MATERIALS SCIENCE AND CHEMISTRY

Molecular Modeling of Hydrothermal Fluids

Principal Investigator:
Prof. Dr. Sandro Jahn

Affiliation:
Institute of Geology and Mineralogy, University of Cologne, Germany

Local Project ID:
hydrothermal

HPC Platform used:
JUWELS CPU of JSC

Date published:

Introduction

Fluids are key agents in many geological processes of the Earth's crust and upper mantle. Despite their importance for geological and technological processes, their thermodynamic and physical properties are not well constrained at many of the relevant conditions, especially in the supercritical state. In this project, we collaborate with experimentalists and thermodynamicists to study properties of hydrothermal fluids in a wide range of densities and temperatures. The main goals of the simulations are the development of molecular structure models including electronic and vibrational properties and prediction of thermodynamic properties such as solute dissociation constants and partial molar volumes.

Results and Methods

Hydrothermal fluids are low viscosity mobile phases made of a volatile solvent such as H2O and CO2, and dissolved components. They are involved in chemical transport and mineral replacement reactions as well as in transport of heat and energy inside the Earth. They influence physical properties such as electrical conductivity, velocities of seismic waves and melting temperatures. In natural environments, hydrothermal aqueous fluids occur at temperatures of up to several hundred °C and densities ranging from less than 0.1 g/cm3 in parts of geothermal systems up to about 1.4 g/cm3 in subduction zones. Densities of carbonate- or silicate-rich fluids can be substantially higher. Physical and thermodynamic properties of hydrothermal fluids such as pressure-volume-temperature (P-V-T) equations of state, viscosity and other transport properties, and molecular speciation models are important input parameters for large-scale reservoir and reactive process models. Despite their importance, the thermodynamic and physical properties of hydrothermal fluids are not well constrained at many of the relevant conditions, especially in the high temperature range above the critical point of water (364 °C). The existing thermodynamic models are only applicable in a restricted range of fluid density, temperature and chemical compositions, and extrapolations bear a large uncertainty. Especially in the high temperature and low-pressure region the commonly used thermodynamic model for aqueous electrolytes at high P and T are not valid anymore. But many natural hydrothermal systems exist exactly in this range of conditions (300 to 700 °C, 0.1 to 100 MPa). These include magmatic-hydrothermal systems, mid-ocean ridge hydrothermal systems and deep roots of high enthalpy geothermal systems.

Concerted and collaborative efforts to address the problem of developing better models for hydrothermal fluids include measurements of physical and thermodynamic fluid properties at high P and T, the theoretical development of advanced thermodynamic models, and molecular-scale numerical simulations. This GCS project is concerned with the molecular modeling part that is aimed at developing molecular structure models for hydrothermal fluids, predicting thermodynamic and physical properties that are not available experimentally, and revealing the underlying mechanisms of structure-property relations.

To achieve these goals, established computational chemistry methods such as classical or ab initio molecular dynamics (MD) simulations are employed. Empirical force fields for the description of particle interactions are suitable for simple model systems, such as H2O-NaCl. They are used in classical MD simulations that are computationally efficient and therefore allow to simulate relatively large systems at many slightly different conditions to investigate the detailed functional form of property variations. On the contrary, ab initio MD simulations model the electronic structure of the molecular systems explicitly and are capable to describe chemically more complex and reactive systems. Their demand of computational resources is orders of magnitude larger than that of classical MD simulations, which limits accessible system sizes and simulation times. The supercomputing time allocated in this project allowed us to perform extensive simulations with dedicated MPI-OpenMP-parallelized codes (e.g. CP2K and LAMMPS)  on the JUWELS CPU cluster at Jülich Supercomputing Centre.

In one of the sub-projects, we studied the thermodynamic properties of aqueous NaCl solutions at T = 400 °C in a wide density range covering conditions from gas-like to liquid-like behavior of the supercritical fluid (see Fig. 1). The decreasing free volume in the fluid with increasing density leads to a cross-over from negative to positive partial molar volume of the solute ions. While at fluid densities above 0.3 g/cm3, the isothermal dissociation constant of NaCl is linearly dependent on density, a more complex behavior is observed in the simulations at lower density (Schulze et al., 2025).

Another case study was aimed at understanding the mechanisms of water dissolution in carbonate melts at conditions of the deep Earth (Schulze and Jahn, 2024). Ab initio molecular dynamics simulations of MgCO3 melt containing 10 wt% H2O revealed various molecular carbon species, of which CO32- and HCO3- were the most abundant (see Fig. 2). 

The formation of the latter requires the dissociation of H2O molecules and the formation of hydroxyl groups. Proton exchange reactions between the different species are observed frequently indicating high hydrogen mobility. 

Our molecular simulations have also been an indispensable complement to in-situ experimental studies of the vibrational (Schmidt and Jahn, 2024) and electronic properties (Sahle et al., 2022; Loges at al., 2024) of aqueous fluids at high pressure and temperature. Once reliable structural models are established, the thermodynamic properties derived from the simulations provide missing parameters for thermodynamic and reactive flow simulations in a wide range of geological settings.

We gratefully acknowledge project funding from DFG (JA1469/12-1, JA1469/13-1), EU project GEOPRO (ID 851816) and GCS (compute project hydrothermal on JUWELS supercomputer at Jülich Supercomputing Centre).

References

Loges, A., Manni, M., Louvel, M., Wilke, M., Jahn, S., Welter, E., Borchert, M., Qiao, S., Klemme, S., Keller, B.G., John, T., 2024. Complexation of Zr and Hf in fluoride-rich hydrothermal aqueous fluids and its significance for high field strength element fractionation, Geochim. Cosmochim. Acta 366, 167-181. https://doi.org/10.1016/j.gca.2023.12.013

Sahle, C.J., de Clermont Gallerande, E., Niskanen, J., Longo, A., Elbers, M., Schroer, M.A., Sternemann, C., Jahn, S., 2022. Hydration in aqueous NaCl, Phys. Chem. Chem. Phys. 24, 16075–16084. https://doi.org/10.1039/d2cp00162d

Schmidt, C., Jahn, S., 2024. Raman spectra of oxidized sulfur species in hydrothermal fluids, J. Volcanol. Geoth. Res. 454, 108146. https://doi.org/10.1016/j.jvolgeores.2024.108146

Schulze, M., Driesner, T., Jahn, S., 2025. Assessing the validity and limits of linear density models for predicting dissociation-association equilibria in supercritical water, Geochim. Cosmochim. Acta, in press. https://doi.org/10.1016/j.gca.2024.10.008

Schulze, M., Jahn, S., 2024. Speciation and dynamical properties of hydrous MgCO3 melt studied by ab-initio molecular dynamics, Geochim. Cosmochim. Acta 368, 147-155. https://doi.org/10.1016/j.gca.2023.12.003

Scientific Team

Prof. Dr. Sandro Jahn, Dr. Maximilian Schulze, Rajorshi Chattopadhyay

Scientific Contact

Prof. Dr. Sandro Jahn
Department of Earth and Environmental Sciences
Ludwig-Maximilians-Universität München
Theresienstraße 41, 80333 München (Germany)
e-mail: s.jahn[at]lmu.de

Webpage:  http://www.geomaterials-modeling.de