Question
Asked 7th Aug, 2021

In pbc calculations for impurity in a supercell, how to minimise the effect of impurities from surroundings supercell images?

How to minimize the effect of impurities from surroundings supercell images on our model supercell? My concern is about the surrounding periodic supercell images where also the modeled impurity is present at same locations. I want to minimize the effects from those supercells only. I can use non pbc calculation in Gaussian09 or similar package but there may be convergence failure as the lattice is stable in pbc only. Another method is to put vacuum around the molecules in the supercell to reduce the interaction but that is taking more resource. Please suggest if there is any other way.

Most recent answer

15th Sep, 2021
Kushagra Agrawal
Cardiff University
Chirantan Pramanik I would suggest looking at any one parameter (say surface energy) and increasing the cell size from 3x3x3 to 4x4x4, and checking the difference. Then I would further increase it to 5x5x5 and check the difference between 4x4x4 and 5x5x5 supercell... and so on. As the effect of impurity reduces, the deviation in the parameter will also start to reduce. Based on your computational resources, you can then decide on the cutoff cell size.

All Answers (5)

8th Aug, 2021
Muhammad Hamza El-Saba
Ain Shams University
Dear Chirantan,
Choose a sharp screening (Coulomb) potential for impurities, such as the Fukuyama potential.
Good luck
1 Recommendation
9th Aug, 2021
Chirantan Pramanik
Weizmann Institute of Science
Thank you Muhammad Hamza El-Saba . Choosing sharp screening potential will reduce the impurity effect in the modeled super-cell as well. My concern is about the surrounding periodic supercell images where also the modeled impurity is present at same locations. I want to minimize the effects from those supercells only. I can use non pbc calculation in Gaussian09 or similar package but there may be convergence failure as the lattice is stable in pbc only. Another method is to put vacuum around the molecules in the supercell to reduce the interaction. Please suggest if there is any other way.
14th Sep, 2021
Kushagra Agrawal
Cardiff University
You need to make sure that you use sufficiently large surface to negate the self-interaction from the impurity of the adjacent cell. To do so, you will have to first quantify this impurity effect in some manner, and then increase the size of the surface till the impurity effect is within your tolerance. Gaussian is very poor for periodic surfaces. I would recommend using CP2K (open-source) or VASP (paid) for periodic catalytic studies.
14th Sep, 2021
Chirantan Pramanik
Weizmann Institute of Science
Kushagra Agrawal thank you. I am using Quantum espresso and using 3*3*3 super-cell for certain lattice
15th Sep, 2021
Kushagra Agrawal
Cardiff University
Chirantan Pramanik I would suggest looking at any one parameter (say surface energy) and increasing the cell size from 3x3x3 to 4x4x4, and checking the difference. Then I would further increase it to 5x5x5 and check the difference between 4x4x4 and 5x5x5 supercell... and so on. As the effect of impurity reduces, the deviation in the parameter will also start to reduce. Based on your computational resources, you can then decide on the cutoff cell size.

Similar questions and discussions

Related Publications

Preprint
We present IrRep - a Python code that calculates the symmetry eigenvalues of electronic Bloch states in crystalline solids and the irreducible representations under which they transform. As input it receives bandstructures computed with state-of-the-art Density Functional Theory codes such as VASP, Quantum Espresso, or Abinit, as well as any other...
Article
ESTEST is a web-based framework for verification and validation of electronic structure computations [1,2]. It enables automatic comparison and post-processing of simulation data obtained using Abinit, Quantum-Espresso, Siesta, Exciting, Qbox and VASP. We present new features of ESTEST that extend its operation to a distributed network of servers....
Article
Most of the recent developments in EELS modelling has been focused on getting a better agreement with measurements. Less work however has been dedicated to bringing EELS calculations to larger structures that can more realistically describe actual systems. The purpose of this paper is to present a hybrid approach well adapted to calculating the who...
Got a technical question?
Get high-quality answers from experts.