Organizers: Marita Thomas, Timm John

Abstract Part A: 

Processes at fluid–solid interfaces fundamentally control mass transfer, reaction kinetics, and permeability evolution in geological materials.

This session focuses on the modeling and mathematical methods for pore-scale mechanisms including fluid-solid element exchange, moving boundary problems during dissolution and precipitation, and fluid liberation from hydrous minerals during metamorphic reactions. Particular attention is given to fluid flow in low-porosity systems and transport in confined geometries, where capillarity, surface energy, and reactive interface dynamics dominate.

Speakers:

11:30h: Xin Zhong (FU Berlin)

Modelling metamorphic and magmatic mineral evolution with machine-learning interatomic potentials and phase-field methods

Metamorphic and magmatic rocks undergo complex reactions and element transport at high pressures and temperatures. Conventional phase-equilibrium modelling predicts stable mineral assemblages by minimizing Gibbs free energy but relies on thermodynamic data that remain poorly constrained for some minerals, particularly under extreme conditions. We first developed a machine-learning interatomic potential for the Mg-Al-Si-O system, which represents several common mantle minerals (Zhong et al., 2026). The potential was trained on density-functional-theory energies and forces for more than 20 minerals and melts using an iterative active-learning workflow. Molecular-dynamics simulations with the resulting potential reproduce experimentally constrained thermodynamic properties and phase relations over wide pressure-temperature ranges. Enhanced-sampling simulations further determine the orientation-dependent solid-melt interfacial free energies of periclase and forsterite and quantify their anisotropy, which is difficult to constrain experimentally.

The usage of the interatomic potential is further upscaled to grain scale in metamorphic and magmatic petrology. Currently, a phase field model is being developed, and preliminary results show that it is possible to simulate the texture evolution of complex chemical systems. The model couples Cahn-Hilliard and Allen-Cahn equations within an extended Kim-Kim-Suzuki framework (Kim et al. 1999), incorporating anisotropic interfacial energies, element diffusion and interface kinetics. Interface propagation is benchmarked against an analytical diffusion-controlled Stefan problem, while predicted equilibrium assemblages are compared with Gibbs-free-energy minimization. The model is then applied to forsterite crystallization from melts using pre-determined interfacial energies, demonstrating that crystal morphology is strongly influenced by undercooling and interfacial anisotropy. A more complex application to garnet growth in rock explores how pressure-temperature evolution and element diffusion control the formation and preservation of garnet compositional zoning.

Kim, S. G., Kim, W. T., & Suzuki, T. (1999). Phase-field model for binary alloys. Physical Review E - Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics, 60(6), 7186–7197. https://doi.org/10.1103/PhysRevE.60.7186

Zhong, X., Li, Y., & John, T. (2026). A general purposed machine learning interatomic potential for Mg-Al-Si-O system suitable for Earth materials at high pressure and temperature conditions. Npj Computational Materials, 1–13. https://doi.org/10.1038/s41524-026-02056-3

12:00h: Liudmila Khakimova (University of Lausanne)

Chemically- and pressure-driven metamorphic (de)volatilization processes: front-tracking analytics

Metamorphic (de)volatilization reactions, such as hydration and dehydration, play a fundamental role in the evolution of rocks. Coupled with fluid transport, these reactions may localize into narrow reaction zones that propagate as sharp fronts, separating reacted and unreacted assemblages. Predicting the propagation of such fronts remains challenging because their characteristic time scales are not necessarily governed by hydraulic or chemical diffusion alone, but also by mass balance across the moving reaction boundary.

This contribution presents a front-tracking analytical framework for both pressure-driven and chemically driven metamorphic (de)volatilization processes in reactive fluid-rock systems. The reaction zone is treated as a sharp moving boundary, enabling Stefan-like analytical solutions for front propagation. The formulation covers both single-front systems and systems with multiple coupled fronts, such as dehydration followed by vaporization of the liberated pore water or cascades of metamorphic reactions propagating through distinct mineral assemblages. The resulting solutions predict an effective diffusivities that control front positions and the characteristic time scale of fluid-rock transformations.

The derived analytical solutions are applied to several geodynamically relevant hydration and dehydration systems. To demonstrate its generality, the fundamentals of pressure-driven reaction-front propagation are first illustrated using a simple brucite-periclase system, in which reaction progress and the associated density change are controlled by variations in fluid pressure. The framework is then extended to chemically driven reactions, in which mineral transformation is controlled by gradients in the concentration of a mobile fluid component. A chemically driven hydration example considers an initially porous forsterite-brucite assemblage exposed to a fluid composition that stabilizes antigorite and talc. Diffusive transport of dissolved components drives mineral replacement and generates a sharp propagating hydration front separating the initial and altered assemblages. This example demonstrates that chemically induced metamorphic reactions can be described using the same front-tracking analytical approach as pressure-driven (de)volatilization fronts.

The framework is also applied to published gypsum dehydration experiments by Fusseis et al. (2012) and Leclère et al. (2018). The analytical solutions provide an alternative interpretation of the observed transformation patterns as a system of coupled sharp fronts, in which dehydration-driven fluid release is followed by vaporization of the liberated pore water. Overall, the presented front-tracking framework enables quantitative prediction of front velocities, effective diffusivities, and transformation time scales in metamorphic fluid-rock systems driven by pressure perturbations, compositional perturbations, or their combination.

Fusseis, F., Schrank, C., Liu, J., Karrech, A., Llana-Funez, S., Xiao, X. and Regenauer-Lieb, K., 2012. Pore formation during dehydration of a polycrystalline gypsum sample observed and quantified in a time-series synchrotron X-ray micro-tomography experiment. Solid Earth, 3(1), pp.71-86.

Leclère, H., Faulkner, D., Llana-Fúnez, S., Bedford, J. and Wheeler, J., 2018. Reaction fronts, permeability and fluid pressure development during dehydration reactions. Earth and Planetary Science Letters, 496, pp.227-237

12:30h: Alexander Koelzer (Utrecht University)

From Millions of Years to Milliseconds – Earthquake sequence simulations in complex tectonic settings

Earthquakes occur on a timescale of seconds and are experienced by humans as sudden, devastating disasters. However, the tectonic conditions that determine where they occur are shaped over time scales from millions of years to years. Deformation in realistic tectonic settings is characterized by a host of viscous, elastic, and brittle processes and the interactions between them. To understand the present-day seismicity and earthquake cycle, these deformation processes need to be considered across all time scales. However, numerical models have not been able to resolve the dynamics of both long-term tectonics and short-term earthquakes.

We present a novel numerical modeling technique that simulates fully dynamic earthquake sequences and slow slip events in complex tectonic settings described by a visco-elast-plastic rheology. We employ an invariant formulation of rate- and state-dependent friction and adaptive time stepping to fully resolve all phases of the seismic cycle. Faults form and evolve spontaneously in the continuum according to heterogeneous, temperature-dependent material parameters and the local stress field during both long-term deformation and the seismic cycle. The total strain budget is dynamically distributed across the faults according to the rheology, structural features, and driving forces. We apply this modeling approach to a subduction zone and the central Apennines.

We generate events covering the slip spectrum from aseismic creep to earthquakes with slip rates in the order of m/s. In the subduction zone setting, we find earthquakes to be largely characteristic despite the complexity and deviating rupture paths in the subduction channel. Sufficiently weak splay faults can be dynamically triggered during a megathrust earthquake from locally amplified trapped seismic waves. This shifts coseismic slip from the megathrust onto the splay fault, resulting in a large surface rupture. Such short-term effects alter the long-term deformation compared to a purely geodynamic model. In the continental setting, results show complex rupture patterns and fault interactions. The interplay between interseismic loading, stress transfer from seismic and aseismic slip, and viscoelastic relaxation governs where subsequent earthquakes occur.

We conclude that our model successfully combines aspects of established geodynamic models and dynamic rupture models, providing a missing link between the long-term and the short-term. It furthermore provides a way of simulating realistic tectonic loading in a geodynamic framework validated against GNSS observations. This opens up new paths of linking observations to seismic hazard analysis via physics-based models.