer, en is meestal parallel met die gelaagdheid georiënteerd. Die K-veldspatiese megakriste is aansienlik kleiner as dié in die Llandudno-granodioriet, maar het ’n soortgelyke samestelling. Die heelrots-samestellings van die gelaagde gesteentes is peralumineus, met A/CNK >1,4. Dit toon ook laer Mg#s (51 tot 58) as die Llandudno-granodioriet (51 tot 65). Die Na2O/CaO-verhouding van die mafiese en leukokratiese gedeeltes is hoër as in die Skiereilandse plutoon. In die mafiese gedeeltes is daar ‘n verryking in spoor- en skaarsaarde-elemente relatief tot die leukokratiese gedeeltes sowel as die Skiereilandse plutoon. Die verskil in heelrots-samestelling sowel as spoor- en seldsame-aardelementsamestelling toon dat die gelaagde gesteentes nié gevorm is deur magmas wat uit differensiasie van die Llandudno-granodioriet ontstaan het nie. Uit die verskille in die biotiet-samestelling van basissnitte uit aanliggende lae word afgelei dat elke laag ’n afsonderlike magmapuls verteenwoordig, terwyl die mafiese gedeeltes van die lae hoofsaaklik ‘n versameling van kristalle verteenwoordig wat tydens inplasing van die magma volume gevorm het, meestal ortopirokseen en biotiet. Dít stem ooreen met die bevindinge rakende die heelrots-chemie van die lae in vergelyking met die Llandudno-granodioriet. Die insluiting van K-veldspatiese megakriste vanaf die moedergraniet by die gelaagdheid, dui daarop dat die gelaagdheid betreklik laat in die kristallisasie-orde van die graniet gevorm het. Dít, tesame met bewyse van subtiele verskille in die chemie van die magmas waaruit afsonderlike lae gevorm is, dui daarop dat die lae ’n bevrore toevoersone in die Skiereilandse plutoon uitmaak wat kort, opeenvolgende pulse van magmatoevoeging vasgevang het. Die gereelde voorkoms van swakker bewaarde, gelaagde biotiet-sliere in die plutoon kan moontlik dui op sones van die gelaagdheid wat gedeeltelik verteer is en tekstureel met die moedermagma ge-ekwilbreer het.

Thesis (PhD)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: In recent years acid mine drainage (AMD) has become the focus on many mine sites throughout the world. The Witwatersrand gold mines have been the main focus of AMD in South Africa due to their extensive impact on especially groundwater resources. The Witwatersrand Basin is a regional geological feature containing the world-famous auriferous conglomerate horizons. It is divided into sub-basins and the East Rand Basin is one of them. Due to the regional scale of the East Rand Basin AMD issues, a systems approach is required to provide a useful tool to understand the pollution source term and fate and transport dynamics and to aid in environmental decision making and to evaluate the geochemical impact of mitigation measures and evaluate future scenarios. The numeric geochemical models, using a systems perspective, show that the mine waste facilities, specifically the tailings dams are significant contamination point sources in the East Rand Basin, specifically for acidity (low pH), SO4, Fe, Mn, U, Ni, Co, Al and Zn. When the AMD solution enters the soil beneath the tailings, ferrous and SO4 concentrations remain elevated, while Mn, U, Ni and Co and perhaps other metals are adsorbed. After ~50 years the pollution plume starts to break through the base of the soil profile and the concentration of the adsorbed metals increase in the discharging solution as the adsorption capacity of the soil becomes saturated. The pollution pulse then starts to migrate to the shallow groundwater where contamination of this resource occurs. Toe seepage from the tailings either first reacts with carbonate, where acidity is neutralised to a degree and some metals precipitated from solution, where after it reaches the surface water drainage, such as the Blesbokspruit, where it is diluted. Some evaporation can occur, but evaporation only leads to concentration of acidity and dissolved constituents, thereby effectively worsening the AMD solution quality. The mixing models have shown that the dilution factor is sufficient to mitigate much of the AMD, although seasonal variability in precipitation and evapotranspiration is expected to have some influence on the mixing ratio and some variability in the initial solution will also be reflected in variation in surface water quality. From a sustainability perspective, a basic cost benefit analysis shows that the costs for the operating mine and society in general is lower when mitigation measures are employed during operation. For a theoretical mine in the ERB with an operating life of 100 years, the cost of operational mitigation measures is ~R 31 billion. This value is 4% of turnover and 19% of profits over the time period. Post closure remediation costs are ~R 67 billion. This value is 8% of turnover and 41% of profit over the time period. Although the initial capital investment in mitigation measures is substantial, although some measures will be implemented during operation, it is a smaller percentage of profits than eventual post-closure mitigation. / AFRIKAANSE OPSOMMING: In die laaste paar jaar het suur mynwater loging (SML) die fokus geword by baie myn areas regoor die wêreld. Die Witwatersrand goud myne het die hoof fokus van SML in Suid Afrika geword as gevolg van die groot impak wat Witwatersrand SML op veral grondwater bronne het. Die Witwatersrand Kom is ‘n regionale geologiese struktuur wat die wêreld beroemde goud-draende konglomeraat horisonne bevat. Die Witwatersrand Kom word onderverdeel in sub-komme, waarvan die Oos- Rand Kom een is. As gevolg van die regionale skaal van die Oos-Rand Kom SML probleem sal ‘n sistemiese benadering ‘n bruikbare middel wees om besoedelingsbron migrasie en eindpunt dinamika te verstaan asook om omgewingsbesluitneming en die geochemiese impakte en mitigerings metodieke asook toekoms senarios te evalueer. Die numeriese geochemiese modelle, wat vanuit ‘n sitemiese oogpunt beskou is, wys dat die myn afval fasiliteite, veral die slikdamme in die Oos-Rand Kom, beduidende kontaminasie puntbronne is van veral suur inhoud, SO4, Fe, Mn, U, Ni, Co, Al and Zn. Wanneer die SML oplossing die grond onder die slikdamme binnedring, bly Fe2+ en SO4 konsentrasies hoog, terwyl Mn, U, Ni en Co asook ander metale geadsorbeer word. Na ~50 jaar begin die besoedelingsoplossingspluim deur die basis van die grond profile breek. Die konsentrasie van die geadsorbeerde metale neem dan toe in die oplossing wat deur die grondprofiel beweeg het soos wat die adsorpsie kapasiteit van die grond versadig word. Die besoedelingspuls begin dan stadig na die vlak grondwater akwifeer migreer waar dit dan besoedeling van die grondwater versoorsaak. Water wat van die slikdam basis sypel reageer eers met karbonaat, waar suur tot so ‘n mate geneutraliseer word, dat sommige metale uit die oplossing neerslaan, waarna dit die oppervlak water lope, soos die Blesbokspruit, beryk en verdun word. Verdamping van die SML oplossing kan plaasvind, maar verdamping veroorsaak die konsentrasie van suur en opgeloste stowwe, wat die SML probleem dus kan vererger. Meng modelle het gewys dat die verdunning faktor genoegsaam is om baie van die SML te mitigeer, alhoewel seisoenale fluktuasies in reënval en verdampings syfers kan verwag word om ‘n mate van invoed op die meng verhouding tussen die SML oplossing en oppervlak water te hê wat dan in die oppervlak water kwaliteit gereflekteer sal word. Vanuit ‘n volhoubaarheidsperspektief is ‘n basiese koste voordeels analise gedoen. Hierdie analise het gewys dat die koste van ‘n operasionele myn en die samelewening as geheel laer is wanneer mitigasie metodieke tydens operasie toegepas word. Vir ‘n teoretiese myn in die Oos-Rand Kom met ‘n operasionele lewe van 100 jaar, is die operasionele mitigeringsmetodieke koste ~R31 miljard. Hierdie waarde is 4% van die myn omset en 19% van winste oor die tyds periode. Na operasionele remdiasie kostes is ~R67 miljard. Hierdie syfer is 8% van omset en 41% van winste oor die tydperk. Alhoewel die inisiële kapitaal uitleg van mitigasie metodes groot is, alhoewel sekere metodes eers tydens operasie implimenteer sal word, is die uitleg ‘n kleiner persentasie van winste as wanneer remediasie na operasie implimenteer word.

Thesis (MSc (Geography and Environmental Studies))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: Supervised classifiers are the most popular approach for image classification due to their high accuracies, ease of use and strong theoretical grounding. Their primary disadvantage is the high level of user input required during the creation of the data needed to train the classifier. One alternative to supervised classification is an expert-system rule-based approach where expert knowledge is used to create a set of rules which can be applied to multiple images. This research compared supervised and expert-system rule-based approaches for forest mapping. For this purpose two SPOT 5 images were acquired and atmospherically corrected. Field visits, aerial photography, high resolution imagery and expert forestry knowledge were used for the compilation of the training data and the development of a rule-set. Both approaches were evaluated in an object-orientated environment. It was found that the accuracy of the resulting maps was equivalent, with both techniques returning an overall classification accuracy of 90%. This suggests that cost-effectiveness is the decisive factor for determining which method is superior. Although the development of the rule-set was time-consuming and challenging, it did not require any training data. In contrast, the supervised approach required a large number of training areas for each image classified, which was time-consuming and costly. Significantly more training areas will be required when the technique is applied to large areas, especially when multiple images are used. It was concluded that the rule-set is more cost-effective when applied at regional scale, but it is not viable for mapping small areas. / AFRIKAANSE OPSOMMING: Gerigte klassifiseerders is die gewildste benadering tot beeldklassifikasie as gevolg van hulle hoë graad van akkuraatheid, maklike aanwending en kragtige teoretiese fundering. Die primere nadeel van gerigte klassifikasie is die hoë vlak van gebruikersinsette wat benodig word tydens die skepping van opleidingsdata. 'n Alternatief vir gerigte klassifikasie is 'n deskundige stelsel waarin ‘n reëlgebaseerde benadering gevolg word om deskundige kennis aan te wend vir die opstel van 'n stel reëls wat op meervoudige beelde toegepas kan word. Hierdie navorsing het gerigte en deskundige stelsel benaderings toegepas vir bosboukartering om die twee benaderings met mekaar te vergelyk. Vir dié doel is twee SPOT 5 beelde verkry en atmosferies gekorrigeer. Veldbesoeke, lugfotografie, hoë-resolusie beelde en deskundige bosboukennis is aangewend om opleidingsdata saam te stel en die stel reëls te ontwikkel. Beide benaderings is in 'n objekgeoriënteerde omgewing beoordeel. Die akkuraatheidsvlakke van die resulterende kaarte was ewe hoog vir beide tegnieke met 'n algehele klassifikasie-akkuraatheid van 90%. Dit wil dus voorkom asof koste-effektiwiteit eerder as akkuraatheid die deurslaggewende faktor is om te bepaal watter metode die beste is. Alhoewel die ontwikkeling van die stel reëls tydrowend en uitdagend was, het dit geen opleidingsdata vereis nie. In teenstelling hiermee is 'n groot aantal opleidingsgebiede geskep vir elke beeld wat met gerigte klassifikasie verwerk is – 'n tydrowende en duur opsie. Dit is duidelik dat meer opleidingsgebiede benodig sal word wanneer die tegniek op groot gebiede toegepas word, veral omdat meervoudige beelde gebruik sal word. Gevolglik sal die stel reëls meer kosteeffektief wees wanneer dit op streekskaal toegepas word. ‘n Deskundige stelsel benadering is egter nie lewensvatbaar vir die kartering van klein gebiede nie.

Thesis (PhD (Earth Sciences))--University of Stellenbosch, 2008. / Tanzanite, a blue/violet gemstone variety of zoisite (Ca2Al2O.AlOH[Si2O7][SiO4]) is only produced in NE Tanzania. The only known locality is an approximately 7km2 deposit in the Merelani area. It is one of the most sought after gemstones in the world with an industry sales value of between hundred and fifty and two hundred million dollars per year. At the current production rates and estimated resources the tanzanite deposit has a life expectancy of around 20 years. Despite the economical and scientific importance as well as the geological uniqueness of the deposit very little research has been conducted on the geology and petrology of the deposit and the characteristics of tanzanite. The primary aim of the research summarised in this dissertation was to gain an understanding of the geological conditions that led to the formation of this unique variety of zoisite. In order to achieve this, a variety of geological disciplines were addressed including the lithostratigraphic setting, the deformational history, the metamorphic history and conditions, the geochemical and isotopic composition, the mineral chemistry as well as the physical and optical properties of the tanzanite. Extensive field work was conducted over a seven year period, which included surface and underground mapping, surface trenching, surface and underground core drilling, structural measurements and an intensive sampling programme. Various analytical techniques were used in order to petrologically and mineralogically investigate both the deposit and tanzanite itself, including optical microscopy, XRF analyses, laser-ablation ICP-MS, quantitative chemical analyses by means of the electron microprobe, XRD analyses, back-scattered electron microscopy, isotope analyses, fluid-inclusion studies, Vis/UV/NIR spectroscopy, IR spectroscopy, and EPR studies. This study revealed an intricate and complex geological history for the formation of tanzanite. The deposition of carbon-rich layers, formed during the development of a sequence of shallow shelf sediments consisting primarily of various organic carbon (δ13C between –22,85 ‰ and –26,74 ‰) -rich mudstone horizons and limestone beds is seen as the first phase of the mineralisation process. These Archaean sediments were deposited in a back- or fore-arc spreading basin. The organic carbon-rich mudstone layers acted as the first phase of vanadium accumulation in the stratigraphic sequence, and are intercalated with a material with a volcanic origin (metabasites), which most likely contributed to the enrichment of amongst others V. Early diagenetic processes were followed by an extended metamorphic and deformation (D1) history, reaching upper granulite facies conditions (10 – 12 kbar and 850 oC to 1000 oC) at ca. 1000 Ma. The high-grade metamorphic history was followed by multiphase retrograde deformation events (D2 and D3) that developed as a result of crustal uplift. The D2 event probably occurred between 850 and 600 Ma at P-T conditions estimated at between 8 to 7 kbar and 700 to 650 oC. A third stage of deformation (D3) resulted in the formation of overprinting structures and occurred during the later stages of the Pan-African (550 – 500Ma) event during amphibolite/greenschist facies conditions (7 to 6 kbar and 600 to 520 oC). The deformational history of the deposit played a critical role in the mineralisation processes. Of the four main deformation events identified, three played a critical role in the mineralisation process. The first event led to the development of early structural features as defined by F1, S1 and L1. This was followed by a multiphase D2 event consisting of three different orders of folding (F2a, F2b and F2c) and the formation of boudinage. The association between tanzanite and boudins as well as the stacking and multiple duplication of the boudinaged ore-zone through isoclinal folding resulted in multiple “ore-shoots”. These ore-shoots follow the plunge of the F2c fold closures and results in mine-able features within the ore-body. The third deformation event led to complex structural overprinting of the earlier fabric as observed in S3 and L3 and resulted in the deformation of the F2 structures through crosscutting F3 folds. Calc-silicate layers developed in the stratigraphic sequence as a result of metamorphic and metasomatic interaction between calcium enriched (boudinaged calc-silicates) and depleted horizons (graphitic gneisses) during a skarn-forming episode. V-rich green grossular garnet (tsavorite) crystallised in tension zones within and in proximity to the boudins during prograde metamorphism. Tanzanite mineralisation occurred during the retrograde stages at ca. 585 ± 28 Ma with P-T conditions estimated at ca. 5 to 6 kbar and 650 ± 50 oC. Two distinctive tanzanite-forming processes are distinguished. The first involves the formation of tanzanite as a result of retrograde reaction of grossular garnet. The second process involves the migration of V and Ca -enriched fluids along brittle shear zones to tension sites where fluids reacted with wall rock during a drop in P-T conditions to precipitate tanzanite. Fluid inclusion and stable-isotope studies concluded that the ore-forming fluids were derived from the dehydration of the metasedimentary sequence and consisted of a mixture of H2O, CH4, H2S and N2. Mineralogical investigation of tanzanite indicated that trace concentrations of vanadium within its crystals structure causes its blue / violet colour. It was proved that the vanadium originated from the abundant organically derived graphite within the deposit. Spectroscopic and EPR analyses revealed the importance of the Ti4+ / Ti3+ ratio within the crystal structure of tanzanite with regard to its colour characteristics. The heating of tanzanite results in a couple valence exchange reaction Ti 3+ + V 4+ → Ti 4+ + V 3+ which causes an increase the blue / violet colour of tanzanite. The research conducted led to the development of a successful geological model for the tanzanite mining and treatment activities in the Merelani area of NE Tanzania. As such the research contributed to the establishment of a successful tanzanite mine, based on sound geological principles, which may act as a role model for other gemstone mines worldwide.

Thesis (PhD (Earth Sciences))--University of Stellenbosch, 2009. / Fluid-present partial melting has generally been regarded a poor candidate for effecting crustal differentiation. In this study I report on anatectic metasediments from the Pan-African Damara Belt in Namibia that have undergone fluid-present biotite melting at a relatively low temperature, yet appear to have lost a significant volume of melt. In situ anatectic features have been identified on the basis of the existence of new generations of cordierite and/or garnet produced as the solid products of incongruent anatexis within or adjacent to leucosomes, that most commonly occur as lens shaped pods at a high angle to the lineation and formed during extension in a direction parallel to the long axis of the orogeny. Within these sites biotite underwent incongruent melting via the reaction Bt + Qtz + Pl + H2O = Melt + Grt + Crd. Cordierite nucleated on preexisting crystals within the bounding gneiss; garnet nucleated within the fracture sites (leucosomes) and typically occurs as individual, large (50 to 120 mm in diameter) poikiloblastic crystals. Thermobarometry applied to the anatectic assemblage yields low-temperature, granulitefacies peak conditions of 750 °C, 0.5 GPa. This temperature is approximately 100 °C lower than the accepted conditions for the onset of fluid-absent biotite melting. This, coupled to the focussing of anatexis on extensional fractures, suggests that anatexis occurred through waterpresent biotite incongruent melting. In order to better understand this process, both fluid-absent and water present partial melting experiments were conducted within the temperature interval 700 to 900 °C at 0.7 GPa. In the fluid-absent experiments, biotite incongruent melting started between 800 and 850 °C to produce melt coexisting with peritectic garnet and cordierite. In contrast, in water-saturated experiments, biotite melted via the reaction Bt + Pl + Q + H2O = Grt + Crd + Melt, between 700 and 750 °C, to produce melt, cordierite and garnet in the proportions 73:24:3.