Energy vs Volume for Silicon Diamond in VASP

[jess]

Hello there,

I am new in VASP; For now, I have to calculate the Ground state energy vs volume for Si Diamond;
I try it for 2 atoms in POSCAR, and I still need some datas to get a parabol for my plots; I change my lattice constants also but I got the same results !!

What should I do for my program so that I can get a parabol in the energy in function of volume ?

Thanks so much for your help !!

Thanks again !!

[Danny]

If you get exactly the same results you are doing a geometry optimisation which also modifies the unitcell. What you wish to do are static calculations (you can allow the atoms only to change positions).
For static calculations set
IBRION=-1
NSW=0
For relaxations set
IBRION=2
NSW=100 (most systems will converge much faster than this)
ISIF= ... (pick the appropriate value http://cms.mpi.univie.ac.at/vasp/vasp/ISIF_tag.html )

Danny

[jess]

Hello,

Thank you !! I change my IBRION = -1 and NSW = 0 in INCAR and varies my lattice constant in the loop.sh from 4 step 0.2 to 6 ; but OSZICAR only gives me one value for the ground state energy ; so I set up my IBRION =2 and NSW = 100 and ISIF = 3 , and it gives 8 values but all values are negative and very close to each other, it does not give me a parabol curve; I need to get a parabol with different values; how can I get that ? What should I change ?

Thank you for your help !!

Thanks again !!

[Danny]

1) the OUTCAR file contains the volume of your cell, or you can calculate the volume of your cell via the lattice vectors used, so you have all info you need.
2) the ground state energies should be negative, if they are not your compound is NOT stable.

Danny

[jber]

You could also use ASE for that. You need to install the latest trunk https://wiki.fysik.dtu.dk/ase/download. ... nt-release ). Then run (see https://wiki.fysik.dtu.dk/ase/ase/cmdli ... lk-systems):

ase vasp bulk Si --cubic --crystal-structure="diamond" --lattice-constant="5.47" --fit="5,1" -p "xc=PBE,kpts=(10,10,10)" -l
ase vasp bulk Si --cubic --crystal-structure="diamond" --lattice-constant="5.47" --fit="5,1" -p "xc=PBE,kpts=(10,10,10)" -s

which results in:
name E E-E0 V0 B
eV eV Ang^3 GPa
Si -43.302 -0.001 163.853 87.702

so the lattice constant a0=163.853**(1./3) = 5.472 A,
what is close to a0=5.469, B0=87.8 reported in
http://jcp.aip.org/resource/1/jcpsa6/v1 ... p154709_s1

For manual analysis, the created Si-bulk-vasp.json file contains energies and so called strains,
dtermined by the --fit keyword.
You can get volume from strain by strain**3 * (initial volume):
$ python
>>> strains=[0.98999999999999999,0.995,1.0,1.0049999999999999,1.01]
>>> a=5.47
>>> volumes = [s**3 * a**3 for s in strains]
[158.80623983957699, 161.22456774580962, 163.66732299999998, 166.13462835264031, 168.626606554223]

Then use eos tool (http://cms.mpi.univie.ac.at/vasp-forum/ ... hp?4.10787 )

You could also follow http://cms.mpi.univie.ac.at/vasp/vasp/diamond.html if you have access to it.

[jess]

Hello again,

Thank you !! I got my results but I still have some questions: I got my results of Ground state energy from the OSZICAR file and the volume cell from the OUTCAR file; while I need to know how to attribute the correct value of my ground state energy to the correct value of the volume cell given in OUTCAR;

I mean: How will I know this value of Ground state energy in OSZICAR will be attribute to this value of volume cell from OUTCAR ? ( because OSZICAR gives me the results of Ground state energy but not attribute these values to the volume cell value; I do not want to make a mistake to attribute the wrong value of ground state energy of the wrong volume cell; so why I am asking that to be sure )

Thank you very much for your help !!

Thanks again !!

[Danny]

SInce you should be doing a static calculation after each relaxation, the volume given in the outcar is the one belonging to the groundstate energy obtained in oszicar. you get a set of data of the type(groundstate energy, volume), and you fit these to an equation of state, which will give you a parabolic (resembling) curve, of which the minimum (again a data of type (groundstate energy, volume) ) will represent the properties of the groundstate within that framework. (these should be quite close to those obtained if you do a good relaxation of your system)

[jess]

Hello,

Thank you very much ( Sorry for late replying ) !! I got all my results satisfactory !!

Thanks again !!