S16 Landing Banner 1400x275

Tutorial TC04—Long Timescale Atomic Simulations in Materials Science

Nov 26, 2017
07:30 AM - 04:00 PM
Hynes, Level 2, Room 202

As microstructural evolution often occurs on timescales that are inaccessible to standard atomistic simulation techniques, such as molecular dynamics, a more appropriate conceptual framework needs to be adopted in order to accurately predict the performance of materials on experimentally relevant timescales. This tutorial will introduce two such sets of tools, one theoretical and one computational, which are invaluable to bridge the timescale gap between theory, computation and experiments.

The first part of the tutorial will introduce the fundamentals of transition state theory (TST) and discuss its applications in materials science as a crucial tool to design materials with the appropriate properties. The second part of the tutorial will show how rate theories can be leveraged to develop simulation techniques that address the timescale problem traditional methods, for example, molecular dynamics, face when dealing with infrequent event processes.

The tutorial targets materials scientists, both students and more senior researchers, who are familiar with the basics of atomistic simulations. No specific knowledge of long time methods is assumed. 

8:30 am – 12:00 pm
Part I: Hannes Jónsson
Introduction to the Fundamentals of Transition State Theory (TST)

TST was developed in the 1930s as a means to explain the reaction rates of elementary chemical reactions. Since then, much development has been carried out in the field with an extension to solid state physics and processes such as atomic diffusion in solids. In this context, harmonic transition state theory has been extremely useful for the community to find the mechanism and estimate transition rates. Powerful methods for finding saddle points on multidimensional energy surfaces have been developed and used in kinetic Monte Carlo simulations of long timescale evolution of materials. Examples will include simulations of grain boundary structure, atomic diffusion at grain boundaries, dislocation formation and glide and crystal growth. A recent extension to magnetic transitions will also be presented. This part of the tutorial will be presented by Professor Hannes Jónsson who has extensive experience on the subject and has developed some of the widely used methods.
(10:00 am – 10:30 am  BREAK)

12:00 pm – 1:30 pm  BREAK

1:30 pm – 5:00 pm
Part II: Danny Perez
Introduction to How Rate Theories can be Leveraged to Develop Simulation Techniques that Address the Time Scale Problem Traditional Methods Face when Dealing with Infrequent Event Processes

The focus will be on so-called accelerated molecular dynamics (AMD) methods, namely Parallel Replica dynamics, Hyperdynamics, and Temperature Accelerated dynamics, with special attention devoted to techniques that can leverage current and upcoming massively parallel platforms, such as Parallel Trajectory Splicing. The theoretical framework underpinning these methods will be discussed and targeted applications relevant to Materials Science, such as crystal growth, plastic deformation and radiation damage, will be presented. These will serve to illustrate how advanced techniques can be used to simulate the evolution of materials over very long timescales, often revealing unexpected insights into the microscopic underpinnings of microstructural evolution, with accuracies approaching that of direct fully atomistic simulations. This introduction will conclude with an overview of the software tools that can be used to carry out AMD simulations.
(3:00 pm – 3:30 pm  BREAK)

Instructors

  • Danny Perez, Los Alamos National Laboratory
  • Hannes Jónsson, University of Iceland