"WARNING: stress and forces are not correct"

[opvu]

Dear all,
To do band structure and DOS calculation, I first relax the cubic unit cell, then generate charge density in a static running (CONCAR -> POSCAR, NSW=0, IBRION=-1). Up to this, there is no problem. But After that, I calculate band structure and DOS (ICHARGE=11).
This is my INCAR:

SYSTEM = 1x1x1SrTiO3
ISTART = 0
NWRITE = 2
INIWAV = 1
ISPIN = 1 !Spin Polarization
ICHARG = 11 !
LREAL = .FALSE. !
LWAVE = .TRUE.
LCHARG = .TRUE. !
PREC = Medium
EDIFF = 1E-4
ISMEAR = 0
SIGMA = 0.05
ALGO = Fast !
NSW = 0 !
NEDOS = 1200 !
LORBIT = 11 !

In vasp.out, there is the following warning:
"WARNING: stress and forces are not correct"

And in the OUTCAR
"WARNING: stress and forces are not correct
(second dervivative of E(xc) not defined)"

Can anybody tell me how to make this warning disappear?

Thanks very much

opvu

<span class='smallblacktext'>[ Edited ]</span>

[admin]

This is an automatic warning which is written if ICHARG=11 or 12, because -strictly speaking- the read CHGCAR (which is kept constant) is the fully scf solution ony for the set of k-points with which is was generated (and usually, the k-grid sampling is changed to do a band structure or DOS run). However, with such a run you are not supposed to converge stress and forces, because they should already be at that stage of a claculation.

[forwordom]

i have meet the same problem !
Could you tell me if this warning will affect results

[admin]

you should not use the forces and stresses of an ICHARG=11 run at all if you have changed the k-point set. Of course the results will change for that ionic step, depending on how well-converged your k-point set of the scf calculations has been..

[admin]

yes, of course, this is exactly what the warning says. you should use ICHARG=11 only for DOS and bandstructure calculations (post-processing of an otherwise converged calculation)