problem with energy convergence

Message

Leny · #1 Post by **Leny** » Sat Apr 21, 2012 2:27 am

Dear Vasp users,
I am running some calculations and getting very strange result in energy decreasing it looks like that :

0 0.84891300 -394.766210 0
1 0.83329900 -394.774032 0.007822
2 0.54520900 -394.915302 -0.149092
3 0.53173100 8641.82598 9036.59
4 0.29672800 2757.622456 3152.39
5 0.29897400 -395.021465 -0.255255
6 0.22045900 33417.01410 33811.8
7 0.20961300 4493.151421 4887.92
8 0.16533100 165.895852 560.662
9 0.14155300 -395.108357 -0.342147

I made my super cell and it is like that on the axis
X 28.8? Y 2.84? Z 39.80?

and my KPOINTS file is like that,

Automatic Mesh
0
Monkhorst-Pack
2 8 1
0. 0. 0.

Does any one knows what is the reason of this mess, I really appreciate any help.
best regards,

liuhongfei · #2 Post by **liuhongfei** » Tue Apr 24, 2012 12:08 pm

That is a very strange problem, you should give more information about your calculation. e.g. your POSCAR and INCAR, etc.

Leny · #3 Post by **Leny** » Thu May 03, 2012 1:31 pm

Hi,
I am calculating adsorption energy of 1H atom on Pt (fcc110) and the surface is big with 11 microfacets, here is my INCAR:

# Regular
SYSTEM = Pt_1H
ISYM = 0
NSW = 200
EDIFF = 0.1E-4
EDIFFG = -0.01
ENCUT = 450
GGA = RP
LWAVE = .FALSE.
LCHARG = .FALSE.
PREC = medium
ISIF = 0
IDIPOL = 3
# ISPIN = 2

# Speedup
NPAR = 8
NSIM = 4
IALGO = 48
LPLANE = .TRUE.
LREAL = Auto

# Optiizers
IBRION = 3
POTIM = 0.0
IOPT = 2

This started happening after I made my super cell bigger to 114 Pt.
Regards,

alex · #4 Post by **alex** » Thu May 03, 2012 3:22 pm

So, how was the structure before?

It's pobably a hexagonal cell, so use Gamma-centred KPOINTS. And change the 2 8 1 to a 1 8 1 set or so. You don't need two k-points for a=28\AA ...

Cheers,

alex

Leny · #5 Post by **Leny** » Tue May 08, 2012 12:50 am

Hi,
My structure was the same except smaller in size, it is cubic cell and I am adding 1 microfacet each time and when I reached this level to have 6 microfacets then this problem started with.
I changed the KPOINTS to 1 8 1 but it did not solve this problem !

Cheers

alex · #6 Post by **alex** » Tue May 08, 2012 9:04 am

Hi,

did you also change Monkhorst to Gamma?

Cheers,

alex

Leny · #7 Post by **Leny** » Tue May 08, 2012 12:28 pm

Hi,
No, I did not change Monkhorst to Gamma, here is my KPOINTS file

Automatic Mesh
0
Monkhorst-Pack
1 8 1
0. 0. 0.

I have always used Monkhorst in fact, it is not been a long time since I started to use Vasp and the field in general .
Now a calculation is running and as you told me I left IDIPOL out and until now not any fluctuation has happened with energy if it continues like that until the end then the problem is solved (hopefully).
I will try Gamma centered KPOINTS too to see what happens .

cheers,

alex · #8 Post by **alex** » Tue May 08, 2012 1:34 pm

That would depend on having a hexagonal unit cell (Gamma-centred) or not (Monkhorst).
So: do you?

Leny · #9 Post by **Leny** » Tue May 08, 2012 11:42 pm

No I did not have a hexagonal unit cell, that is good to know, Thank you.

cheers,