An experimental study into the behaviour of Titanium-rich and Chromium-rich oxide mineral phases at high pressures and temperatures

The composition of fluid phases in the Earth's mantle are significantly affected by oxygen fugacity, and mantle fluids can be reasonably modelled by the C-H-0 system. Using a broad generalization, under oxidized conditions (at -FMQ) the fluid phase is dominated by C02-H20-rich fluids, under relatively reducing conditions (at -FMQ-2-3 log units) by Cf4-H20-rich fluids, and under strongly reducing conditions (below IW) by CH4 and H2. Thus by controlling fluid speciation in the Earth's mantle, oxygen fugacity influences the P-T position of a mantle solidus and the composition of mantle-derived melts. Oxygen fugacity also constrains the oxidation state of elements and the stability of minerals. For example, natural . diamonds and natural fluids approaching H20 are at equilibrium in the mantle under reduced conditions at -IW+l-2 log units. The Fe3+f2.Fe ratios of mantle-derived oxide minerals, such as Cr-rich spinel and Mg-rich ilmenite, in equilibrium with silicates can be used to indicate the oxygen fugacity (f02) under which mantle derived mineral assemblages formed. The spinel-based oxygen geobarometer has received a significant amount of attention and appears to be experimentally well constrained for all but the most Cr-rich spinel assemblages. The ilmenite-based oxygen geobarometer, however, was fomiulated on the basis of existing thermodynamic data and has not been experimentally tested. An important prerequisite that needs to be demonstrated when applying oxide mineralbased geothermometers and oxygen geobarometers to natural samples with confidence is mineral stoichiometry of the oxides and a knowledge of their ferric and ferrous iron contents. The availability of a synthetic multilayer crystal and accurately calibrated oxide and silicate standards make it possible to use the electron microprobe for precise oxygen analyses of spinels. A requirement of the oxygen measurement routine described is the use of repetitive statistical analyses of the oxygen standards and subsequent corrections and recalibration. A representative set of oxygen analyses for each spinel population studied is essential to obtain reliable data, and the danger of using single data is emphasized. Magnesium-rich kimberlitic ilmenites have been analysed for oxygen content and, within the accuracy of the technique (± 2 rel. % ), are stoichiometric. Magnesiochromite spinels, having broad compositional similarities but different petrogenetic and cooling histories, have been analysed for oxygen and their stoichiometry has been assessed. Diamond-indicator spinels from the Aries, Roberts Victor and Makganyene kimberlites and Argyle lamproite are stoichiometric. Spinel inclusions in olivine phenocrysts from low-Ti tholeiites from the Hunter Fraction Zone and high-Ca boninites from the Tonga Trench show a range of non-stoichiometry. Low Fe3+fi,Fe values calculated assuming stoichiometry for such spinels are invalid. Spinels from metamorphosed volcanics from the Peak Hill-Glengarry Basin and the Heazlewood River Ultramafic Complex are also non-stoichoimetric, having significant Feg1304 - Crs1304 components. The results demonstrate that high-chromium spinel non-stoichiometry is a common feature. This has important implications when using ferric and ferrous iron concentrations to estimate the oxidation state or temperature of formation. In the second section of this study, synthetic ilmenite-olivine-rutile±orthopyroxene assemblages were equilibrated between 900 and 1200 oc and 15 to 35 kbar, under controlled oxygen fugacity. Oxygen fugacity was controlled using the double capsule technique with inner sample capsules consisting of graphite, olivine, platinum or AgPd alloy. The outer capsule was Pt or AgPd alloy, and the oxygen buffers used in the experiments were WC-W02-graphite (WCWO), graphite-water/graphite-C02-CO (GW/GCO), Ni-NiO (NNO), and Fe203-Fe304 (HM). Experiments were performed with a 0.5 inch piston-cylinder apparatus under fluid-excess conditions. Within the P-T -X range of the experiments, ilmenite compositions showed redox ratios, Fe3+fi,Fe, were a linear function of oxygen fugacity (0.03 at WCWO, 0.09 at GW/GCO, 0.20 at NNO and 0.73 at HM). The redox ratio of ilmenites at high oxidation conditions is found to be sensitive to temperature, and Fe3+f:LFe decreases with increasing temperature at HM buffer conditions. This temperature dependence is less noticeable under more reducing conditions. Similarly, rutile showed increasing iron contents with increasing oxygen fugacity. Investigations demonstrated that the major substitution was (Fe3+Q.OH) and iron contents (expressed as Fe203) varied linearly with oxygen fugacity from 1.6 wt.% at WCWO, 2.3 wt.% at GW/GCO, 3.9 wt.% at NNO and 6.7 wt.% at HM). The Fe203 content increases with increasing pressure, and to a lesser degree with temperature. Experimental results were applied to test existing ilmenite-based geothermometers. Both existing geothermometers progressively underestimated temperatures with an increasing ilmenite redox ratio. A least-squares fit to the experimental data, taking into account and correcting for the mole fraction of hematite in ilmenite, Xbem. determines Margules parameters defining the non-ideality terms and gives improved ilmenite-based geothermometers. Using the improved Margules parameters, a least-squares fit to the experimental data gives two ilmenite-based oxygen geobarometers: an improved olivineorthopyroxene- ilmenite geobarometer and a new ilmenite-rutile geobarometer. The barometers are applicable to a large spectrum of ilmenite compositions occurring in mantle rocks and gives reasonable results to temperatures as low as 900 °C. Mg-rich ilmenites, often with elevated Cr203 contents and generally with low Fe3+ contents are often found in association with diamonds and diamond-bearing rocks, or assemblages that are believed to have originated in mantle source regions in close proximity to the diamond stability field. The ilmenite-based geothermometers and oxygen geobarometers described above ignore the effect of minor cations present in natural ilmenites, and the validity of their use with such natural mineral assemblages is untested. In the third part of this study, chromium bearing Mg-ilmenites are synthesized in the presence of olivine-orthopyroxene±spinel±rutile assemblages between 1050 and 1200 OC and 17 to 35 kbar, under controlled oxygen fugacity, using oxygen buffering techniques described above. Oxygen buffering was achieved with the use of Fe-FeO (IW), WCW02- graphite (WCWO), Ni-NiO (NNO), and Fe203-Fe304 (HM) buffers under fluidexcess conditions. Within the P-T-X range of the experiments, ilmenite compositions showed increasing redox ratios and Cr203 contents as a linear function of oxygen fugacity. The Cf203 content of ilmenites at constant oxygen fugacities is strongly positively correlated with pressure. Experimental results were used to test existing geothermometers and oxygen geobarometers. A Monte Carlo fitting procedure to the experimental data, taking into account the Cr203 content of ilmenite, has modified previously determined Margules parameters expressing non-ideality in ilmenite solid solution and gives improved ilmenite-based geothermometers and oxygen geobarometers. Existing spinel-based geothermometers and geobarometers are also corrected using experimental data of high Cr# spinel compositions. The revised oxygen geobarometers are applied to natural ilmenite and spinel upper mantle-derived xenocryst assemblages from kimberlites and lamproites from various regions. A consistent picture emerges from the ilmenite and spinel data which suggests that the oxygen fugacity of the lithospheric upper mantle is broadly bounded by the IW and FMQ buffers. A further significant topic of this study relates to the incorporation of H+ into nominally anhydrous mineral phases as a coupled substitution with trivalent elements such as Cr3+ and Fe3+. The observations on iron and chromium substitution in rutile and their dependence on oxygen fugacity (for iron) led to the use of infra-red spectroscopy to confirm the presence of (OH) in both natural and synthetic rutiles. The presence of (OH) in rutile has previously been described. To investigate both the (CrOOH) and (FeOOH) solid solutions synthetic Ti02, Ti02-Cr203, Ti02-Fe203, Ti02-Fe304 and Ti02-FeO mixes were run in H20-saturated conditions to determine maximum saturation of the trivalent element in rutile. Attempts were also made to run Ti02-Cr203 mixes under H+absent conditions in nominally anhydrous conditions. In the Ti02-Cr203 experiments there was little difference in Cr203 substitution into rutile between H20-saturated and 'anhydrous' experiments, and in both cases there is a positive correlation with pressure and temperature. Infrared spectroscopy showed a strong (OH) absorption peak around 3200 cm-1 in rutiles synthesized under H20-saturated conditions, but no absorption peaks for the 'anhydrous' experimental rutiles were detected over the mid-IR region. The H20-saturated Ti02-CI203 experiments demonstrated that the incorporation of water in rutile is strongly controlled by pressure, temperature, and the redox state of the cations. At 1100 ·c, H20 contents of Cr-rich rutiles increased from 0.79 wt.% at 5 kbar to 1.08 wt.% at 20 kbar and similarly at 20 kbar, H20 contents increased from 1.08 wt.% at 1100 ·c to 1.65 wt.% at 1400 ·c. The incorporation of H20 into rutile is not favoured by the absence of trivalent cations. For example, rutiles synthesized in the presence of Fe203 at 20 kbars and 1100 ·c contained 1.18 wt.% H20, whereas rutiles synthesized at the same physical conditions in the presence of Fe304 contained 0.46 wt.% H20. Also, rutiles synthesized under hydrous conditions at increasing oxygen fugacities from Ti02-FeO compositions show increasing amounts of water and cation contents. Rutiles synthesized under hydrous conditions from Ti02-only compositions did not contain any water. Also, rutiles synthesized from the Ti02-FeO mix under varying oxygen fugacity conditions showed an increase in total iron content and water content with increasing f02. These studies reveal a role for rutile in assessing both activity of water and oxidation state. They also show that rutile, an accessory mineral in eclogite, MARID and some peridotite samples from the upper mantle, provides a possible mechanism for recycling significant water contents from shallow to deep mantle conditions