-
- MR13A-04: Thermoelasticity of Geologically Relevant Silica Phases (SiO2) Using Machine Learning Potentials
-
NOLA CC
Author(s):Generic 'disconnected' Message
Jessica Santos Rego, Columbia University of New York (First Author, Presenting Author)
Chenxing Luo, Princeton University
Caetano Rodrigues Miranda, University of São Paulo
Renata Wentzcovitch, Columbia University in the City of New York
Silica (SiO2) is a major component of the oceanic crust that sinks deep into the Earth through subduction. Under the extreme pressures and temperatures of the lower mantle, silica can transform into different crystal structures. These phase transitions affect the material’s density and elasticity, influencing how seismic waves travel through the Earth’s interior. Two key transitions, from stishovite to a CaCl2-type structure, and from CaCl2-type to seifertite, are expected to occur in subducted mid-ocean ridge basalt (MORB). These transitions are thought to contribute to seismic anomalies detected near the core-mantle boundary (CMB), particularly in the D' region. However, most previous studies used simplified models that ignore anharmonic effects, which become important at high temperatures and can shift phase boundaries. To address this, we trained a machine learning model using high-accuracy quantum mechanical data and simulated how silica behaves under these extreme conditions. This approach enabled us to calculate temperature-dependent elastic properties and free energies, incorporating anharmonic effects. The resulting phase boundaries are more accurate and relevant to deep mantle conditions. Our findings improve the understanding of silica’s behavior in subducted MORB and help refine models of the deep Earth and the seismic features observed near the CMB.
Scientific DisciplineSuggested ItinerariesNeighborhoodType
Enter Note
Go to previous page in this tab
Session
