Navigation and service

Multiscale Modelling Methods for Applications in Materials Science

Contents and objectives

The knowledge and experience with the new multiscale techniques (such as sequential/hierarchical modelling and hybrid methods) and tools (such as the tools for sequential modelling developed in MMM@HPC) should be disseminated to further groups of the community. In particular postgraduate students and postdoctoral researchers entering the field are targeted by this tutorial.

This tutorial will include five half-day lecture sessions each with four lecture hours. In these sessions methodological introduction to the theory will be made and in some lectures supplemented by use cases of several current high-profile applications, such as modelling of materials for energy conversion and storage and materials for next generation electronics.

The lectures will be comprehensive but still require substantial knowledge in at least two of the theoretical methods employed in multiscale modelling (e.g. quantum mechanics, statistical physics, theoretical chemistry, theoretical solid state physics, Monte Carlo simulations and molecular dynamics). Particularly, the lectures will focus on combining different approaches to treat concrete applications (organic light emitting diodes, Li ion batteries, 1D and 2D carbon nano-devices, polymer electronics, molecular electronics etc.). The lectures will sufficiently explain the physical interrelations between different scales and introduce best practices in combining the methods. In order to cover the coarse-grain and the continuum scales we have considered involving established experts from materials research (alloys, catalysts, polymers).

The multiscale modelling often requires novel implementations and practical approaches for performing computer simulations in a more complex software – and sometimes even hardware – environment. In contrast to standard approaches that have been used for many years in the community these new approaches, e.g. such based on the UNICORE system, expose the physics aspects of the models while hiding the technicalities of the underlying computer infrastructures and still allowing experts for the included sub-models to extend and optimize the methods. First practical experience with such an approach is essential to undergird the acquired knowledge during the lecture parts. For this reason, two half-day training sessions are organized in cooperation with the MMM@HPC project. The MMM@HPC project has developed an integrated platform for multiscale materials modelling. A full set of training material was developed, which is available upon request and which will be part of the proposed tutorial. The Jülich Supercomputing Centre, which is part of the Jülich CECAM Node, provides high performance computing resources (JUGENE, JUROPA), services (UNICORE) and qualified support which are necessary for the realisation of the training.

We emphasize that it is planned to publish Lecture Notes, which will be freely available upon request or via Web access.