Atomistic simulations for materials science

Course goals

Upon completion of the course, students will have developed the ability to interpret results from atomistic simulations and judge whether a materials chemistry research question can be reliably addressed with simulations. They will also have obtained basic working knowledge in the usage of standard simulation software. Students will complete five applied projects, and learning progress will be determined based on their project reports. 100% of the course grade will be based on the reports. There is no exam. The report for each project will count 1/5 of the final grade. 1/5 of the final grade will be determined based on a presentation of the results of project 4.
 

Content

Many materials properties and chemical processes are governed by atomic-scale phenomena such as phase transformations, atomic/ionic transport, and chemical reactions. Thanks to progress in computer technology and methodological development, now there exist atomistic simulation approaches for the realistic modelling and quantitative prediction of such properties. Atomistic simulations are therefore becoming increasingly important as a complement for experimental characterization, to provide parameters for meso- and macroscale models, and for the in-silico discovery of entirely new materials. In recent years, these conventional simulation methods have been further complemented and enhanced by data-science and machine learning. This course aims at providing a comprehensive overview of cutting-edge atomistic modelling and machine-learning techniques that are frequently used in materials chemistry. Students will learn the theoretical background of various simulation techniques ranging from empirical to first-principles approaches and will get hands-on experience in using state-of-the-art (free and open source) simulation software. Practical examples will focus on inorganic crystalline materials and their surface science, though the methods taught are transferable to other materials classes as well. Simulation results will be analysed with Python scripts and open-source data-science packages. Contact with industry will be established through guest lectures. Course grades are based on reports for practical assignments.

Programme and schedule
The course consists of four modules: (1) Fundamental Atomic Interactions, (2) Density-Functional Theory for Computational Surface Science, (3) Molecular Dynamics Simulations, and (4) Data Science and Machine Learning for Materials Chemistry. In each module, the required technical background is first briefly introduced, which is followed by a project assignment. Students have roughly two weeks to complete each project.