Bridging-Time Scale Techniques and their Applications in Atomistic Computational Science

Workshop Report

 

The BRITS International Workshop was held from 11-15 September 2016 with 50 participants from 14 different countries. In total we had 39 contributions (1 Colloquium, 30 oral presentations and 8 Posters. The theme of the meeting was the subject of how to perform dynamical simulations in atomic systems over time scales that are longer than those computationally accessible by Molecular Dynamics (MD). At the meeting were a mixture of disciplines and talks and posters varied from descriptions of new algorithms and theoretical underpinning of the techniques, to exciting new applications in materials science, chemistry, biology and physics. The applications included chemical reactions on surfaces, transport mechanism across cells and the long-term evolution of radiation damage.

 

Many techniques were discussed including new ideas of parallel trajectory splicing, using the power of modern computers with their thousands of processors, adaptive kinetic Monte-Carlo (KMC) methods, stochastic surface walking techniques, course graining using metadynamics and free energy methods.

 

For the KMC methods, the determination of saddle points on n-dimensional surfaces is still the subject of current research since the algorithms that are used are always a compromise between speed and robustness. Algorithms described by mathematicians that work for small systems are not always practical to use for large systems. Even after the meeting it was still not clear how to extend precise results obtained using transition state theory on small systems to the larger practical systems that many materials scientists wish to study, despite this being a topic of the discussion session.

 

For the larger systems where free energy methods are used rather than detailed atomistics, the use of collective variables and the metadynamics technique was discussed in detail. An interesting new development in this area was the discussion of how an algorithms might be used to determined the best collective variables for the systems but more work needs to be done to provide a consistent link to the atomistic level.

 

Presentations by Established Scientists

The colloquium talk was given by David Landau of the University of Georgia, who gave an excellent overview of the Wang-Landau method and its application to Monte Carlo techniques. This technique was applied in a number of problems described by the other participants of the meeting. The colloquium was very well attended by many of the local members of the MPI and lively discussion followed the talk.

A large group of participants from Los Alamos National Laboratory, led by Art Voter presented significant advances in the use of multiprocessors to advance new techniques while others applied these techniques to important systems in materials science.

Similarly, Normand Mousseau described recent advances with his ART method, which was used in other presentations in practical applications.

 

Presentations by Young Scientists

There were a number of presentations that were from younger participants. There were excellent talks from Fudan university from chemists who had used their in-house methods, to determine reaction pathways and to investigate the structure of molecular crystals.

The metadynamics methodology mentioned above was the idea of a young researcher, Pratyush Tiwary about to start his first academic staff position at the University of Maryland.

In addition Abhijit Chatterjee from Mumbai presented and interesting variation of temperature accelerated dynamics where the temperature was raised in steps to speed up the rare event processes.

Posters from PhD students included a description of how events with low transition barriers can be included in a mean rate theory in a lattice-based KMC method to model the growth of thin films.

 

The organisers received a number of positive comments from the workshop participants many of whom have uploaded their presentations to the workshop web site. However, the success of the meeting was not just because of the quality of the scientific programme but also in a large part due to the MPI for the wonderful facilities offered, including a very nice outing along the river with dinner at the Fortress Koenigstein and especially to Amy Wright for her perfect administrative input.