Computational Materials Research

Unlike many other FRI streams, all of the research performed in the computational materials stream is done on computers. This research can be divided into two major categories: applications and method development. Applications uses predeveloped software packages such as VASP and EON to simulate atomic systems and gather information about their chemical propeties and structure. Method development projects increase the efficiency and/or accuracy of the computational algorithms used to make predictions. As this research is computationally expensive, FRI students are given remote access to supercomputers to perform these calculations. Below highlights the major research projects and research results from the Computational Materials stream.

Computational Materials Database

One of the goals of the computational materials stream is to use high-throughput computing to find promising material candidates for catalytic reactions that occur in alternative energy resources. Catalytic activity can be predicted by calculating the binding energy of key intermediates of chemical reactions. Oxygen is a key intermediate for the oxygen reduction reaction (ORR) and Hydrogen is a key intermediate for Hydrogen evolution reaction (HER). Below are links to two databases created by the stream:

Random Bimetallic Database

Combining two transition metals in one material is a good way to create new material candidates as an alloy can have drastically different properties than either of the metals individually. Predicting the catalytic activity of random alloys can be difficult since various binding sites occur on the surface of materials. In this database we consider the hollow site of candidate materials and the various ensemble of atoms at the site (see figure below where A=green and B=blue).

If you would like to view how the catalytic activity changes with the ensemble of atoms, you can click Menu->Any material->Graph. Following this, click on a box for a metal pair and a figure should show up at the bottom of the screen. Below is an example figure for a PdAg alloy. The red line represents the ideal binding energy for optimal catalytic performance for the ORR.

Core-Shell Nanoparticle Database

Alongside the randomly alloyed materials, another common way to combine metals is in the form of a core@shell nanoparticle. In a core@shell nanoparticle, one metal in the core of a nanoparticle is completely surrounded by a shell of another metal (see figure below where the brown spheres are core atoms, the grey spheres are shell atoms, and the blue circle is the binding site). Unlike randomly alloyed nanoparticles, only one metal is present on the surface, meaning that we do not have to take into account multiple binding sites. To learn more about this project check out the FRI publication below called Computational Screening of Core-Shell Nanoparticles for the Hydrogen Evolution and Oxygen Reduction Reactions.

Computational Materials Software for Global Optimization

Another goal of the computational materials stream is to explore the the efficiency and accuracy of existing methods for global optimization and develop new methods for finding the global minima. We have created software which combines two global optimization methods, basin hopping and minima hopping, into one package and allows the user to either run the original methods or a combination of these two methods. You can find more about this software at the following links:

Software for Global Optimization

Global Optimization Database

'Feel the Force' Haptic Device Software

Our stream is creating a graphical user interface for a project called ‘feel-the-force’ in which forces from chemical and material systems are translated to a haptic control device where the user can construct and investigate materials at the atomic scale. The haptic feedback will allow the user to gain intuition about the function of materials and feel the mechanics of chemical bond breaking and formation. We are also currently working on developing an atomic force simulator in the back end, based upon machine learning models fitted to quantum simulations of atomic interaction. You can find the current version of our software below:

'Feel-the-Force' Software`

Kenitic Database

The Henkelman group is creating a kinetic database, the KDB. Kinetics study is important in that it is critical in determining reaction rate, by finding activation energy of reactions, a.k.a. energy barriers, through finding saddle points on the energy surface. This can be up to several orders of magnitude more costly compared to a single optimization calculation. However, not all ab-initio saddle searches need to start at scratch. One can speed up the saddle search by starting from known saddles of similar geometries. Therefore, we need a database of saddles (transition states) so that when saddles need to be found, one could check with the KDB to hope to start with a converged saddle to reduce computational costs.
Instructions on how to participate in the project is here.

Other Projects

Other areas of research in the computational materials include exploring local optimization methods in order to find a more efficent method of finding local minimums. Current work includes research into optimizing Newton's method of local optimization for use in chemical systems.

The computational materials stream also explores methods for calculating rates or recations in solid state systems. Determing these rates by simulating these reactions is computational expensive, so we have explored methods which are more efficent but sacrifice some accuracy. This work has mainly focused on improving Harmonic Transition State Theory.

The stream has also done work with other streams including the nano materials lab, where computational results were compared with experimental results from the lab.

Recent Student Publications and Presentations

FRI students are in bold below:

Presentations

1. Le, M. and Stockton, G. Modifying the Atomistic Machine-learning Package for Real-time Atomic Simulations. Poster presentation at the Texas Advanced Computing Center (TACC) Symposium for Texas Researchers. September 26^th, 2019; Austin, TX.
2. Clarke, R. Improving Computational Methods for Finding Reaction Rates. Poster presentation at the Texas Advanced Computing Center (TACC) Symposium for Texas Researchers. September 26^th, 2019; Austin, TX
3. Spear, L. Reactivity and Tunability of Bimetallic Materials. Poster presentation at the Texas Advanced Computing Center (TACC) Symposium for Texas Researchers. September 27^th, 2019; Austin, TX.
4. Bandaranaike, B. , Biswas, S. , Rajkitkul, P., and Harlan, L. Generating Potential Energy Surfaces through Machine Learning. April 12^th, 2019; Austin, TX. Poster presentation at the Spring Undergraduate Research Forum at UT Austin.
5. Spear, L. Reactivity and Tunability of Bimetallic Materials. Poster presentation at the UTSA College of Sciences Research Conference. October 5th, 2018; San Antonio, TX Winner of the best visiting student award
6. Spear, L. Reactivity and Tunability of Bimetallic Materials. Oral presention presented at the Fall Undergraduate Research Forum at UT Austin. September 29^th, 2018; Austin, TX. Winner for the best chemistry presentation
7. Clarke, R. Improving Computational Methods for Finding Reaction Rates. Oral presention presented at the Fall Undergraduate Research Forum at UT Austin. September 29^th, 2018; Austin, TX.
8. Sansom, J. Investigating Methodology for Global Optimization. Poster session presented at the Undergraduate Research Forum at UT Austin. April 13^th, 2018; Austin, TX. Winner of the FSTI Award for Excellence in Chemistry
9. Edmonds, G., Spear, L., Olivares, K., and Sun, R. Reactivity and Tunability of Bimetallic Materials. Poster session presented at the Undergraduate Research Forum at UT Austin. April 13^th, 2018; Austin, TX.
   
10. Bandy, R. Investigating Methodology for Global Optimization. Poster session presented at: the American Association for the Advancement of Science (AAAS) Annual Meeting; February 18^th, 2018; Austin, TX. Winner of the 2018 Student E-Poster Competition in the Technology, Engineering, and Math category!
11. Edmonds, G. Reactivity and Tunability of Bimetallic Materials. Poster session presented at the American Association for the Advancement of Science (AAAS) Annual Meeting; February 18^th, 2018; Austin, TX.
12. Zhou, K. Bounded Reversible Work Transition State Theory Poster session presented at: the American Association for the Advancement of Science (AAAS) Annual Meeting; February 18^th, 2018; Austin, TX.
13. Bandy, R. Investigating Methodology for Global Optimization. Poster session presented at: Institute of Pure and Applied Mathematics workshop on Optimization and Optimal Control for Complex Energy and Property Landscapes; October 2^nd, 2017; Los Angeles, CA.
14. Harlan, L. and Edmonds, G. Reactivity and Tunability of Bimetallic Materials. Poster session presented at: Institute of Pure and Applied Mathematics workshop on Optimization and Optimal Control for Complex Energy and Property Landscapes; October 2^nd,2017; Los Angeles, CA.
    

Publications

1. K. Barmak, J. Liu, L. Harlan, P. Xiao, J. Duncan, and G. Henkelman, Transformation of Topologically Close-Packed β-W to Body-Centered Cubic α-W: Comparison of Experiments and Computations, J. Chem. Phys. 147, 152709 (2017). DOI
   
   
2. B. Corona, M. Howard, L. Zhang, and G. Henkelman, Computational Screening of Core-Shell Nanoparticles for the Hydrogen Evolution and Oxygen Reduction Reactions, J. Chem. Phys. 145, 244708 (2016). DOI