NEB

[ccccc]

Hi everyone,

I want to learn the usage of NEB method and i have been trying to do a NEB calculation for the transition state of a molecule on (100) surface.
But i faced with some problems. I prepared an INCAR file :
SYSTEM = Butonol
IMAGES = 8
ISTART = 0
SPRING = -5
ICHARGE= 1
ISYM = 2
NELM = 60
NELMIN = 8
PREC = HIGH
LREAL = .TRUE.
NSW = 150
IBRION = 2
NPAR = 1

I read something in some papers like these: "In the first levels, typically more than 100 iterations of NEB were carried out until the forces perpendicular to the reaction path were considerably small." I want to know how i can specify the forces acting perpendicular to the reaction path. Actually i want to test whether the TS changes if i change that forces. But i dont know whether i can interfere it! can i do? Or can i just read its value from OUTCAR?

I found some parameters from vasp manual like EDIFFG, ISIF and LCORR. Can one of them help me what i want?
Thank you very much..

[graeme]

If you are using our code, you can just read the force from the OUTCAR. This contains both the perpendicular (real) and parallel (spring) forces. By grepping for NEB, you can also see the components. By specifying the convergence criteria, (e.g. ediffg=-0.01) you will require that the sum of these forces on each image is less than 0.01 eV/Ang.

A couple of suggestions for your INCAR file: use PREC=Normal, ISYM=0, LREAL=Auto, and IBRION=3, POTIM=0.2. You can use more aggressive optimizers (see http://theory.cm.utexas.edu/vtsttools/optimizers/) but the one optimizer that you can't use is the IBRION=2 - the one that you have specified.

[ccccc]

Dear Graeme
Thank you for your attention to my problem. I've changed the parameters and I'll check the results again according to your advices.

[ccccc]

I divided my initial and final geometries into 8 mid-geometries. And I finished my neb calculation with this INCAR
SYSTEM = neb
IMAGES = 8
ISTART = 0
SPRING = -5
ICHARG = 1
ISYM = 2
NELM = 60
NELMIN = 8
PREC = HIGH
LREAL = .TRUE.
NSW = 150
IBRION = 2
NPAR = 1
EDIFFG =-1e-03


But now at the end of the work, all final energies for this 8 geometries have gone to the same value. I see the same total E0 at the end of the all OSZICAR. files. What do you think the problem is??

Thank you:)

[graeme]

Did you freeze the bottom layers of your surface? If not, the entire slab may be translating.

Also, you don't need high precision, and you should not use vasp's conjugate gradients for the NEB.

[ccccc]

In that calculation the two bottom layers were frozen.
Now i've prepared a new INCAR file

SYSTEM = neb
IMAGES = 8
ISTART = 0
SPRING = -5
ICHARG = 1
ISYM = 0
NELM = 60
NELMIN = 8
PREC = Normal
LREAL = Auto
NSW = 150
IBRION = 3
POTIM = 0.2
NPAR = 1
EDIFFG =-1e-03

It's written in manual that EDIFFG-tag specifes the components of forces on this system when it's set to a negative value. So can i see the components of forces at the end of this run? So, will the transition barrier change if i change the EDIFFG tag?

Thank you...

[graeme]

You should always use the force criteria (negative ediffg). If you have linked with our NEB code, you can see the magnitude of the forces by grepping the outcar for "FORCES:"