What garnet, clinopyroxene, and diamond potential can tell us about the evolution of sub-cratonic mantle sections: a case study of the Zagadochnaya kimberlite (Yakutia)

This PhD provides major and trace element compositions for a new suite of Cr-rich garnet xenocrysts and associated minerals from the diamond-free Zagadochnaya kimberlite, Daldyn Field, Yakutia (Russia). Interpreting the nature and evolution of the underlying lithospheric mantle from these samples requires a good understanding of relationships between pressures and temperatures of formation, upper mantle phase relations and trace element signatures of petrochemical processes. I addressed these issues by integrating three main lines of research. i) Evaluation of single-clinopyroxene geobarometry for garnet peridotites. Single-clinopyroxene thermobarometry represents the most reliable method among single-mineral thermobarometric techniques for mantle rocks, but the geobarometer tends to produce considerable scatter in P–T estimates when applied to clinopyroxenes with unfavourable compositions. Multiple electron microprobe analyses on compositionally diverse clinopyroxenes, using different analytical conditions, demonstrated that this scatter is mostly related to propagation of analytical errors on the calculated Cr-in-cpx pressure. The results of this analytical tests were used to calculate model analytical errors and propagated P uncertainties for a large set of published analyses of mantle-derived, xenolith-borne clinopyroxenes. I found that the parameter aCr/Cr# [where aCr = Cr–0.81•Na•Cr#, Cr#=Cr/(Cr+Al), atoms per 6-oxygen formula unit] can be used to discriminate clinopyroxenes for which analytical errors alone will propagate unacceptable P uncertainties (i.e., higher than ±0.25 GPa) for several combinations of analytical conditions. I therefore defined a new optimized analytical procedure for single-clinopyroxene geobarometry, which significantly decreases the pressure uncertainties and allows a better definition of clinopyroxene-based geotherms. ii) Thermodynamic modeling of natural peridotitic systems. Calculations of phase equilibria in Cr-bearing peridotitic systems were used to predict the effects of P, T and bulk compositional variations on garnet–spinel relations in fertile and depleted mantle compositions. Calculations showed that in cratonic lithospheric sections the width of the garnet–spinel transition strongly depends on bulk composition: in a fertile mantle, spinel can coexist with garnet to about 120 km depth, while in an ultra-depleted harzburgitic mantle it can be stable to over 180 km depth. In the garnet+spinel stability field the calculated modes of spinel are very low (0.1–2.8 %), suggesting that spinel grains may be easily overlooked in mantle xenoliths. The model also suggests a significant potential role of P–T conditions on the distribution of garnets in the popular Ca–Cr discrimination diagram. iii) Interpretation of the mantle record in the Zagadochnaya kimberlite. A suite of 28 garnet xenocrysts, often containing polymineralic inclusions, were selected from a heavy-mineral concentrate of the Zagadochnaya kimberlite for detailed electron microprobe analysis and laser-ablation inductively-coupled mass spectrometry. Robust P–T estimates for the garnets and associated clinopyroxenes and the Ca and Cr contents in the garnets indicate that the kimberlite sampled only a shallow mantle portion ( 130 km, the absence of diamond, the Mg-rich composition of the kimberlite, and the pervasive reactions that produced the secondary assemblages