Page 1 of 1
NEB calculation input
Posted: Wed Dec 27, 2006 10:55 pm
by molesimu
hello,
I am a new user for NEB method.
Could anyone please tell me whether the endpoint configurations should be put in different directories?
Is there any typical INCAR file for NEB available?
Many thanks!
Steve
Posted: Thu Dec 28, 2006 1:45 am
by andri
Yes, they should be in different directories. The first one should be placed in directory 00 and the last one in XX where XX is the number of images.
The simplest thing to do is to use the "nebmake.pl" script from the vtstscripts package (
http://theory.cm.utexas.edu/vtsttools/scripts/). Just place your endpoints in POSCAR_1 and POSCAR_2 and run the script as
<nebmake.pl POSCAR_1 POSCAR_2 #numberofimages>
This will place POSCAR_1 in 00/POSCAR and POSCAR_2 in XX/POSCAR. The rest of the images are a equidistant linear interpolation between the endpoints.
INCAR (
http://theory.cm.utexas.edu/vtsttools/neb):
-------
# General
SYSTEM = Ru
ISYM = 0
NSW = 2500
IALGO = 38
EDIFF = 0.1E-4
EDIFFG = -0.005
LWAVE = .TRUE.
LCHARG = .FALSE.
PREC = med
ISIF = 0
# NEB
IBRION = 3
POTIM = 0.10
IMAGES = 15
NPAR = 60
LCLIMB=.TRUE.
SPRING = -5.0
problem for running
Posted: Sat Dec 30, 2006 9:14 pm
by molesimu
Thank andri for the reply,
When I submit a job, the job ended shortly since no POSCAR founded.
In the directory of /vaspneb, there are INCAR KPOINTS and POTCAR and sub-directory 00 01 ..., which contain POSCARS. The directory of /vaspneb was the executable directory.
Could anyone please give me some advices?
Posted: Sat Dec 30, 2006 9:38 pm
by andri
Are the ICHAIN and IMAGES tags correctly set in the INCAR? The default for
ICHAIN is 0 which specifies NEB so technically it doesn't need to be set but it doesn't hurt either. IMAGES should be set to the number of movable images, not include the fixed endpoints.