[CPMD-list] Energy components in PIMD

[Axel Kohlmeyer]

On Sun, 18 May 2008, ukasz Walewski wrote:

LW> Dnia 2008-05-14, ÿÿro o godzinie 12:35 -0400, Axel Kohlmeyer pisze:
LW> > 
LW> > dear lukasz,
LW> 
LW> Dear Axel,

dear lukasz,

LW> > one thing that came to my mind, is that you may improve the
LW> > situation by turning the "outside" hydrogens into deuterium.
LW> > their bonded vibrations seem to be of pretty high frequency and 
LW> > thus slowing them down by increasing their mass may help improving 
LW> > the stability w.r.t. timestep and ficitious mass.
LW> 
LW> I performed this test - the situation gets even worse than with "light"
LW> outer hydrogens. In the attached graph black plot is the EKINC from
LW> previous run with VDB/35Ry/3au. and the red plot is EKINC from similar
LW> run with the outer protons substituted with deuterium (see attached
LW> input file).

interesting. i didn't expect that. but that could be a hint as 
to what makes it so difficult to reduce the drift in the fictitious
kinetic energy. with the additional mass of the deuterium, i would
expect the whole structure to "bend" and "twist" more. perhaps, but
this is pure speculation, these changes in the geometry induce fairly
large changes in the electronic structure which the fictitious 
electron dynamics cannot follow well. that would explain why lowering 
the fictitious electron mass reduces the drift. 

LW> > you can rescale more often, say every 25th step? that would make
LW> > the disruptions smaller (but more frequent).
LW> 
LW> This seems to me negatively verified - on the second attached graph the
LW> black plot is the total energy with COM motion removal every 10'000
LW> steps and the red plot - every 25 steps. The plots speak for themselves.

yep. so either you have to live with what you have using the 
translation/rotation correction infrequently, remove the subtract
COMVEL/ROTVEL altogether after equilibration or use 3 or 4 _very_
soft position _restraints_ on some of the outer carbon atoms that
will also make sure that the system stays where it should.

cheers,
   axel.

LW> 
LW> Regards,
LW> Lukasz
LW> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.