V42A-01 INVITED 10:20h
Mantle Oxidation State and Oxygen Fugacity: Constraints on Mantle Chemistry, Structure and Dynamics
Mantle oxidation state is reviewed based on Mössbauer measurements of mantle xenoliths and synthetic high-pressure phases, and suggests relatively low Fe$^{3+}$/$\Sigma$Fe in upper mantle and transition zone phases, but high Fe$^{3+}$/$\Sigma$Fe in lower mantle (Mg,Fe)(Si,Al)O$_{3}$ perovskite, even under reducing conditions, with significant implications for physical and chemical properties. Whole rock Fe$_{2}$O$_{3}$ concentrations for pyrolite mantle are calculated to be low in the upper mantle and transition zone (ca. 0.3 wt% Fe$_{2}$O$_{3}$), but high in the lower mantle (ca. 4 wt% Fe$_{2}$O$_{3}$). High Fe$_{2}$O$_{3}$ concentrations may be balanced according to the iron disproportionation reaction 3 Fe$^{2+}$ = Fe$^{0}$ + 2 Fe$^{3+}$. Oxygen fugacity is relatively high at the top of the upper mantle (near the fayalite-magnetite-quartz buffer) due to the concentration of Fe$^{3+}$ in modally minor phases, but probably low in the transition zone and lower mantle (close to metal equilibrium). Material in subducting slabs may be more oxidised, however, and the expected increase in Fe$_{2}$O$_{3}$ concentration due to an increase in majorite garnet proportion and the garnet-perovskite transformation may be partly balanced by reduction of oxidised species such as carbonate in the subducting slab. Such coupled redox reactions may contribute to the genesis of diamonds in the lower mantle. Disproportionation of iron in the lower mantle to produce ca. 1 wt% Fe-rich metal may have occurred in the early Earth, with important consequences for mantle geochemistry. If the metal phase does not segregate appreciably, material can move across the lower mantle boundary without a net enrichment or depletion in oxygen.
V42A-02 INVITED 10:35h
The Influence of Oxygen Environment on Kinetic Properties of Minerals
In this presentation, we review experimental determinations of the influence of oxygen environment on the physical properties of olivine. Kinetic properties of solids such as electrical conductivity and viscosity that are often functions of ionic diffusivity depend on the oxygen fugacity. Ionic diffusivity in solids can be expressed as the product of point defect concentration and point defect diffusivity. While thermogravimetry experiments reveal a pronounced dependence of point defect concentration on oxygen fugacity, electrical conductivity and thermogravimetry point defect relaxation experiments yield the point defect mobility. Results from these two types of experiments on olivine demonstrate that the power law dependence of ionic diffusivity on oxygen fugacity arises from the oxygen fugacity dependence of the point defect concentration alone. We also discuss experimental determination of the influence of oxygen environment on dislocation structures (i.e., deformation mechanisms) of olivine. Results from creep tests on single crystals of olivine indicate that for crystals compressed along a direction $45^{\mathrm{o}}$ to the [100] and [001] crystallographic directions, the deformation mechanism changed from glide or cross slip of screw dislocations to the motion of near-edge dislocations as the oxygen fugacity was increased from $10^{-11}$ to $10^{-3}$ atm.
V42A-03 INVITED 10:50h
Iron isotope fractionation in mantle minerals and the effects of partial melting and oxygen fugacity
The oxygen fugacity of the mantle exerts a fundamental influence on mantle melting, volatile speciation and the development of the Earth's atmosphere. As changes in mantle oxidation state are generally reflected in the ferric iron contents of mantle minerals, the iron isotope signatures of mantle rocks should provide information about spatial and secular changes in mantle oxidation state. However, the exact nature of the processes governing iron isotope fractionation in igneous rocks remains poorly understood, limiting the use of iron isotopes as a proxy. We have investigated the relationships between Fe isotope fractionation, oxygen fugacity, melting and metasomatism with a combined MC-ICPMS Fe isotope and M\"{o}ssbauer spectroscopy study of spinels and silicate minerals from mantle xenoliths and massif samples originating from different tectonic settings. There exist large variations in the iron isotope compositions of olivines, pyroxenes and spinels. Clear correlations exist between the $\delta$$^{57/54}$Fe values of coexisting minerals, implying equilibrium isotope fractionation. Spinel $\delta$$^{57/54}$Fe values correlate negatively with relative oxygen fugacity, spinel Fe$^{3+}$/$\Sigma$Fe and Cr number; clinopyroxene $\delta$$^{57/54}$Fe values correlate negatively with clinopyroxene MgO content and Cr number. There do not appear to be strong relationships between the $\delta$$^{57/54}$Fe values of the minerals and chemical or isotopic indices of metasomatism. Taken together these observations imply that variations in the iron isotopic compositions of mantle rocks and minerals are largely a function of melting coupled with changes in mantle oxidation state.
V42A-04 11:05h
Relationship between Diamond Dissolution Features, Oxygen Fugacity and Temperatures of Lac de Gras Kimberlite Magmas
Natural diamonds recovered from kimberlites usually show a variety of morphological forms and surface features, reflecting the complex history of diamond growth, dissolution and transformation during its residence in the mantle and ascent in kimberlite melts. High-temperature kimberlite magma is a reactive media for diamonds. The reaction with kimberlite is an important mechanism of diamond destruction that may greatly influence the grade and value of diamonds in a kimberlite pipe. Crystallization temperatures (T) and oxygen fugacity (fO2) of kimberlite melts may notably influence the kinetics of diamond dissolution process and the character of the resorption features developed on the diamond surface. In order to see how notable is the impact of the parameters of the host kimberlite melt on the characteristics of the diamond population we compare the T and fO2 of kimberlite melt crystallization with the diamond contents and the characters of resorption in diamonds from several kimberlite pipes in the Lac de Gras region (NWT, Canada). Exceptionally fresh kimberlites from the Lac de Gras region allowed us to apply olivine-spinel (Ol-Sp) thermometry and oxygen barometry to chromite inclusions in olivine phenocrysts in these kimberlites. We extended our previous estimates to a larger number of kimberlite pipes from different clusters within the Lac de Gras area. A detailed description of the morphology and surface features of individual stones was compiled for five kimberlite pipes from the both clusters. The T and fO2 values calculated at an assumed pressure of 1 GPa are in the range 970 - 1140 \pm 50 \deg C and 2.8 to 4.4 log fO2 units below the nickel-nickel oxide (NNO) buffer. Kimberlites from one cluster show very similar T and fO2 values, whereas variations between the clusters are within 150 \deg C and 1 log unit respectively. These differences correlate well with variations in the characteristics of the diamonds. An increase in diamond resorption in the kimberlites corresponds to increase in T. The development of various surface dissolution pits and structures is more extensive in the more oxidized kimberlites. The two processes of diamond dissolution - volume resorption and surface etching, do not show a strong correlation with each other. The diamond grade correlates with decrease in kimberlite fO2, while the quality of diamonds decreases with T. The established correlations between T- fO2 values of kimberlites and diamond resorption features suggest that conditions of the host kimberlite may notably influence the entrained diamonds. Our data shows that the development of the surface etching mainly occurs in kimberlite melt and therefore is controlled by the conditions of the melt, whereas the volume resorption depends on the conditions in both the kimberlite and in the mantle source and its relationship with the kimberlite conditions are obscure. The small but distinct variation in the fO2 between the kimberlites from the different clusters may have notably influenced their diamond populations.
V42A-05 11:20h
Evidence for methane-derived diamond from a single lherzolite xenolith (Premier, South Africa)
We report the first concentrations and isotopic compositions of carbon and nitrogen of 24 diamonds released from a single lherzolite xenolith from the Premier kimberlite in South Africa. This nodule of about 40 cm$^{3}$ contained 59 colourless sharp-edged octahedral diamonds, ranging from 0.05 to 0.17 carats. Their nitrogen concentration range from 47 to 1221 ppm (n = 59) and are clearly correlated with nitrogen speciation (10 to 85 % of IaB defects, the most stable state for nitrogen in a diamond cristalographic network). Nitrogen aggregation is a diffusive process depending on residence time, and temperature in the mantle, together with total nitrogen concentration. For Premier xenolith diamonds, the correlation between N-content and N-aggregation state confirms the synchronised formation at T = 1180 $\pm15\deg$C for a mantle residence time of 750 Ma (age of 1.2 Ga). $\delta^{13}$C values vary in a small range (from -4.26 to -0.84 $\permil$) and are positively correlated to both $\delta^{15}$N (from -1.12 to +7.18 $\permil$), and nitrogen contents (between 47 and 1221 ppm). $\delta^{13}$C and N-contents reported in the present work are within the ranges previously measured in diamonds from Premier (e.g. Deines, 1984). Positive correlations among the parameters cannot result from any mixing between two realistic end-members (e.g. mantle-derived and metasedimentary carbonates). In contrast, they could be explained by an isotope fractionation related to diamonds crystallisation. Most authors suggest that diamonds formed from oxidized carbon. However, the present carbon isotopic compositions are incompatible with a diamond crystallisation from mantle-derived carbonate-bearing fluids/melt. Parameters required by such a model are indeed not supported by previously determined isotope fractionation factors. Alternatively, we show that diamonds $\delta^{13}$C-$\delta^{15}$N-N-content covariations are compatible with an open-system crystallisation from reduced carbon bearing fluids. We believe to have therefore a first strong evidence for methane-related diamond formation.
V42A-06 11:35h
Similar oxygen fugacities in arc and MORB mantle source regions: evidence from V/Sc systematics
V/Sc systematics in peridotites and lavas are investigated to constrain the fO2 of the asthenospheric mantle. V/Sc ratios may see through certain processes that can modify barometric fO2 determinations in mantle rocks and/or magmas: early fractional crystallization, degassing, crustal assimilation, and mantle metasomatism. Melting models are combined here with a literature database on peridotites, arc lavas, and mid-ocean ridge basalts along with new, more precise data on peridotites. V/Sc ratios in primitive arc lavas from the Cascades magmatic arc are subtly correlated with fluid mobile elements (e.g., Ba and K), indicating that fluids may influence fO2 during melting. Despite evidence for fluid control in arcs, the average V/Sc-inferred fO2s of arc basalts, MORBs, and peridotites are remarkably similar (-1.25 to +0.5 log units from the FMQ buffer). This suggests that the upper part of the Earth's mantle may be strongly buffered in terms of fO2. Higher barometric fO2s of arc lavas and arc-related peridotite xenoliths may be respectively due to magmatic differentiation processes and to exposure to large time-integrated fluid fluxes incurred during the long term stability of lithospheric mantle.
V42A-07 INVITED 11:50h
Electron Microprobe Analysis of Fe$^{2+}$/Fe$^{3+}$ in Minerals With low Total Iron Concentrations
The development of the `flank method' by H\"{o}fer et al. (1994) has made it possible to quantify ferrous and ferric iron concentrations in minerals using the electron microprobe. The flank method makes use of the changes in both the wavelength and intensity of soft Fe L$\alpha$ and Fe L$\beta$ X-ray emission lines of minerals containing Fe$^{3+}$ and Fe$^{2+}$.By measuring at energies off the peak maxima (on the peaks flanks) the differences due to variable ferric iron ratios are maximized, thus making the flank method much more sensitive than methods relying on either peak shifts or peak area ratios. Using a correction for self-absorption, the Fe$^{3+}$/Fe$^{2+}$ ratio of minerals may be accurately and precisely determined. The original flank method was developed for minerals with high total iron concentration e.g. Fe-rich garnet end-members (almandine, andradite, and skiagite), and Fe-oxides (w\"{u}stite). To make it applicable to minerals with total iron concentrations of less than 10 wt.% as is common in mantle-derived minerals, we have modified, in three ways, the flank method to significantly improve the precision. Firstly, we have increased the number of analyses per mineral grain to 400 thereby providing a far more representative mean. Secondly, because it is necessary to accurately reproduce the exact position of the spectrometer crystal for each flank measurement, we have eliminated the need to reposition the spectrometer by serially measuring each flank position. Thirdly, we compensate for instrumental drift by measuring two 10 $\mu$m by 10 $\mu$m grids (1 $\mu$m spot size and spacing) for both flank positions. Each mineral grain is measured sequentially, collecting 200 analyses on the first flank position. The procedure is identically repeated for measurements on the second flank position. The elapsed time between each grid measurement is the same for all samples and drift for either flank position is eliminated by averaging. Using these modifications to the flank method, we have been able to precisely measure ferric iron concentrations of less than 1 wt.% in mantle-derived pyrope garnets with true microprobe spatial resolution. The accuracy of the flank method is currently being verified with spectrophotometric wet-chemical analysis of ferric iron in pyrope garnet standards. References H\"{o}fer, H.E., Brey, G.P., Schulz-Dobrick, B., and Oberh\"{a}nsli, R., 1994. Eur. J. Min., 6: 407-418
V42A-08 12:05h
Valence state of iron in mantle phases at high pressures and high temperatures
Starting from orthopyroxene with composition Mg$_{0.90}$Fe$_{0.10}$SiO$_3$, the valence states of iron in its high pressure polymorphs (clinopyroxene, ringwoodite, ilmenite, and perovskite) have been studied using synchrotron Mössbauer spectroscopy (SMS) at pressures to 31 GPa and temperatures to 2000 K. The orthopyroxene sample was loaded in a diamond anvil cell. The high-pressure phases were synthesized with laser heating at 1500-2000 K and at pressures of 12 GPa (clino-pyroxene), 19 GPa (ringwoodite), 22 GPa (ilmenite), and 31 GPa (perovskite). Mössbauer spectra were recorded in situ before, during, and after laser heating at each pressure with the newly developed SMS techniques. All phases were identified by x-ray diffraction at room temperature under high pressures. Data for the orthopyroxene starting material show only Fe$^{2+}$ in two octahedral sites, in agreement with literature data. After compressing the orthopyroxene to 12 GPa, SMS data can be interpreted by Fe$^{2+}$ in a single octahedral site. During heating at 1400 K, Mössbauer spectra show stark difference from that at room temperature. One possible explanation for the high temperature data is that an intermediate spin state for Fe$^{2+}$ could be formed because the high-spin and low-spin configurations could be in the thermally accessible range. After temperature quench at 12.5 GPa, the transformed clinopyroxene shows Fe$^{2+}$ in a single octahedral site. Data for ringwoodite at 19 GPa after heating show small amount, yet detectable, Fe$^{3+}$. Significant amount of Fe$^{3+}$ is observed at higher pressures for ilmenite and perovskite in which Si is six-coordinated. Implications for the oxidation state of the upper mantle, the transition zone and the lower mantle will be discussed.