This project is aimed at the investigation and quantification of rare earth element speciation in hydrothermal fluids using molecular modeling approaches. From the simulations species stability constants are derived. They are important parameters for thermodynamic modeling of hydrothermal ore-forming systems. So far, these parameters are largely unconstraint or they have been extrapolated from low temperature experiments with large uncertainty. The speciation models are used to predict under which conditions rare earth elements are mobilized (and concentrated) in hydrothermal systems. Within the project we will develop new interaction potentials based machine-learning algorithms. Due to their high efficiency and accuracy the new potentials will be used increasingly in molecular dynamics simulations and at least partially replace ab initio methods used to far. The focus of the investigation will be on the elements yttrium and lanthanum as representatives of the heavy and the light rare earth elements. For the complex formation we will concentrate on mixed chloride/fluoride-hydroxide ligands as well as on sulfate and carbonate ligands.