Evidence for a liquid silicate layer atop the Martian core

Understanding the internal structure and composition of Mars is crucial for deciphering its formation and evolution. Previous seismic studies using InSight data suggested a large, low-density core (6.0–6.3 g cm⁻³), implying significant light element enrichment (S, C, O, H). This low density poses a challenge, as the abundance of these volatile elements in Mars's likely building blocks (chondritic meteorites) is insufficient to explain such a large density deficit. The main research question is whether the initially proposed model of a large, low-density core is consistent with all available data (seismic, geophysical, and cosmochemical). The purpose of this study is to reconcile the conflicting data by proposing a new model of Mars's interior. The importance lies in providing a more accurate representation of Mars's internal structure that can be used to test and refine theories of planetary formation and evolution. A more accurate model allows for a better understanding of Mars's geological history, including processes like core formation, mantle differentiation, and volcanic activity.

Previous estimates of Martian core composition have relied primarily on geochemical arguments or on seismic and geophysical observations focused on the mean core density and tidal response. However, these methods have limitations. Geochemical models struggle with variability and poor constraints on volatile element abundances in Martian precursor materials. Seismic studies using core-transiting phases provided estimates of core size and density, but could not directly constrain composition. The abundance and identity of light elements within Mars's core remained uncertain due to the limitations of both geochemical and earlier seismic interpretations. The availability of more comprehensive seismic data, and advancements in ab initio modeling of material properties under high pressure and temperature, provided a new opportunity to refine Mars's core model.

The authors combined InSight seismic data with ab initio molecular dynamics (AIMD) simulations. First, they compared AIMD-predicted P-wave velocity and density profiles for pure liquid iron with InSight data. This revealed a significant density deficit at the core-mantle boundary (CMB), which cannot be fully explained by light element enrichment. To address the discrepancy, they performed inversions based on differential body wave travel time data, aiming to resolve the CMB region and the core's radial structure. The inversion involved the use of a newly updated differential travel time dataset and stacked P-to-S waveform data, incorporating constraints from the planet's mean mass and moment of inertia. The AIMD simulations calculated the thermoelastic properties of Fe-Ni-X alloys (X = S, C, O, H) at Martian core conditions. These simulations are essential for converting seismic observations (velocity and density) into compositional constraints. By creating senary Fe-Ni-S-C-O-H mixtures, the authors generated a broad range of possible compositions to test against the seismic data. A key aspect was the identification and interpretation of seismic phases that interact with the proposed molten silicate layer. This included identifying P and S wave reflections and diffractions from the layer boundaries and the core, to test the validity of the revised model. Cosmochemical constraints on the abundances of light elements were used to refine the possible compositions, eliminating those considered geochemically implausible.

The core density and seismic velocities at the CMB could not be matched by any single Fe-Ni-S-C-O-H mixture, implying that the previously defined CMB is instead the lower boundary of a molten silicate layer. This layer significantly modifies the estimated core radius and density. Inversion of seismic data reveals a smaller, denser core (radius 1,675 ± 30 km, density 6.65 ± 0.1 g cm⁻³) overlain by a 150 ± 15 km thick molten silicate layer (density 4.05 ± 0.05 g cm⁻³). This model reconciles previously conflicting geophysical and cosmochemical data. The presence of the molten silicate layer is supported by the identification of unique seismic phases (SdS, PLS, PCMB, Pa^LSL Pair, SDKDS) in the InSight data that are consistent with the model's predictions. The best-fitting core composition includes 85-91 wt% Fe-Ni and 9-15 wt% of light elements, chiefly sulfur, carbon, oxygen, and hydrogen. The hydrogen abundance is higher than expected from cosmochemical constraints, suggesting either a higher-than-expected partition coefficient during core formation or the presence of other light elements (N, P) not considered in the model.

The findings address the long-standing discrepancy between seismic estimates of core density and cosmochemical constraints on light element abundances in Mars. The new model, with a smaller, denser core and overlying molten silicate layer, resolves this discrepancy by reducing the required amount of light elements within the core itself. The significantly increased density of the core within this new model remains consistent with seismic data. The presence of the molten silicate layer suggests a complex dynamic interaction between the core and the mantle, potentially influencing thermal and chemical evolution. The identification of previously unaccounted for seismic phases adds robust support to the proposed model, highlighting the value of detailed waveform analysis in interpreting seismic data. The inferred core composition is consistent with models that incorporate cosmochemical constraints, but with the notable exception of higher-than-expected hydrogen abundances.

This study presents compelling evidence for a molten silicate layer atop the Martian core. This revised model resolves the long-standing conflict between seismic and cosmochemical data regarding the core's composition. The findings highlight the importance of considering complex interactions between the core and mantle in understanding planetary evolution. Future work should focus on refining the model, particularly through improved experimental measurements of the physical properties of liquid Fe-Ni-X alloys and further analysis of InSight data to confirm the properties of the molten silicate layer. Analysis of Martian volcanic rocks may offer insights into the chemical characteristics of the molten silicate layer.

The model assumes a global, interconnected molten silicate layer. The possibility of local variations in layer thickness cannot be entirely excluded, which could affect the interpretation of some seismic phases. The AIMD simulations, while valuable, require further experimental verification. The model's reliance on the accuracy of seismic data and interpretations also introduces some degree of uncertainty. Further research, particularly experimental validation of the AIMD simulations, and acquisition of more extensive seismic datasets, would enhance the confidence in the interpretations and refine the proposed model.