[mehr]

Self-organization and assembly processes are crucial steps in the making of a wide range of materials and, in turn, have a great impact on their performance. For instance, the crystal structure, or polymorph, that forms during nucleation often dictates the bioavailability of pharmaceutical drugs, or the mechanical and catalytic properties of metal alloys and inorganic nanoparticles. In biology and medicine, protein folding and aggregation processes play a major role in the onset of many neurodegenerative disorders. Similarly, active, self-propelled, objects can form unexpected structures such as colloidal rotors on the micron scale, or bacterial biofilms, bird flocks and swarms of unmanned aerial systems on the macroscopic scale. While recent advances in experimental, theoretical & computational methods have allowed for unprecedented insights into the behavior of nonequilibrium systems, a complete understanding of these processes has remained elusive so far. For example, it is still impossible to predict which crystal structure forms when a liquid crystallizes. Similarly, the elucidation of the rules of life of swarms and active assemblies remains an outstanding challenge, although it is a necessary starting point to the successful development of soft matter robotics. In this talk, I discuss how my research group leverages computational materials science and artificial intelligence to shed light on assembly, cooperativity, and emergence in hard, soft and active matter. I show how recent advances in statistical mechanics and ML-guided simulations shed light on assembly pathways in materials and biological systems. I finally highlight how data science and machine learning methods provide a new way to accelerate discovery in soft autonomous robotics technology. [mehr]

Innovative solar system missions must become increasingly innovative and elaborate since "the low-hanging fruits have already been picked." Solar sails, which are propelled solely by solar radiation pressure, are among the key technologies for the future exploration of the solar system because they make missions possible that would otherwise be infeasible due to their immense propellant requirements. The optimization of solar sail trajectories, however, is a difficult task. In the talk, a method is presented that is based on machine learning, fusing artificial neural networks and evolutionary algorithms. Such optimization methods may also be applied for subsurface ice melting probes, as they are required to explore Jupiter's and Saturn's icy moons, which may harbor life in the oceans beneath their thick ice crusts. Such ice melting probes have been developed at FH Aachen and successfully tested in Antarctic ice. It will be interesting to discuss whether those methods can also be applied in polymer research. [mehr]

Accurate structural models of biological systems can be obtained by integrative approaches that properly combine multiple sources of information, such as experimental data and a priori physico-chemical knowledge. In this talk, I will give an overview of the methodological approaches that we have been developing as well as a series of applications to systems of outstanding biological importance. Specifically, I will focus on the determination of structural ensembles of intrinsically disordered systems using NMR data and on the use of cryo-electron microscopy data to unravel the continuous dynamics of flexible parts of ordered systems. Finally, I will present an open-source, freely-available module of the PLUMED library (www.plumed.org), which enables the simultaneous determination of structure and dynamics of conformationally heterogeneous systems by integrating experimental data with a priori information. [mehr]

[mehr]

[mehr]

Block copolymers are known for their elaborate microphase separating capabilities. Due to the many tuning parameters a predictive modelling approach is required. Molecular dynamics for such large system sizes is expensive. On the other hand, self-consistent field theory is capable of simulating much larger systems, but its mean-field based approximation only becomes correct when the chain density becomes unrealistically high. As a result some experimental effects such as a first order order-to-disorder phase transition are not reproduced. Field-theoretic simulations, where fields are not constrained to their mean field value but allowed to fluctuate, bridge the gap between these methods. I will give an introduction into this method, with some of the associated phenomena such as the ultraviolet divergence, and successful applications in symmetric block copolymers and block copolymer-homopolymer blends. [mehr]