Page 1 of 1

convergence problem

Posted: Tue Jul 05, 2011 4:36 pm
by asaf
Hi,
I am trying to do a single point calculation, using the am05 functional with the standard PBE pseudo-potentials of VASP. The energy is diverged.
The content of the INCAR file is:
GGA = AM
PREC = accurate
ISTART = 0
ICHARG = 2
EDIFF = 0.1E-05
IBRION = -1
NSW = 0
LREAL = .FALSE.
ISMEAR = 0
SIGMA = 0.05
NPAR = NPARVALUE
NWRITE = 0


and the energy goes as:
N E dE d eps ncg rms rms(c)
DAV: 1 0.320445905868E+06 0.32045E+06 -0.23887E+07 490 0.356E+05
DAV: 2 0.777338145438E+05 -0.24271E+06 -0.23997E+06 980 0.795E+04
DAV: 3 0.458900884544E+05 -0.31844E+05 -0.31404E+05 980 0.291E+04
DAV: 4 0.350904732635E+05 -0.10800E+05 -0.10654E+05 980 0.156E+04
DAV: 5 0.297551456700E+05 -0.53353E+04 -0.52827E+04 980 0.115E+04 0.374E+02
DAV: 6 0.184732584169E+11 0.18473E+11 -0.58181E+10 887 0.181E+09 0.241E+02
DAV: 7 -0.507873419265E+13 -0.50972E+13 -0.44588E+13 652 0.372E+10 0.707E+02
DAV: 8 0.642437905016E+11 0.51430E+13 -0.20541E+11 490 0.636E+09 0.662E+02
DAV: 9 -0.562690850598E+11 -0.12051E+12 -0.34214E+11 586 0.214E+09 0.635E+02
DAV: 10 0.202773161757E+12 0.25904E+12 -0.94984E+11 959 0.166E+10 0.559E+02
DAV: 11 0.376614273057E+13 0.35634E+13 -0.52844E+11 964 0.221E+10 0.348E+02
DAV: 12 0.861021035792E+13 0.48441E+13 -0.94884E+10 980 0.722E+09 0.187E+02
DAV: 13 0.526080837064E+13 -0.33494E+13 -0.12521E+11 972 0.496E+09 0.129E+02
DAV: 14 0.283037484795E+13 -0.24304E+13 -0.78110E+11 980 0.902E+09 0.829E+01
DAV: 15 0.238990546731E+13 -0.44047E+12 -0.16569E+10 980 0.109E+09 0.762E+01
DAV: 16 0.166136531120E+13 -0.72854E+12 -0.71656E+10 972 0.175E+09 0.637E+01
DAV: 17 0.105343397333E+13 -0.60793E+12 -0.13202E+11 980 0.196E+09 0.496E+01
DAV: 18 0.923192067781E+12 -0.13024E+12 -0.29542E+08 980 0.364E+08 0.462E+01
DAV: 19 0.701967908672E+12 -0.22122E+12 -0.14474E+08 980 0.244E+08 0.405E+01
DAV: 20 0.398128811306E+12 -0.30384E+12 -0.92717E+07 980 0.182E+08 0.310E+01
DAV: 21 0.192059008215E+12 -0.20607E+12 -0.87700E+07 980 0.146E+08 0.224E+01
DAV: 22 0.139228877288E+12 -0.52830E+11 -0.45458E+07 980 0.126E+08 0.199E+01
DAV: 23 0.104132441980E+12 -0.35096E+11 -0.30936E+07 980 0.871E+07 0.177E+01
DAV: 24 0.911491449202E+11 -0.12983E+11 -0.22118E+07 980 0.809E+07 0.168E+01
DAV: 25 0.104988468197E+12 0.13839E+11 -0.13867E+07 980 0.713E+07 0.177E+01
DAV: 26 0.113118500470E+12 0.81300E+10 -0.11740E+07 980 0.578E+07 0.182E+01
DAV: 27 0.114880344783E+12 0.17618E+10 -0.95872E+06 980 0.596E+07 0.183E+01
DAV: 28 0.123761331339E+12 0.88810E+10 -0.95327E+06 980 0.516E+07 0.191E+01
DAV: 29 0.137377339728E+12 0.13616E+11 -0.86141E+06 980 0.585E+07 0.199E+01
DAV: 30 0.141077336066E+12 0.37000E+10 -0.75968E+06 980 0.515E+07 0.201E+01
DAV: 31 0.145080553380E+12 0.40032E+10 -0.62922E+06 980 0.502E+07 0.204E+01
DAV: 32 0.936882682446E+11 -0.51392E+11 -0.36076E+06 980 0.379E+07 0.171E+01


Dose anyone knows what is the problem ?
thanks

convergence problem

Posted: Sat Jul 09, 2011 3:34 am
by forsdan
From when I visited Dr. Mattsson we attributed the divergence to numerical issues in the low density regions (at least for systems with vacuum such as slab), which caused a divergence in the the potential. This was however in 2009, so Dr. Armiento and Dr. Mattsson could have found another reason by now. In any case, there are some procedures you can try to see if they work. I have listed them in a previous post:

http://cms.mpi.univie.ac.at/vasp-forum/ ... php?4.6563

Cheers,
Dan