Nudget Elastic Band Theory for oxygen vacancy problem

[berna_akgenc]

Dear All,

I would like to calculate activation energy path for Yttria Stabilized Structure (%8 mol Yttria) using Nudget Elastic Band method as implementation VTSTTOOLS.I am using the INCAR given below. I have some questions and confusions which i am putting as under.

1. If my INCAR tags are appropriate for the calculation or not.
SYSTEM = NEB-Calc
ISYM = 0 # switch on symmetry
Start parameter for this Run:
ISTART = 0
ICHARG = 2

Electronic Relaxation 1
ENCUT = 500 eV
IALGO = 38
EDIFF = 1.0E-6
NELMIN = 5

Ionic Relaxation
EDIFFG = -1.0E-3
NSW = 500

DOS related values:
ISMEAR = 0
SIGMA = 0.05

Cell Volume + Cell shape
ISIF = 3

Other Calculation parameters
PREC = ACCURATE # Advised to do so
LREAL = .FALSE.
LWAVE = .FALSE.
LCHARG = .FALSE

NEB Related
IMAGES = 9 # Number of NEB images between the fixed endpoints
SPRING = -5.0 # The spring constant, in eV/Ang^2 between the images; negative value turns on nudging
ICHAIN = 0 # Indicates which method to run. NEB (ICHAIN=0) is the default
LCLIMB = .TRUE. # Flag to turn on the climbing image algorithm
LNEBCELL = .TRUE. # Flag to turn on SS-NEB. Used with ISIF=3 and IOPT=3

IOPT = 3 # QM (Quick-Min) force based optimizers is used and good for high force far from minmum

IBRION = 3 # Disbale VASP default optimizers which are basesd on Energy
POTIM = 0 # Disbale VASP default optimizers which are basesd on Energy

TIMESTEP = 0.1 # Dynamical time step (Default comes with IOPT=3)

LSCALAPACK = .FALSE.


2. I am moving to oxygen vacancies, How does VASP understand between intersititial atoms site and vacancy differences. Other words, I am creating oxygen vacancy as delete some oxygen. What is different of moving of oxygen vacancy and atoms at bulk structure. I just wanna make sure my results accuracy form you.

Many Thanks!

[berna_akgenc]

According to my acknowledge, POSCAR of reactant and POSCAR of product should be optimized,  I have optimized them in other directory with a classic method (ISIF=3). I got and error message and then I used ISIF=2 for optimization. Then, I have copied CONTCAR and OUTCAR files as POSCAR00 and POSCAR10.

Then I run this command,

nebmake.pl POSCAR00 POSCAR10 9

The command creates 00,01...09 directory which included just different POSCAR file.

Now, how I can run the program? Should I copy KPOINTS POTCAR INCAR at every single sub-directory. And run them individually?
Could you take a look my files please?

Thanks

[graeme]

I think you basically asked this question previously. You are not approaching this NEB calculations correctly. Here are some simplified instructions:

1. Minimize the endpoint initial and final states.
2. Use nebmake.pl to generate an initial band. Note, do not do this in your submit script; just run it from the command line.
3. Put your INCAR, POTCAR, and KPOINTS file in the parent directory, i.e. where the 00 to 10 directories are, each containing a POSCAR file.
4. Set the IMAGES=9 tag in your INCAR (as you have).
5. Run a _SINGLE_ copy of vasp from that parent directory, allocating sufficient cores. With 9 images, the number of cores should be an even multiple of 9. With the IMAGES tag set, MPI will spawn copies of vasp to run in the subdirectories. Again, I emphasize, this will be done for you by vasp, do not start many copies of vasp in the 00 - 10 subdirectories.

[berna_akgenc]

Dear Graeme,

I really appreciated your help, I catch up every single step, except one thing, what does it mean of minimize the endpoint initial and final states, How I can do that? Do you mean that use of InterpolatesPOSCAR?

Step 1: I just relaxed my initial and final states via ISIF=2 and I copied CONTCAR files as POSCAR00 and POSCAR10
Step 2: then I use nebmake.pl command in the parent directory.

Should I do something in these two step? And how?

Thank you so much,
Berna

[graeme]

You minimize the initial and final state (endpoint) structures as you would any other structure using vasp. It sounds like you did that, so perhaps the only problem was how you were running your neb calculation.

[berna_akgenc]

I didn't actually. I just used my CONTCAR files as POSCAR00 and POSCAR10. The only problem is this. How I can minimize the initial and final state's endpoint?

Could you explain it to me ?

Thank you,
Berna

[graeme]

You do a vasp calculation where you minimize the energy of the initial and final configurations. Those are the geometries that you want to generate an initial NEB path.

One more note: for the very basics of doing these calculations, it may be better to sit down with someone local at TAMU who does this and can step you through it. Students in Prof. Perla Balbuena's research group, in Chemical Engineering, for example, calculate reaction barriers routinely.

[berna_akgenc]

Thank you!
I've figure it out ;)

Berna