neb results analysis

[lanjh]

hello, Sir!
I have performed a neb run now. The Incar file is as follows
SYSTEM=NH2-H
ENCUT=420
ENAUG=700
ISTART=1;ICHARG=2
ISMEAR=2; SIGMA=0.2
NSW=600
POTIM=0.05
IBRION=3
ALGO=V
GGA=91
PREC=medium
LWAVE=.FALSE.
LCHARG=.FALSE.

ICHAIN = 0
IMAGES = 4
SPRING= -5.0
LCLIMB = .TRUE.
LTANGENT= .TRUE.
.

Now I checked the results, and found the neb.dat file gives the following results :
0 0.000000 0.000000 0.000000 0
1 1.655292 0.794083 -2.991783 1
2 3.314092 1.983367 -0.269760 2
3 4.973704 2.863259 -0.004253 3
4 5.980563 2.106869 0.630218 4
5 6.986842 1.841780 0.000000 5

And the nebef.dat gives
0 0.00976500 -437.78920000 0.00000000
1 0.03649400 -436.99511700 0.79408300
2 0.03529800 -435.80583300 1.98336700
3 0.03076600 -434.92594100 2.86325900
4 0.02990300 -435.68233100 2.10686900
5 0.00996100 -435.94742000 1.84178000


Can you tell me whether my results is good ? Thank you very much.

--------------------------------------------------------------------------------

[graeme]

That looks just fine. It would be good to set a convergence criteria, such as
EDIFFG=-0.01

If you do this and run a little longer, you can make sure that your barrier of 2.86 eV does not change.

Also, please don't cross-post to the vasp forum. This is a good place for transition state calculations using our code, and all other questions can be address on the vasp forum.

[lanjh]

Thank you very much. In fact, the covergance of force and energy didn't change after 300 steps. I want to do find the transition, whether I can use the high point in this curve, or I need to run a dimer calculation ?

[graeme]

Since you are using the climbing-image, the highest energy image along the band will converge to the saddle point.