SSNEB is not converging

[tanmoy]

Hi,
My name is Tanmoy and I am working as a PhD student in ICAMS, Bochum, Germany. I was doing some SSNEB calculations in my system (Titanium-Tantalum) and I had some nice results. But, recently, regarding one my SSNEB calculations, I faced some problem. What I observed that my SSNEB calculation is not converging. In my other calculations NSW reached max 50 steps, but in this calculation, it goes to 500 steps i.e, it is not converging. The energy is also not changing at all in all the steps. I have tried by changing the spring forces and the MAXMOVE tag, but didn't see any improvement. I am posting my INCAR file. Please give me some suggestions.

SYSTEM = Ti2Ta2
ENCUT = 225 energy cutoff [eV]
EDIFF = 1e-4 convergence of electronic relaxation[eV]
EDIFFG = -0.010 convergence of ionic relaxation [eV]
ISMEAR = 1 smearing (-5=tetrahedron with Bloechl, N=Methfessel-Paxton order N)
SIGMA = 0.2
ISIF = 3 stress tensor calc (0=no)
IBRION = 3
POTIM = 0.0
LCHARG = .FALSE.
LWAVE = .FALSE.
NSW = 500
ISPIN = 1 spin-polarized calc (1=no, 2=yes)
IALGO = 48
ISTART = 0 start mode (0=from scratch, 1=restart, 2=restart/constant basis set)
NPAR = 4
IOPT = 3
ICHAIN = 0
IMAGES = 5
SPRING = -4.0
LCLIMB = .FALSE.
LTANGENTOLD = .FALSE.
LDNEB = .FALSE.
LNEBCELL = .TRUE.
TIMESTEP = 0.05
MAXMOVE = 0.2
ISYM = 0

[graeme]

If you can post a .tar.gz file of the calculation, excluding WAV*, CHG*, we can see what's going on.

[tanmoy]

Hi Graeme,

Thank you so much for your reply. The calculation folder is too big, hence I am sending the .tar.gz file parts by parts. Please find it in the attachment.

Regards,
Tanmoy

[tanmoy]

Here is the second part, which includes no 2,3 and 4 images calculations. Last part will be followd this thread.

[tanmoy]

And here is the last part which includes the last part i.e, final image calculations and also a folder containing the directory of the images where the ssneb calculation goes on.

[graeme]

You don't appear to have our vtstcode linked into your vasp binary.

[tanmoy]

What should I do then? Kindly give me some suggestions.

[graeme]

You have to link our vtstcode into vasp for the ssneb to be functional.

See:

http://theory.cm.utexas.edu/vtsttools/download.html

and

http://theory.cm.utexas.edu/vtsttools/installation.html

[tanmoy]

Hi Graeme,

Thank you for you reply. I didn't even check that although I had calculated several SSNEB calculations before. Just by mistakenly, here I took the job file from a different calculation and hence it shows vtstcode is not linked with VASP. Anyway, after your reply I changed the job file as before and SSNEB worked fine.

Thank you once again.