Hello,
I need perform integration of the local DOS from the valence-band bottom to the Fermi level, for get the number of electrons per atoms in local DOS in this interval.
I'm using VASP.
I have calculate DOS following the procedure suggested in vasp documentation.
I plotted DOS using p4vasp.
Also is mention that there are two type of Oxygen in the structure studied, but I do not have any idea how I can identify two different atoms type of the same element using DFT.
How can I solve those calculations?
Thanks in Advance!
Dear Maximo,
if you need to integrate the local DOS in an energy interval from bottom of valence band (E1) to the Fermi energy (E2), you can simply compute it by using for example a trapezoidal integration, which returns a very good approximation of the number of electrons contained in that energy interval. In MATLAB there is a function called trapz(). I'm attaching a tutorial for the latter in the following.
You can expect that this integration has to return the total number of electrons contained in your unit cell (data which should be available in every DFT simulation starting info).
About the second question, I am not able to understand the reason why you have two different types of Oxygen atoms in the same unit cell. Do you mean different labels for two oxygen, for example, belonging to different sites (different sublattices, for example)? Could you give me more details?
Best,
A.
http://de.mathworks.com/help/matlab/ref/trapz.html?requestedDomain=www.mathworks.com
Dear Angelo,
Thanks your suggestion about trapz() help me very much. I used Octave that has this function, too. But the value is still far from reported values.
Let me explain better about the second question. I attached of paper that I'm using. In the paper Electronic structure section (page 2), mention the author mention about that, cite
"there are two types of O atoms, one with one Mn and two Cr neighbors, the other with one Cr and two Mn neighbors"
Part of the analysis of the structure is though this two types of O atoms.
My questions are:
How can I identify those types of the same atoms?
and one more complex thing, after the calculation How it suppose that I can calculate LDOS for those types of oxygen?
http://everest.iphy.ac.cn/papers/prb67.115130.pdf
Dear Maximo,
in order to identify the two "different" oxygen atoms involved in the same unit cell, you can simply label them as if they are two different chemical species or different types of atom. So your number of atoms has to be increased of 1, and you can choose to label O1 and O3 the two respective oxygen, belonging to different sites.
This allow you to see the possible different behavior in the: charge distribution around the two different sites, to map them. Although a separated LDOS representation of the two contributes it seems a little bit complicate, since I guess that unless their position and atoms bonded to, their behavior could be the same (physically), I mean, the orbitals involved will be the same, as well as the occupancy and the number of electrons involved.
In the paper, specifically in the Figure 4, they talk about the plot of "one" Mn atom, and the same for the 2 oxygen atoms. What did they mean by saying "one"?
Best,
A.
Label in POSCAR? and use the same pseudopotential two times in POTCAR for each O type?
I try this option, but VASP override the POSCAR with the atom found in POTCAR. There are another alternative to label? Can you give a real example?
The structure of the Figure 4 has 56 atoms, 8 Li, 8 Mn, 8 Cr and 32 O (O1, O2), but in LDOS plot is only take in account one atom of each species. Is about of symmetry, if you plot for one atom (Mn, O1, O3) for the other atoms of this type you have the same value.
For answering to your questions.
I'm not very familiar with VASP setting, but in principle it should be possible to run a simulation under this condition. Take a look on this link, I hope it can be helpful for your case:
http://atomsk.univ-lille1.fr/doc/en/format_pos.html
About the plot, if it's possible to represent the same structure reporting just the LDOS of one atom for each species, labeling them differently and mapping them as belonging to different specific sites, in the end the LDOS of an isolated O1 or O2 in the same relaxed position of the original structure could be provide the LDOS that you're looking for, without taking into account the presence of the other elements.
Should be possible, but not appear so trivial. I believe there are some tricks for perform this simulation that we ignore.
Thanks for your help. I'm going to check this link.
I agree with everyone else about how to do the integration. I usually open the DOSCAR in excel and do the trapezoidal integration myself.
However, as far as the two different oxygen atoms go I use the FINDSYM website from Brigham Young University
http://stokes.byu.edu/iso/findsym.php
It's a little confusing at first, but essentially you upload your POSCAR file and it creates a CIF file for you by determining which atoms are "equivalent" due to symmetry (e.g. only differ by simply symmetry transformation like translation or reflection). You can then compare their CIF to your POSCAR and see which atoms are unique. I use VESTA to do this. The atoms in the CIF are plainly labeled as Atom 1 or 2 in VESTA.
Dear Erik,
Thank for the recomendation, but when I upload my POSCAR and run I get the following error:
Reading CIF file... This program has bombed:
Error: no data found in CIF file. read_cif 39 exit
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs