Browsing by Author "Otto, Tahnee"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    A phase equilibrium modelling investigation of the consequences of entrainment of components of the source on the crystallisation of mantle-derived magmas in the upper crust, with specific relevance to the petrogenesis of chromitite layers in the Rustenburg Layered Suite of the Bushveld Complex, South Africa
    (Stellenbosch : Stellenbosch University, 2023-03) Otto, Tahnee; Stevens, Gary; Moyen, Jean-Francois; Mayne, Matt; Stellenbosch University. Faculty of Science. Dept. of Earth Sciences.
    ENGLISH ABSTRACT: Thermodynamic modelling techniques offer several advantages over experimental studies in investigations of the partial melting of mafic and ultramafic rocks under upper mantle conditions, but may not be as reliable as experimental studies in accurately predicting rock behaviour. The two main thermodynamic datasets that are in common use for such thermodynamic investigations have different levels of coverage. The Berman (1988) dataset is utilised by the MELTS family of software (Asimow & Ghiorso, 1998; Ghiorso & Sack, 1995; Ghiorso et al., 2002), and along with routines for handling activity-composition relations for solid and liquid solution phases, is embedded within the software. In contrast, the Holland & Powell (2011) dataset is used by a wide range of different software with different computational strategies. Activity-composition relations are handled by published models that are selectable and transparent. One of the programs using the Holland & Powell (2011) dataset, Rcrust (Mayne et al., 2016), has been developed to allow modelling of phase stabilities with changing bulk composition, which makes it particularly powerful for studying processes involving fractionation. In this study, the results of sets of experiments on ultramafic and mafic compositions under upper mantle conditions were compared with outputs from MELTS and Rcrust. Differences between the modelling and experimental results were quantified, and the Holland & Powell (2011) dataset combined with appropriately selected activity-composition models produced the best match with experimental results when melt fraction, melt composition, and the assemblage of coexisting phases were considered. The results demonstrate that modelling is a viable alternative to experimental techniques in investigating melting of the upper mantle. Thermodynamic modelling was used to map the envelope of uncertainty that exists around experimental results due to often unavoidable experimental pitfalls, and demonstrated that factoring in these uncertainties produced a better fit between modelling and experimental results. Next, this study developed a thermodynamic modelling methodology to investigate the stoichiometry of the melting reactions in an eclogite and peridotite as a function of pressure, temperature, and bulk compositional change due to melt loss. Pressure-temperature regions where the partial melting reactions produce peritectic crystals that increased in abundance with melt were identified. The entrainment of these crystals to the magma on segregation produces distinct shifts in the chemistry of the extracted magmas, for both sources. The results illustrate that variable amounts of peritectic crystal entrainment represents a previously unrecognised source of compositional heterogeneity in basaltic and intermediate magmas. The modelled results highlight how peritectic crystal entrainment may be a principal mechanism for creating the magmas that produce layered mafic and ultramafic complexes, because the entrained phases react with melt on ascent to produce magmas with a high crystal load once they reach the upper crust. During the partial melting of peridotite, the most common entrainable peritectic mineral is orthopyroxene, which is Cr2O3- bearing. On ascent, entrained peritectic orthopyroxene reacts to form olivine and chromite, and such magmas will enter high-level magma chambers as a melt-olivine-chromite slurry. Density segregation on intrusion allows the formation of magmatic layering if the magma body has appropriate geometry. Thus, these results provide an explanation for the presence of chromitite seams in layered mafic complexes, as well as average bulk rock Cr2O3 contents well above mantle values in some layered intrusions, such as the Bushveld Complex. Chromitite layers are a consequence of peritectic orthopyroxene entrainment and do not require any complex phase equilibrium magma chamber processes.