Week 2: Getting the Ball Rolling!

Aryan R - March 7, 2025 4:09 pm

Hello everyone! Thank you all for reading my previous blog post introducing the overall goals of my future research.

Introduction

I am excited this week because I was able to receive the .cif files (Crystallographic Information Files) from Professor Botana containing the material characteristics of four different filled skutterudites. I used VESTA – a 3D visualization program for structural models – to convert those .cif files to POSCAR files in the VASP format. I have previous experience with this simulation package but opted to use Quantum Espresso in that previous experience due to copyright restrictions. I was able to join the VASP group within the Sol Supercomputer so now I have access to the VASP package in Linux, but also hopefully have access to proprietary material pseudopotentials.

VASP Specifics

With the POSCAR being provided by the Professor Khoury, VASP needs two other initial files to be able to run .scf (self-contained field) calculations. POTCAR files are the proprietary information that contain the pseudopotentials of specific elements. To create the POTCAR file needed to run my calculation, I combined the POTCARS of Sm, Sn, and Ru in that order. Additionally, I created the most basic INCAR file possible, setting ISMEAR = 2 and SIGMA = 0.2. (For more information about these specific inputs, reference the above link)

Results

In the Linux command terminal, I ran a preliminary test simulation using the “vasp_std” command to run the .scf calculation with the pre-prepared files above. No other files were used, but other important files such as the KPOINTS and CHGCAR will likely be implemented in the full-scale simulations in the future. Unfortunately the test case failed, with a slew of warnings and errors popping up. However, this is normal, as I expect multiple iterations to fail before being able to tackle the more complex cases. Right now, the most concerning error is the the “large supercell” error which suggests the atom is too large for the scope of our simulation.

However, this is a problem for next week. Look forward to learning about my Sol Supercomputer settings and the different parameters I can toggle within a VASP simulation to try to resolve errors.

Thanks, and see you all next week!

Comments:

All viewpoints are welcome but profane, threatening, disrespectful, or harassing comments will not be tolerated and are subject to moderation up to, and including, full deletion.

heet_d

Hey Aryan, nice blog post! You mention transitioning from Quantum Espresso to VASP for your simulations. Have you noticed any key differences or advantages between the two that could impact your results? By this I mean would you expect any differences in computational efficiency, accuracy, or any errors that you might face if you would have used QE here?

March 9, 2025 at 12:21 pm - Reply

aryan_r

That's a great question Heet! Both Quantum Espresso and VASP rely on the same DFT equations to initiate their simulations. As long as I use the same type of pseudopotential for each element (the input), the computational result and accuracy would be the same. There may be slight differences in timing, but most things stay constant.

March 10, 2025 at 10:02 pm - Reply