Theory & Simulation

We study viscous liquids, melts and a wide range of soft- and active-matter systems with computer simulation and non-equilibrium statistical physics. A major focus is to understand how dynamical processes on the microscopic scale influence macroscopic material properties, and how these dynamical processes are affected by strong external forces (such as gravity-driven flow) that cause the system to be far from thermal equilibrium and to show strongly nonlinear response.

The systems that we study are characterized by strong collective effects that cause time-delayed response. We use a theoretical framework that predicts the ensuing history-dependent material properties starting on the microscopic scale, and we combine this theoretical description with continuum-mechanics simulations to predict the material response across the scales up to the macroscopic level.

Active matter is a material class inspired by biological systems where the individual constituents of the material possess an internal driving mechanism that can convert energy into directed motion. These processes play an important role in many aspects of life, be it for wound healing in tissues or for the transport of cargo such as pharmaceutical drugs in living beings. We study collective effects among active particles, performing also experiments on microgravity platforms to access the regimes predicted by theory free from sedimentation or other unwanted gravity-induced effects. We also pursue physical theories for biological systems in the context of muscle dynamics, in close collaboration with the Institute of Aerospace Medicine of the DLR.

Molecular-dynamics simulations are used to study microscopic processes in model metallic melts, using effective interaction potentials that are derived using machine-learning approaches and calibrated against available experimental data. The simulation then allows to explain the trends that are observed in experiment, for example upon changing the composition of the molten alloy, and to rationalize these in terms of the relation between dynamics and liquid structure. The origin of the underlying microscopic processes, thermodynamic or kinetic, as well as entropic or energetic, can thus be addressed. Various empirical material laws, such as those connecting self- and interdiffusion, of diffusion and viscosity, can be tested to understand the region of respectively the limits of their validity. Within MD simulations, we also address the effect of flow on the solidification of metallic melts, revealing a delicate interplay between thermodynamic driving forces and kinetic processes in the flowing melt.

Theoretical descriptions are based on the mode-coupling theory of the glass transition (MCT). Combined with MD simulation, this theoretical modeling gives key insights into the generic aspects of liquid dynamics. MCT describes the dynamics of a liquid based on information about its static structure; we combine the theory with experimentally measured static-structure information to provide "first-principles" predictions of the mass-transport processes in realistic model systems for multi-component alloys. The theory is also used to describe the dynamics of soft-matter systems under flow, providing a microscopic description of the non-linear rheology of colloidal suspensions or of agitated granular matter. These systems are prone to strong nonlinear-response effects that cause qualitatively novel phenomena as compared to the near-equilibrium linear response.

In order to understand how these microscopic phenomena affect the macroscopic material properties, we combine MCT with meso- and macro-scale simulation techniques such as the finite-element method (FEM). In particular in glass-forming melts and suspensions showing pronounced visco-elastic response, memory effects become important. These give rise to material properties that depend on the processing history of the material, and that need to be captured by the material laws that enter the continuum-mechanics equations and that we derive from microscopic principles. A prominent example are frozen-in residual stresses in glassy solids that are caused by the flow prior to solidification and that dramatically change the toughness of the final material. To achieve such a multi-scale description of material properties starting from the microscopic equations of motion, is a major numerical challenge in computational fluid dynamics.

Current research topics include:

• Machine-learning of interaction potentials for metallic melts and complex fluids.

• Non-linear rheology of visco-elastic fluids, prediction of residual stresses in non-trivial flow geometries.

• From droplets to 3d-printing: visco-plastic flows in microgravity.