School of Earth Sciences - Theses
Permanent URI for this collection
Search Results
1 - 2 of 2
-
ItemThe geochemistry and petrogenesis of MARID and PIC xenoliths( 2019)MARID (Mica-Amphibole-Rutile-Ilmenite-Diopside) and PIC (Phlogopite-Ilmenite-Clinopyroxene) rocks are rare mantle samples, most often occurring as xenoliths in kimberlites. Both MARID and PIC xenoliths contain high modal abundances of phlogopite, which does not occur in the “depleted mantle” – i.e. the residue that represents the protolith of Archean mantle roots, complementary to the formation of continental crust. Consequently, these rocks are frequently interpreted as the products of extreme mantle metasomatism. Previous studies of these rocks have focussed on acquiring isolated datasets on small sample suites. This approach has made it difficult to explore possible relationships within the spectrum of metasomatised, phlogopite-rich xenoliths, even from a single locality. It therefore remains unclear how MARID and PIC rocks are formed (whether by metasomatism of peridotites, or by melt crystallisation in mantle veins). The nature of MARID parental melts/fluids also remains unidentified. Finally, MARID geochemical compositions (e.g., high K2O content) and “enriched mantle” radiogenic isotope signatures are often cited as evidence for their presence in the source of mantle-derived potassic magmas, such as lamproites. This study seeks a greater understanding of the causes and effects of mantle metasomatism by studying the petrographic, geochemical, and isotopic (radiogenic Sr-Nd-Hf-Pb and stable N-O-S-Tl) features of 26 MARID and PIC samples from southern African kimberlites (Kimberley) and orangeites (Newlands). Previous inferences of a genetic link between PIC rocks and kimberlite melts are substantiated by clinopyroxene trace element compositions in equilibrium with kimberlite melts, and the similar radiogenic isotopic compositions of PIC clinopyroxene (87Sr/86Sri = 0.7037–0.7041) and southern African Cretaceous kimberlites (87Sr/86Sri = 0.7032–0.7048). Based on petrographic observations of kimberlite groundmass minerals (e.g., calcite, perovskite) as inclusions in secondary clinopyroxene rims, PIC rocks must be the result of kimberlite metasomatism, and not the melt source for kimberlite magmas. Furthermore, PIC rocks appear to be closely related to some phlogopite-bearing peridotites and wehrlites from the lithospheric mantle beneath Kimberley, due to similarities in their mineral geochemical compositions. It can therefore be inferred that these lithologies are the result of variable degrees of interaction between kimberlite melts and the lithospheric mantle. This study also presents mineralogical, geochemical, and isotopic evidence that kimberlites infiltrate and affect the mineral and bulk-rock compositions of the MARID rocks that they host, complicating interpretations of MARID genesis. By filtering out such contamination effects, new interpretations regarding these rocks can be reached. Although many studies support a magmatic genesis, the ranges in primary MARID mineral major element compositions presented in this thesis overlap with those in extremely metasomatised peridotites, opening the possibility of MARID genesis by metasomatic overprinting of peridotite protoliths. Some lherzolites studied herein have similar geochemical and isotopic signatures to MARIDs, suggestive of a common parental metasomatic fluid. Prior to kimberlite infiltration, MARID rocks appear to have had “enriched mantle” radiogenic isotopic signatures, which may indicate that the source of MARID parental melts contained recycled material – comprising both oceanic and continental crust, based on stable isotope considerations. Finally, bulk-rock reconstructions of MARID compositions were conducted to eliminate secondary features, from which the trace element composition of MARID-derived melts was modelled. These melts closely match the compositions of mantle-derived magmas, such as orangeites, lamproites, and some ultramafic lamprophyres, suggesting that MARID-veined mantle could indeed be the source of alkaline ultramafic magmas in intracontinental settings.
-
ItemIsotopic disequilibrium in granitic systems: the origins of heterogeneity in granites and implications for partial melting in the crust and petrogenetic models( 2017)Unravelling the processes involved in granite magmatism is essential to our understanding of the continental crust, its formation and evolution. Radiogenic isotope systems are commonly employed to this end, but the behaviour of these systems may not be as simple as is often assumed. Understanding the causes of isotopic complexity in granite suites is the aim of this research. By modelling the disequilibrium partial melting of isotopically heterogeneous protoliths the different compositions of the melt, source and restite for a range of hypothetical scenarios have been calculated. Results demonstrate that the melt produced may have Sr, Nd and Hf isotope compositions distinct from both the protolith and restite. A corollary is that restite-bearing magmas may exhibit different isotope compositions than their melts, a feature which should be preserved as a difference between the Hf isotope compositions of bulk-rock samples and their magmatic zircon populations. The same modelling also suggests that a single source rock can produce melts with diverse isotope compositions. The predictions of this modelling have been tested by analysing S- and I-type granites from the Lachlan Fold Belt, southeastern Australia, including iconic examples of restite-bearing rocks. Comparisons of Hf isotope compositions between bulk-rocks and their magmatic zircons reveal discrepancies (ΔεHfbulk-zircon) ranging from -0.6 to +2.5 ε units for I-type granites. This intra-sample Hf isotopic heterogeneity is interpreted to represent disequilibrium between the melt and restite assemblage. The ΔεHfbulk-zircon values are consistent with calculated ΔεHfmagma-melt values (from -4.2 to +7.4) based on the disequilibrium amphibole dehydration melting of 0.5-1.0 Ga meta-igneous protoliths. S-type granites also record differences between their bulk-rock and magmatic zircon Hf isotope compositions; however, the disparity is more subtle. Both positive and small negative ΔεHfbulk-zircon values are observed, consistent with modelling the partial melting of isotopically heterogeneous meta-sedimentary protoliths. In addition to low-temperature granites, case studies of two high-temperature I-type granitoid suites (Boggy Plain and Wallundry) have also been conducted. Both display a weak coupling between geochemical parameters that have been interpreted previously to indicate the involvement of assimilation and fractional crystallisation (AFC) processes. Positive ΔεHfbulk-zircon values obtained in the Boggy Plain Suite support the existing petrogenetic model in which basaltic melt becomes variously contaminated by material derived from the continental crust. The positive value is explained by retention of earlier-crystallised, more radiogenic phases in isotopically evolved, more felsic samples. In contrast, the Wallundry Suite is characterised by negative ΔεHfbulk-zircon values caused by the presence of unmelted components of its contaminant. A complex interplay of contamination, crystallisation, melt segregation and interaction between magma batches is required to account for the Wallundry Suite isotope data. The results of this study indicate that disequilibrium partial melting can produce within-suite isotopic variability without recourse to assimilation or mixing processes (1) in mafic to felsic samples caused by the progressive separation of melt from its isotopically distinct restite assemblage; and (2) via the extraction of multiple batches of isotopically distinct melts produced from a single source as anatexis proceeds. Furthermore, the isotope variation resulting from restite unmixing may be distinguished from magma mixing by decoupling of the Rb-Sr, Sm-Nd and Lu-Hf isotope systems. Importantly, the isotopic discrepancy between bulk-rock granite samples and their magmatic zircon populations suggests that the most mafic bulk-rock granite samples of a given suite, not magmatic zircon, provide the most accurate estimate of source rock Hf isotope compositions. This raises concerns regarding the ubiquitous use of zircon Hf isotope data to constrain crustal growth models.