[CPMD-list] I get (NAN) in trajec.xyz

Axel Kohlmeyer axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Mon Feb 21 23:19:36 CET 2005


On Sat, 19 Feb 2005, Younes Ansari wrote:

YA> 
YA> Dear sir :
YA> 
YA> I)     I have done a MOLECULAR DYNAMICS CP for SiO2-Trymidite (50  atoms) with MAXSTEP 100 but ofter about 31 steps I got NAN but not the numbers in the TRAJEC.xyz for example :
YA> 
YA> 1
YA> 
YA> 1.0            1.3000    4.700
YA> 
YA> 1.01    1.78        5.1300
YA> 
YA> 
YA> 33
YA> 
YA> NAN    NAN    NAN
YA> 

it would be interesting to see the energies from the step 
_before_ the NaNs, but it will be most likely, that your
electrons are getting 'too hot' and the whole CP dynamics
is breaking down. 

YA> So I wont see the complete trajectory in the VMD so please would you
YA> tell me which keyword could help me have a complete step of
YA> caculations.

you can always delete the last steps in the .xyz file up to the point 
where the NaNs start and read that into VMD. but this is not
your problem right now. you either have:
- a not too well converged wavefunction to start from
- a non-equilibrium geometry or atoms that are too close, 
  so they become 'too fast' for the electrons
so you should check, your cell size and geometry (e.g. by
doing a wavefunction optimization with RHOOUT and then
run cpmd2cube on the resulting DENSITY file and view this
with the latest VMD version with periodic display of the
density isosurface and the atoms.
you can also try a smaller timestep.

YA> II) Is there any keywords which separate a slab (for example SiO2
YA> slab) from other cases in the input file.

your question is not very clear. you alway calculate the 
electron structure for the whole cell, i.e. you cannot
have a 'high-quality' and a 'low-quality' region, and
you also don't have an option to 'glue' two inputs together.
you always have to merge the inputs into one.

if you are referring to decoupling the electrostatic
interactions between the periodic images of the slabs
in z-direction, then you should check out the SURFACE
option and have a look at the notes section of the manual.

axel.
-- 

=======================================================================
Dr. Axel Kohlmeyer   e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-26673
Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 (0)234/32-14045
D-44780 Bochum  http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.




More information about the CPMD-list mailing list