I'm studying DFT calculation using Quantum Espresso. It has some limitations in calculating more accurately than other paid programs such as VASP. I saw some published papers using Quantum Espresso. However, the result was quite different when I calculated my material and compared it with the reference paper that used VASP.
I calculated the DOS of Lithium Vanadium Phosphate using Quantum Espresso. The structure of LVP was drawn by analyzing the XRD results of the synthesized LVP using Fullprof. The pink line in the figure represents the calculated result of the LVP structure, and the background figure is reference data. I guess the difference comes from the different atomic structures. However, I think it also comes from the difference of program, pseudopotential, and something like that. Does it actually matter?
These are the pseudopotentials that I used
O_ONCV_PBE-1.2.upf
V_ONCV_PBE-1.2.upf
Li_ONCV_PBE-1.2.upf
P_ONCV_PBE-1.2.upf
The reference paper is https://doi.org/10.1007/s10008-020-04808-7
Hello Yoonju Oh.
There are many factors that can change the DOS result in Quantum ESPRESSO. The number of K points, occupation type and energy cuttof can lead to results different from the expected. Could you share your input and the correct DOS result, so I can help you with this?
Best regards,
Ricardo Tadeu
Hello, Ricardo Tadeu.
I really appreciate your answer, sir. Although I have studied this field for more than two years, I am still a beginner. I'm recalculating it by changing the pseudopotential and K-points. Anyway, I have attached my input to this comment. Thank you for your time and reply again.
Sincerely,
Yoonju Oh
I think the structure you provided is not correct, as I show in these pictures below. I made another one, and I would advise you use this new input I attached. When we build Quantum ESPRESSO inputs, we should definitely rely on symmetries. Quantum ESPRESSO identified 2 symmetries elements in your input, while in this new, QE identified 4, and this contributes to lower time of calculation. I tried to do a DOS calculation, but I found it very expensive, once there are few dozens of atoms. Do you use any cluster to do it?
So, this new structure I made using the CIF file attached, and a software called wxDragon. It's not a very famous software, but it's free and can convert from CIF to Quantum ESPRESSO input. Please tell me if this new input works better.
Bests
Ricardo
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication