D.Phil. (Geology) / Please refer to full text to view abstract

Ultramafic rocks – that are usually located deep below the Earth's surface – are occasionally exhumed by the motion of tectonic plates. The massive chemical disequilibrium that exists between these exposed rocks and the surface waters and atmosphere leads to geologically rapid reactions that consume water and CO₂, binding them to form secondary hydrated/carbonated solid minerals that are found extensively in continental exposures (ophiolites) and at the seafloor near mid-ocean ridges. Pervasive fracturing and faulting in oceanic lithosphere generates pathways for fluids to access and react with rocks that are in some cases located down to depths of tens of kilometers. Over time, the large volumes of fluids and volatiles that are bound up in crustal and upper mantle rocks via such reactions are eventually subducted to extreme depths where subsequent fluid release can trigger melting, arc volcanism and seismic activity. In addition to their geophysical importance, these reactions are also considered to be critical for the survival of organisms in deep sea hydrothermal systems, and a potential source in the origin of life hypothesis. The natural transfer of atmospheric CO₂ to stable, solid carbonate minerals has, in recent years, motivated a large research effort towards investigating its potential as a large-scale carbon sequestration alternative. Understanding the geophysical impact and environmental potential of these reactions and their related processes requires knowledge of their basic physical and chemical behavior. Because of the difficulties of observing these processes in real-time, either experimentally or in the field, there has been a heavy reliance on hypothetical arguments that have been driven by observations in natural rocks. The observations paint a very complex picture – involving an interplay between reaction, fluid flow and fracturing – that is not easily explained by simple model descriptions. Although there has been increasing interest in modeling this class of problems in recent years, to date there remains a considerable gap between the theory and computational framework that is required for a consistent model description. A major theme in said models is their omission of poro-mechanical effects and complications arising from clogging of pore space with precipitating minerals. Both of these are necessary ingredients for a consistent model; however, they require a more complex description that is based on coupled multiphase continuum mechanics, reactive transport, and potentially brittle failure. Each of these components is a technical challenge in its own right, requiring development of novel theory and computation that integrates them in a suitable manner. The overall goals and themes of this thesis are aimed at closing this gap. To this end, I develop a modeling framework and computational tools that are capable of describing reactive flow in brittle media, with a specific focus on fluid-mineral reactions in near-surface ultramafic rock environments. The exposition of this framework is split into 3 separate chapters that build on one other in increments of complexity. Specifically, Chapter 1 presents a poromechanics-based description of coupled fluid flow, mass transfer and solid deformation for a simplified hydration reaction. This model is extended in Chapter 2 to incorporate cracking by adopting modern developments in computational fracture mechanics. Finally, in Chapter 3 I extend the set of reactions to support mixed H₂O-CO₂ fluids by leveraging recently developed tools in computational thermodynamics. Along the way I present a number of numerical model simulations that develop intuition and draw comparisons with natural observations, whilst remaining mindful of its limitations and areas for improvement. Overall, this work represents progress towards better understanding of physical and chemical feedbacks of reactive-brittle processes in the Earth's near-surface and the potential for large-scale carbon sequestration.

From the abstract: In the layered sequence of the Bushveld Complex a number of distinct, but possibly genetically related groups of transgressive, postcumulus, ultramafic and mafic rock are recognised. The main part of this thesis investigates a suite of postcumulus rocks for which the name iron-rich ultramafic pegmatite is proposed. The majority of iron-rich ultramafic pegmatite bodies examined are from the upper critical zone of the layered sequence at Rustenburg Platinum Mines Amandelbult Section, in the northern sector of the western Bushveld Complex. Field relationships imply that the iron-rich ultramafic pegmatites should be considered as an integral feature of the layered sequence, even though they transgress the cumulates. Consequently, this thesis also includes a study of the cumulate sequence at Amandelbult. A second group of postcumulus, ultramafic rocks which is investigated comprises latiniferous ultramafic pipes; the Driekop pipe has been selected as a case­ study. This thesis is presented in four sections, namely, an introduction and verview, and studies on the Driekop pipe, the cumulate sequence at mandelbult and the iron-rich ultramafic pegmatite suite. A new classification scheme of discordant bodies of postcumulus, ultramafic rock in he Bushveld Complex is proposed (see also Viljoen & Scoon, in press). In he scheme presented here, two main varieties of postcumulus, ultramafic rock re recognised, namely, non-platiniferous magnesian dunites and iron-rich ltramafic pegmatites.

The Townlands iron-rich ultramafic pegmatite is a relatively large pipelike body situated in the western corner of Rustenburg Section, Rustenburg Platinum Mines. It is characterised by a strong negative magnetic signature and transgresses the noritic layered sequence of the upper critical zone of the Bushveld Complex. The layered rocks are downwarped in the vicinity of the pipe and are in sharp contact with the pegmatitic material. The pegmatite varies in composition between dunite and wehrlite, with the marginal zones being more wehrlitic in composition. Olivine (Fo₃₀ - Fo₅₂) and clinopyroxene (Wo₄₅En₃₀Fs₂₅ - Wo₄₅En₃₇Fs₁₈) are the dominant constituents and accessory phases include ilmenite, Ti -magnetite, apatite, amphiboles, chlorite-group minerals, biotite, ilvaite and a host of unusual ore minerals. The Fe-Ti oxides exhibit exsolution textures typically found in slowly cooled igneous rocks and temperatures of formati on are consi dered to be in excess of 800°C. The UG2 chromitite leader layers intersected by borehole TLP.l are enriched in Fe and Ti and exhibit compositions intermediate between chromite and Ti-magnetite. The ore mineral assemblage includes a primary sulphide assemblage consisting of troilite, chalcopyrite, cubanite and pentlandite, and an array of unusual phases formed by late-stage secondary processes. The unusual sulphides mooihoekite and haycockite, that occur in certain parts of the pegmatite, are considered to have formed by partial replacement of the primary assemblage and a possible paragenetic sequence is discussed. Mineral compositions and whole rock geochemical data are consistent with an origin for the pegmatite by crystallization from a fractionated melt. It is suggested that intercumulus fluids, trapped during the crystallization of the noritic layered sequence, accumulated in an area of structural weakness, in response to an increasing overburden pressure and/or tectonic activity. Evidence is also presented that indicates that the Townlands pegmatite may consist of at least two separate, but adjoining pegmatite bodies. / KMBT_363 / Adobe Acrobat 9.54 Paper Capture Plug-in

The purpose of this study is to characterize the metasomatism that has taken place as a result of the chemical incompatibility between mafic and metaultramafic bulk compositions during high-grade regional metamorphism in the Tobacco Root Mountains, southwest Montana. Metasomatism of these rocks took place by both diffusionand infiltration-dominated processes. The result of these processes are characterized mineralogically and geochemically in the rocks.

Les terrains ultramafiques de Nouvelle-Calédonie, riches en Ni, Co, Mn et Cr sont exploités par des mines à ciel ouvert, ce qui génère l'émission de poussières riches en métaux. L'objectif de ce travail est de développer des approches pour estimer le risque environnemental lié aux poussières émises par les mines à ciel ouvert et les usines de traitement du minerai de nickel. L'estimation de la fraction biodisponible des métaux contenus dans deux fractions granulométriques de poussières, celle inférieure à 100 ~m mobilisable par le vent (F<1 OO~m.) et celle susceptible de pénétrer le système respiratoire (PM1 0), a été réalisée par extractions cinétiques à I'EDTA. L'obtention des PM10 a nécessité la mise au point d'une technique de tri par transport des particules dans un tube horizontal grâce à un flux d'azote. Les extractions cinétiques ont permis de discrimtner trois fractions de métaux: rapidement extraite, lentement extraite et non biodisponibles. Les concentrations en métaux potentiellement biodisponibles sont toujours très élevées et la fraction lentement extraite est toujours la plus concentrée. Pour F<1 00 ~m. les constantes cinétiques de la fraction lentement extraite sont plus faibles pour les poussières de sols miniers que celles de sols forestiers. Les poussières issues de sols miniers seraient alors un réservoir en éléments métalliques biodisponibles à plus long terme. La bioindication lichénique avec traitement en données compositionnelles des concentrations en métaux permet de définir un indice de dispersion des poussières. Cette méthodologie pourrait appuyer les réseaux de surveillance de la qualité de l'air en Nouvelle-Calédonie. / Bioavailability estimation of nickel, cobalt and manganese in dust from ultramafic soils likely to be mobilized by wind and~eve lopment of a bioindication tool using lichen for dust emitted by mining activities in New Caledonia New Caledonian altered ultramafic soils, particularly rich in Ni, Co, Mn and Cr, are extracted by opencast mines which generale dust rich in metals. The objective of th is work is to develop approaches for environmental risk assessment of dust emitted by opencast mines and nickel ore metallurgical plants. The assessmentof metals' bioavailable fraction from two dust granulometrie size fractions, one less than 100 IJm which is mobilizable by wind (F<1001Jm,) and another one able to penetrate the respiratory system (PM 1 0), has been determined by kinetic extraction with EDT A. The development of a new separation deviee based on particle transport subjected to a nitrogen flux in a horizontal tube has been necessary for PM1 0 segregation. Kinetic extractions le ad to the distinction of th ree metal pools: rapidly labile, less rapidly labile and non-bioavailable. Trace metal potentially bioavailable concentrations were always high and the less rapidly labile pool is always the most concentrated pool. Concerning F<1 001Jm, the less rapidly kinetic constant of the less rapidly labile pool is weaker for mining soils than forest soils. F<1001Jm fractions from mining soils representa more durable reserve in trace metal than the same fraction from forest soils. Bioindication using lichens with compositional data analysis of their metal concentration allow defining an indicator of emission dispersion. This methodology could support air quality monitoring networks in New Caledonia.