Projects
Kinetic Theory of Decentralized Learning
Over the past two decades, active matter has emerged as a new central field within soft matter and statistical physics. A defining characteristic of active matter is its capacity to harness energy from its environment to facilitate self-propulsion. Despite these non-equilibrium forces, usual active matter remains inherently “dead”, lacking the capability to learn and process information and adapt to its environment beyond predetermined policies. In this project, based on kinetic theory, we introduce a novel statistical physics perspective on smart matter, which possesses the ability to dynamically adjust its policy.
Collaborators
Glass Transition
The transition from a liquid to a glass is characterized by a drastic slow-down of structural relaxation and transport, while the structure remains amorphous and seemingly unchanged. In this project we investigate connections between structural properties of supercooled liquids and their emerging dynamics using computer simulations, machine learning and mode-coupling theory.
Specific topics and collaborators
- Machine-Learning Glasses: Ludovic Berthier, Giulio Biroli
- Mode-Coupling Theory for Confined Glasses: Thomas Franosch
Dynamic Coarse-Graining and Non-Markovian Dynamics
Coarse-graining describes the process in which a high-dimensional microscopic system is reduced to a model with much fewer degrees of freedom. One important property of the equations of motion of this coarse-grained model is their non-Markovianity: The missing knowledge on the motion of the neglected degrees leads to the emergence of memory in the dynamics of the coarse-grained variables. We study how these memory effects can be quantified in both equilibrium and non-equilibrium systems. Furthermore we develop tools to simulate particles with non-Markovian dynamics. (See Introducing memory in coarse-grained molecular simulations for a recent review.)
Specific topics and collaborators
- Reconstructing Memory Kernel: Friederike Schmid, Martin Hanke-Bourgeois
- Simulating Non-Markovian Coarse-Grained Models: Friederike Schmid
Non-Equilibrium Dynamics
Non-equilibrium systems, such as active suspensions or externally-driven particles, show phenomenology which goes far beyond equilibrium physic. Importantly, dynamical properties such as friction and memory do not only affect liquid transport in non-equilibrium systems but can also influence the emerging structural properties of liquids. We are interested, in how complex dynamical properties, such as viscoelasticity, affects collective motion of active particles, but also more abstract concepts known from equilibrium physics, including fluctuation-dissipation theorems.
Specific topics and collaborators
- Fluctuation-Dissipation Theorems in Non-Equilibrium: Friederike Schmid
- Reconstructing Memory Kernels in Active Systems: Jeanine Shea and Friederike Schmid
- Active Particles in Viscoelastic Media: Ryoichi Yamamoto
Confinement
Liquids in confinement behave very different from their bulk counterparts. A well known phenomenon is for example the wetting at the boundaries, which leads to non-homogeneous densities and layering, but also dynamical properties are strongly affected by confinement. We study structural properties of liquids in extreme confinement by fundamental-measure theory and simulations. We also investigate confined glasses and confinement-induced crystallization by theory, simulations and experiments.
Specific topics and collaborators
- Mode-Coupling Theory for Confined Glasses and Liquids in Extreme Confinement: Thomas Franosch
- Confinement-Induced Crystallization: Charlotte Petersen
- Colloids in Confinement (Confocal Microscopy): Stefan Egelhaaf and Alejandro Villada-Balbuena
- Slip and Yielding in Confined Soft Glasses: Suzanne Fielding