My Community

I am doing a 1 D relaxation for a kind of polymer,

ISIF=3.
ISTART = 1
INIWAV = 1
LWAVE = .TRUE.
ICHARG = 1
LCHARG = .TRUE.

Due to some reason, the job was stopped. Then I restart the job without any change of INCAR, but paste the CONTCAR to POSCAR. Then I found that the lattice vectore(along Z axis) changed abnormally after a few ionic steps. The result is abviously not right.
Your answer to this problem is highly appreciated.

Tian

is this supposed to be a single polymer chain in vacuum?
along which direction do you align the molecule?
by setting ISIF=3, the box shape and volume are relaxed, this is certainly NOT appropriate for a 1D polymer. If you do not expect large relaxation effects in the direction along which the polymer relaxes, it will be best to choose ISIF=4
To get reliable results, PREC=High must be set in addition, and the cutoffs have to be increased, ENCUT= at least 1.3*max(ENMAX)

Thank you for your answer.
Actually, one of my concerns is just the relaxation along the polymer chain.
I put the polymer chains in a simple tetragonal cell, with the chain along with z-axis (also, the lattice vector c). Also I set the initial value along a, b direction very large, like 5 A.
Am I right to use ISIF=3? Can VASP do single polymer chain in vacuum without put it in a unit cell?
Thanks again.

1) in all codes making use of translational invariance, you HAVE to define a 3D unit cell.
2) of course depending on the lateral extension of the polymer, a 5 x 5 x z unit cell seems too small to guarantee that neighboring polymer chains do not interact. This may be the reason for the large changes in z-direction (provided that your input geometry was reasonable) In such a case, also ISIF=3 obviously is not a good choice.
3) make sure you use a 1 x 1 x Nz k-mesh

5 A is not large. It's definitely not "very large". 10 or 15 A sounds like it might be reasonable, depending on what the polymer is (i.e. how thick it is). You'll need to do some tests, increasing the box size until you the energies no longer change significantly with the box size.

ISIF = 4 or 7 seem appropriate (maybe 5 if you're careful and think about what you're doing).

Of course, if you are trying to simulate an ordered aggregation of these polymer chains, ignore everything that's been said! Then you need to treat the system just like a normal bulk material.
<span class='smallblacktext'>[ Edited Thu Nov 16 2006, 01:26PM ]</span>