The sulfur content at sulfide saturation (SCSS) in silicate melts plays a pivotal role in governing the behavior of chalcophile elements in planetary magma oceans. Numerous high-pressure experiments have been conducted to determine SCSS, employing various regression methods to capture the thermodynamic characteristics of the system. However, existing empirical equations have shown limited predictive accuracy when applied to laboratory measurements. In this study, we have compiled and analyzed 542 experimental datasets encompassing diverse sulfide and silicate compositions under varying pressure-temperature (P-T) conditions (up to 24 GPa and 2673 K). Employing empirical equations, linear regression, Random Forest algorithms, and a novel hybrid approach combining empirical fits for P-T conditions with Random Forest modeling for compositions, we have developed multiple SCSS models. These models have been rigorously compared with laboratory measurements. Our findings reveal that the Random Forest and hybrid models exhibit exceptional predictive performance (R2 = 0.82–0.91, mean average error [MAE] < 746 ppmw S, residual mean standard error [RMSE] < 972 ppmw S) in comparison to previous empirical models (R2 = 0.28–0.69, MAE = 622–1,170 ppmw S, RMSE = 1,070–1,744 ppmw S). Linear regression falls in between the performance of classical and machine learning models. Furthermore, we have applied our hybrid model to predict SCSS during the solidification of magma oceans on Earth, Mars, and the Moon. A comparison of our model results with expected sulfur contents in residual magma oceans, calculated through mass balance, offers valuable insights. Our analysis confirms that sulfides precipitated during the early accretion phases of Earth and Mars, but not on the Moon. Subsequently, evolving compositions of magma oceans offset increasing sulfur concentrations, preventing sulfide precipitation during intermediate stages of crystallization. Late-stage sulfide precipitation, contributing significantly to the bulk-silicate sulfur abundances of Earth, Mars, and the Moon, occurred at shallow depths (120–220 km, 40–320 km, and <10 km, respectively) within their respective magma oceans. This study sheds light on predicting SCSS under a range of conditions, advancing our understanding of chalcophile element behavior in planetary magma oceans.

How to cite: ZhangZhou, J., Li, Y., Chowdhury, P., Sen, S., Ghosh, U., Xu, Z., Liu, J., Wang, Z., and Day, J.: Predicting Sulfide Precipitation in Magma Oceans on Earth, Mars, and the Moon Using Machine Learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2982, https://doi.org/10.5194/egusphere-egu24-2982, 2024.

The compositional diversity of the crust of Mercury revealed by NASA’s MESSENGER spacecraft is interpreted to result from partial melting of a heterogeneous sulfur-rich Mercurian mantle. Major magmatic activity and the building of its secondary volcanic crust are restricted to the first billion year of the planet evolution. In order to understand the causes for the production of diverse lavas and the early death of major volcanism, we have performed a suite of high-pressure and high-temperature partial melting experiments under reduced conditions at temperature and pressure conditions relevant the mantle (1450-1750°C; 5, 3.5, and 1.5 GPa) of potential primordial S-free and S-saturated mantles and obtained crystallization sequences, solidus and liquidus of the residual mantle of Mercury with the Mg/Si ratio of 1.02 (Mer8) and 1.35 (Mer15) contains under above conditions. Our experimental data reveal that the majority of chemical composition of the highest Mg/Si region (HMR) on the Mercury’s surface can result from ~25±15 wt.% melting of a deep primitive mantle. Additionally, the possibility that garnet was abundant in the deep mantle could explain that Mercury rocks but the High-Mg province overlap the terrestrial “Al-undepleted” array, consistent with low-pressure melting of a garnet-free mantle.

How to cite: Wu, P., Xu, Y., Lin, Y., Saracino, F., Namur, O., Cartier, C., and Charlier, B.: Melting the mantle of Mercury and the compositional diversity of the crust, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8112, https://doi.org/10.5194/egusphere-egu24-8112, 2024.

The Mars is thought to have been covered by a deep magma ocean after its formation. In the past, the solidification of this ocean was modeled by one-step experimental¹ and numerical models² assuming a Fe-rich (~18.7 wt% FeO) mantle composition³. Estimates of the mantle composition and depth of the Martian magma ocean were recently updated by using the latest cosmochemical model (~13.7 wt% FeO)⁴ and the newest seismic observations from the InSight mission (the depth of the mantle, ~1560 km)⁵. Here, we present an experimental crystallization study of a nominally dry experimental Martian magma ocean (MMO), simulating up to ~50 percent fractional crystallization of the updated MMO composition and refined core-mantle boundary condition. A ‘two-stage’ model of magma ocean solidification is assumed, which features early efficient crystal suspension up to 50% solidification in magma and corresponding equilibrium crystallization, followed by fractional crystallization of the later residual magma ocean. For the first stage experiments at pressures of 1.5, 5, 10, 13 and 16 GPa and a constant temperature were designed to represent MMO equilibrium crystallization. Results indicate formation of a cumulate pile of olivine, orthopyroxene, clinopyroxene, garnet, spinel, periclase and quartz. Our preliminary result significantly differ from the previous experimental and numerical studies^1,2, likely due to the updated mantle composition and interior structure of Mars. Further second-stage experiments will start at the averaged residual magma composition resulting from the first stage, and we will provide more detailed results at the conference.

References and Notes

1. Bertka, C. M. & Fei, Y. Mineralogy of the Martian interior up to core-mantle boundary pressures. J. Geophys. Res. Solid Earth 102, 5251–5264 (1997).

2. Elkins-Tanton, L. T., Zaranek, S. E., Parmentier, E. M. & Hess, P. C. Early magnetic field and magmatic activity on Mars from magma ocean cumulate overturn. Earth Planet. Sci. Lett. 236, 1–12 (2005).

3. Dreibus, G. & Wänke, H. Mars, a volatile-rich planet. Meteoritics 20, 367–381(1985).

4. Khan, A., Sossi, P. A., Liebske, C., Rivoldini, A. & Giardini, D. Geophysical and cosmochemical evidence for a volatile-rich Mars. Earth Planet. Sci. Lett. 578, 117330 (2022).

5.  Stähler, S. C. et al. Seismic detection of the martian core. Science 373, 443–448 (2021).

How to cite: Wang, X., Lin, Y., and van Westrenen, W.: Solidification evolution of a dry Martian magma ocean: Constraints from high pressure-temperature experimental petrology, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8318, https://doi.org/10.5194/egusphere-egu24-8318, 2024.

The aluminum (Al), lithium (Li) and phosphorus (P) in zircon, combining their partitioning coefficients between zircon and silicate melt, may show important insights into the evolution of Earth’s crust and habitability over geological timescales, but their partitioning coefficients have not been experimentally studied in detail. In this study, we conducted high-temperature experiments to constrain the partition coefficients of Al (D_Al), Li (D_Li) and P (D_P) between zircon and silicate melt. Our experiments firstly show the amounts of P has the identifiable effects on D_Al and D_Li, but not on itself. The positive and negative correlations of D_Al and D_Li as P content in zircon increasing enables us to constrain the aluminum saturation index (ASI) and Li content of the parental melt of zircon, respectively, by comparing their contents in zircon. The re-evaluated ASI values over time yields two significant increases at 3.6 Ga and 1.0 Ga, whereas, the calculated Li content of Earth’s crust non-linearly decrease with time. Meanwhile, the P content of early Earth’s crust is constrained to 1779 ppm before 3.0 Ga, similar to the present-day crust level of 2052 ppm. These indicate that the global scale tectonic events occurred on the Earth around 3.6 Ga and 1.0 Ga, and that the early Earth could have had a life habitability, and it did not get receded during the terrestrial crust evolution.

How to cite: Shang, S. and Lin, Y.: Re-experimentally constrained aluminum, lithium and phosphorus partitioning between zircon and melt: implication for early Earth’s crust evolution and possible habitability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7525, https://doi.org/10.5194/egusphere-egu24-7525, 2024.

A weakly temperature dependent isotope effect, i.e., the nuclear field shift, can produce isotope anomalies in the deep mantle due to redox condition changes. This effect can produce several to tens of ppm isotope anomalies or mass-independent fractionations at 2000 or even higher temperatures for many heavy metal isotope systems. This effect can be confirmed easily by comparing several isotope anomaly results of the same isotope system. For example, for W isotope anomalies, e.g., 180W and 183W, there are unique relationships between tthem if they are caused by the nuclear field shift effects. This work suggests a method to use heavy metal isotope systems as the tracers to detect the redox changes in the deep mantle. 

How to cite: Tang, M. and Liu, Y.: Redox-induced isotope anomalies in the deep Earth, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18384, https://doi.org/10.5194/egusphere-egu24-18384, 2024.

The redox state of the Earth’s interior (i.e., the oxygen fugacity, fo₂) is related to the Fe speciation (Fe²⁺, Fe³⁺) in mantle rock-forming minerals and controls the speciation of volatiles like carbon at depth. To date, the fo₂ of the lower mantle has been mostly constrained by HP-T experiments, due to the extreme rarity of natural samples represented by mineral inclusions in sub-lithospheric diamonds. Experimental evidence suggests that the lower mantle is reduced and saturated in Fe(-Ni) metal (about 1 wt%). However, coexisting minerals like ferropericlase and bridgmanite are predicted to contain 0.02 and 0.6 of Fe³⁺ /∑Fe, respectively. A slight increase of Fe³⁺ /∑Fe (less than 1 wt%) is expected in the case of fo₂ > iron-wüstite buffer. This would imply the complete oxidation of Fe(Ni) alloys promoted by reduction of carbonates (either fluids or melts). The finding of carbonates trapped in sub-lithospheric diamonds is natural evidence of the (local) oxidative redox state of the deep and inaccessible lower mantle and this is enhanced by the lack of metallic inclusions coexisting with Fe³⁺-poor ferropericlase in sublithospheric diamonds. Moreover, the variation of Fe³⁺ in bridgmanite appears, at least currently, to be better explained by its crystal chemistry while the effect of pressure and fo₂ remains unclear, mainly due to the lack of oxybarometers applicable to lower mantle assemblages.

In this study, we combined an experimental investigation of the Fe³⁺/∑Fe in ferropericlase and bridgmanite equilibrated at known high pressure, temperature and oxygen fugacity conditions with Fe³⁺ /∑Fe measurements conducted on bridgmanite(-like) and ferropericlase inclusions in sublithospheric diamonds from Rio Sorriso and São Luís (Brazil) and Kankan (Guinea). Some inclusions are composite for which the Fe³⁺/∑Fe was determined by in situ synchrotron Mössbauer source spectroscopy and the bulk Fe³⁺ /∑Fe determined.

Our preliminary results show a discrepancy between natural inclusions and experimental products in terms of i) modal abundance of ferropericlase and bridgmanite, likely related to their diverse role in diamond formation (redox) processes; ii) chemical compositions expected for both peridotitic and metabasaltic parageneses; and iii) Fe³⁺ /∑Fe content.

How to cite: Stagno, V., McCammon, C., Kaminsky, F., and Marras, G.: The redox state of the Earth’s lower mantle: combined evidence from natural mineral inclusions in superdeep diamonds and experimental predictions , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19097, https://doi.org/10.5194/egusphere-egu24-19097, 2024.

The redox state of the sub-arc mantle is critical in contraining the behavior of redox-sensitive elements, particularly volatile species, within the magmatic system. However, the long-term evolution of redox conditions in the sub-arc mantle and its impact on atmospheric oxygen levels remain unclear. In this study, we address two key challenges in understanding the sub-arc mantle redox throughout geological history: 1) the need for more effective methods to identify arc samples in different tectonic settings, and 2) the lack of constraints on the evolution of redox-sensitive trace elements from the mantle source. To overcome these challenges, we used an XGBoost machine learning model trained on multiple elements and elemental ratios (e.g., Nb, Ta, Ti, Pb, Th/Nb, Nb/La, and U/Nb) to accurately classify arc basalts and basalts (N ≈ 14,000) derived from other tectonic settings. Subsequently, we modeled the evolution of trace elements (V, Ti, and Sc) partitioning in the hydrous and dry depleted mantle beneath the arc using non-modal near-fractional melting. Finally, we applied the trained machine learning model and redox-sensitive elemental ratios to a refined global dataset of basalts (N ≈ 19,000) to calculate the redox evolution since 3.4 Ga. Our findings reveal that there have been minimal fluctuations in the melting T-P conditions since 3.4 Ga, indicating a consistent oxidized state in the sub-arc mantle. The average ΔFMQ of 0.96 ± 0.52 (1 SD) observed in our study is similar to that of the modern arc mantle since the early Archean. Interestingly, it is unlikely that the long-term oxidation of the sub-arc mantle has significantly impacted the progressive increase in atmospheric oxygen levels over time.

How to cite: Liu, C.-T., Ye, C.-Y., and Zhang, Z.: Sub-Arc Mantle Redox Evolution since the Archean, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3301, https://doi.org/10.5194/egusphere-egu24-3301, 2024.

The redox state of the convective asthenospheric mantle governs the speciation of volatile elements, such as carbon, therefore influences the depth at which (redox) melting can occur with implications for seismic signals. Geophysical observations suggest the presence of carbonatite melts at depth of 200–250 km, however, thermodynamic models indicate that the onset of (redox) melting would occur at 100–150 km for a mantle with 3–4 % of Fe³⁺/∑Fe. Here we present a new oxybarometer based on the V/Sc exchange coefficient between olivine and melt, which is insensitive to surficial alteration, volatile degassing, electron exchange reactions and fractional crystallization. By applying this method to primary mid-ocean ridge basalts (MORBs) from the Southwest Indian Ridge and East Pacific Rise, we demonstrate that the average oxygen fugacity (fo₂) of MORBs, corrected for the depth of formation, is 0.88±0.24 log units above the fayalite-magnetite-quartz (FMQ) buffer, which is slightly more oxidized than previously estimated (near FMQ buffer). Our findings indicate that the convective asthenospheric mantle exhibits a higher oxygen fugacity than continental lithospheric mantle. Along an adiabat, carbonatitic melts can form from a CO₂-bearing source at depth of 200–250 km explaining, therefore, the electrical conductivity and seismic velocity anomaly of the asthenospheric mantle.

How to cite: zhang, F., Lai, S., and Stagno, V.: The redox state of asthenospheric mantle and the onset of melting beneath mid-ocean ridges, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2592, https://doi.org/10.5194/egusphere-egu24-2592, 2024.

How to cite: Choudhary, S., Tao, R., Das, S., Sen, K., Pattnaik, J., Lu, Y., Kumar, S., and Viljoen, F.: Redox heterogeneity in lithospheric mantle evidenced by C-O-H-S fluid inclusions entrapped in Kerguelen mantle-plume associated pyroxenites, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8480, https://doi.org/10.5194/egusphere-egu24-8480, 2024.

The study of mineral inclusions trapped within lithospheric diamonds stands as one of the fundamental approaches to directly study the Earth’s interior, enhancing our comprehension of its chemistry and deep geological processes. The inclusions provide crucial insights into the pressure (P), temperature (T), and redox conditions (fO₂) occurring during the diamond nucleation and growth. Among the various mineral inclusions observed in lithospheric diamonds, those characterized by a peridotitic mineral assemblage (P-type), frequently exhibit the presence of Mg-chromite (Stachel et al., 2022) that suggests their involvement in redox-driven diamond formation from CO₂-bearing melts. Since the current understanding of the Earth’s interior redox state primarily relies on peridotite samples from the shallow upper mantle based on the Fe³⁺ content of spinels (Ballhaus et al., 1991), investigating the chemistry of spinel mineral inclusions in diamonds would extend the knowledge of the mantle redox state to greater depths.

In this study, we focused on a suite of nine diamonds extracted from the Udachnaya kimberlite pipes, located within the Siberian craton. These diamonds show multiple dark and transparent inclusions with sizes between 30 and 200 µm in diameter. The selected diamonds are polished to expose some of the inclusions. The chemical composition was determined by electron microprobe while textural features were observed by scanning electron microscopy. The Fe³⁺/ΣFe ratios of both encapsulated and polished inclusions were measured by in situ synchrotron Mössbauer spectroscopy at the ID18 beamline of the ESRF synchrotron (Grenoble, France), employing a 6 x 15 µm² focused beam.

The analyzed Mg-chromites exhibit FeO contents of 15-17 wt% and Cr# and Mg# of 0.85-0.92 and 0.54-0.61, respectively. The Mössbauer data collected on the Mg-chromites show a large variability with Fe³⁺/ΣFe ratios ranging from 0.07 to 0.28. Interestingly, variable chemical composition for the Mg-chromites among inclusions trapped in the same diamond along with fO₂ ranging within 2 log units below the fayalite-magnetite-quartz buffer suggest possible changes in the chemistry of the diamond growth medium providing, therefore, evidence of redox heterogeneities in the lithospheric mantle underneath the Siberian platform.

How to cite: angellotti, A., Marras, G., Logvinova, A., Mikhailenko, D., and Stagno, V.: Revealing Redox Variability in the lithospheric mantle: Insights from Mg-chromite inclusions in diamonds from the Udachnaya kimberlite pipes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-418, https://doi.org/10.5194/egusphere-egu24-418, 2024.

The potential of using natural Fe-Ti-rich heavy mineral sand, as an abundant, inexpensive, and ecologically acceptable raw material for the processing of pseudobrookite-based ceramics for thermoelectric applications is presented. The as-received raw powder was used in untreated form or after pre-oxidation at 600°C, 700°C, and 800°C for 1h in an air. The starting powders were consolidated into dense compacts via conventional solid-state sintering or spark plasma sintering. The phase composition, ferrous:ferric ratio and microstructural characteristics of the powders and sintered compacts were investigated by XRD, Mössbauer spectroscopy and SEM. The samples with a higher fraction of pseudobrookite exhibit relatively low thermal conductivity ranging between 2 and 3 W m^-1 K^-1, which is promising for practical thermoelectric applications. The Seebeck coefficient depends on the ferrous:ferric ratio and is rather low, leading to low power factor values for all the samples which will have to be improved for practical applications.

How to cite: Hanžel, D., Daneu, N., Radoševič, T., Bernik, S., Mazaj, M., Maček Kržmanc, M., Verhovšek, D., Kocjan, A., Vrabec, M., Spreitzer, M., and Guilmeau, E.: Pseudobrookite-based ceramics with low thermal conductivity prepared from Fe-Ti-rich heavy mineral sand concentrate, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19755, https://doi.org/10.5194/egusphere-egu24-19755, 2024.

In traditional high pressure-temperature assembly design, priority has been given to temperature insulation and retention at high pressures. This limits the efficiency of cooling samples at the end of experiments, negatively impacting many studies in the field of high-pressure earth and planetary science. Inefficient cooling of experiments containing molten phases at high temperature leads to the formation of quench textures, which makes it impossible to quantify key compositional parameters of the original molten phase, such as their volatile contents.

Here, we designed a novel, low-cost experimental assembly for rapid cooling in a six-anvil cubic press. This assembly not only retains high heating efficiency and thermal insulation, but also enables a very high cooling rate (~600 °C/s from 1900 °C to the glass transition temperature). Without using expensive materials or external modification of the press, the cooling rate in an assembly (~600 °C/s) with cube lengths of 38.5 mm is about ten times faster than the traditional assembly (~60 °C/s). Experiments have shown that the heterogeneous quenched textures produced with the traditional assembly, does not shown with the novel rapid cooling assembly design.

How to cite: Xu, Y., Wu, P., and Lin, Y.: A novel rapid cooling assembly design in a high-pressure Cubic Press apparatus, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8063, https://doi.org/10.5194/egusphere-egu24-8063, 2024.

We provide a numerical analysis aimed at quantifying uncertainty associated with methane (CH₄) production following geogenic hydrogen (H₂) and carbon dioxide (CO₂) generation. Our study stems from the observation that naturally generated H₂ can potentially be (i) reduced through, e.g., mineral-based (abiotic) geochemical processes and/or (ii) consumed through (biotic) methanogens. Both scenarios yield methane (CH₄) as a product. Some studies suggest relying on the H₂/CH₄ ratio as a straightforward indicator to assess the origin of methane in the subsurface. For example, Oze et al. (2012) rely on laboratory experiments of serpentinization associated with a given temperature/pressure condition and rock/fluid compositions and suggest that values of H₂/CH₄ larger than 40 are likely to indicate abiotic origin of CH₄. Otherwise, values H₂/CH₄ less than 40 suggest contribution of biotic activity to methane generation. Here, we consider the same types of (abiotic) geochemical reactions analyzed by Oze et al. (2012) and conceptualize the subsurface system as a natural chemical reactor within which a mixture of H₂ (generated from serpentinization) and CO₂ (generated from carbon-clay-reactions) yields a mixture of H₂, CO₂, and CH₄. Our analysis considers that complete mixing of the various chemical species is attained and that geochemical reactions can be evaluated under thermodynamic equilibrium conditions. We then perform a modeling study framed in a stochastic context and relying on a numerical Monte Carlo framework. We aim at quantifying the way uncertainties associated with hydrogen loss (as reflected through the H₂/CH₄ ratio) due to geochemical reactions at reservoir equilibrium condition can depend on corresponding uncertainties related to (i) composition of the fluids residing in the system, (ii) depth of a reservoir (i.e., as reflected through temperature/pressure conditions), and (iii) characterization of the thermodynamic equilibrium model. With reference to the latter point, uncertainties in terms of values of reaction equilibrium constants stem from the observation that temperature and pressure values associated with significant burial depths may fall outside ranges of validity of commonly employed thermodynamic databases and typically used geochemical software. Our stochastic simulation results suggest that (on average) almost 43% of native H₂ is consumed due to the geochemical reactions analyzed. This would correspond to an average value of H₂/CH₄ of about 13 (with first and third quantiles corresponding to 7 and 20, respectively). The ensuing sample probability density function of the H₂/CH₄ ratio displays a clear positive skewness. Our results may practically be used as a simple criterion to identify the probability associated with CH₄ production from geochemical processes involving natural H₂ under reservoir thermodynamic equilibrium conditions.

Keywords: serpentinization, methane, geochemical reactions, uncertainty quantification, H₂/CH₄ ratio.

References

Oze, C., Jones, L. C., Goldsmith, J. I., and Rosenbauer, R. J. (2012). Differentiating biotic from abiotic methane genesis in hydrothermally active planetary surfaces. Proceedings of the National Academy of Sciences, 109(25), 9750-9754. https://doi.org/10.1073/pnas.1205223109.

How to cite: Ranaee, E., Inzoli, F., Riva, M., and Guadagnini, A.: Quantification of uncertainty related to methane production associated with geogenic hydrogen and carbon dioxide, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7879, https://doi.org/10.5194/egusphere-egu24-7879, 2024.