I am working on an example for InAs. The ultimate goal is to study the thermal properties from the phonons. But first, I must optimize the structure with VASP. It seems like a very easy calculation however the results from the paper I am following shows the lattice parameter to be 6.06 angs and my results show 6.21 angs. It is very crucial for me to be very good at optimizing my structure in order to get accurate results from the phonons. Below I will provide some info of my input files:
INCAR:
PREC = Accurate
ENCUT = 311 --> at least 30% above ENMAX from POTCAR
EDIFF = 1e-7
SYMPREC = 1e-6
IBRION = 1
NELMIN = 1
ISIF = 3
NSW = 99
ISMEAR = -5 -> semiconductor
KPOINTS:
MP - 11x11x11 k mesh
POTCAR:
psuedopotential -> PAW_PBE
POSCAR:
In As
1.0
0.000000000000 3.018000000000 3.018000000000
3.018000000000 0.000000000000 3.018000000000
3.018000000000 3.018000000000 0.000000000000
In As
1 1
Selective dynamics
Direct
0.000000000000 0.000000000000 0.000000000000
0.250000000000 0.250000000000 0.250000000000
It is very interesting to me that I start from a structure that is very close to the ideal structure and yet my results seem to get farther from the ideal lattice parameter. If anyone has some insight on this problem I would greatly appreciate it.
Hey Robert Appleton ,
You're right that the experimental lattice parameter of InAs is 6.06 Å. A good reference for you in the future is the Materials Project, which contains DFT-calculated data for too many materials to count! Here is the page for the ground state of InAs (https://www.materialsproject.org/materials/mp-20305/#). While this page reports the primitive lattice parameter of 4.371 Å, you can download the conventional standard cell by clicking on this button (see attached image). The DFT-calculated lattice parameter of InAs is 6.18148 Å, which is in good agreement with what you've calculated already. This could just be the case of a material for which the PBE XC functional isn't great.
With that being said, I have one concern with your input file. Your ENCUT is a little small; Materials Project typically uses 520 eV. If you want to use a smaller ENCUT, then you should do a convergence test. If PBE doesn't accurately predict the lattice parameter, then I recommend trying the SCAN XC functional (Article SCAN: Strongly Constrained and Appropriately Normed Semiloca...
), which is a natural extension of PBE that doesn't cost much more but works well for solid-state materials.Please let me know if you have any further questions.
Best,
Rob
Daniel Hashemi
Hello thanks for replying and yes I actually did what you are suggesting. I ran a script that looped through lattice parameters starting around 6.0 - 6.5 angs and calculated the energy. When I plotted the Energy vs Lattice Parameter it showed a minima about ~6.21 angs. I attached the plot as an image.
Robert Appleton Thanks, the plot seems fine. As Robert Wexler mentioned above increase the ENCUT (or use PREC=High which only increases the energy cutoff by 30%) but you have more control by just increasing ENCUT. But I wouldn't expect much difference.
Okay then for my understanding, it is good practice to do pick a reasonable k-mesh and a reasonable encut and relax the volume. Then do k-point convergence and encut convergence tests. Then to be very precise we should do another volume relaxation at the new k-mesh and encut from the convergence test.
These calculations are not very computationally expensive but I know my optimization is very crucial for studying the phonons.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Semiconductor