[CPMD-list] MD calculation on yttrium complex

[Axel Kohlmeyer]

On 6/15/06, Yi LUO <yi_luo at riken.jp> wrote:
>
>
> Hi, Dear all,

dear luo,

before commenting on your calculation i would generally recommend
to first go over the available CPMD tutorial material and pick a less
challenging system to practice with. CPMD is not a 'black box'
kind of program and probably will never become one.

tutorials are at:
http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/cpmd-tutor/
http://www.pci.unizh.ch/gruppe.hutter/e/information.html

i also recommend reading the suggested reviews on the method and
especially the new book by jorge kohanoff:
http://www.amazon.co.uk/exec/obidos/ASIN/0521815916/
http://www.amazon.com/gp/product/0521815916/

>   I performed CPMD calcualtion on Yttrium organometallic complex and found
> that the structure goes wrong after step 10. Is the pseudopotential or input
> parameters improper?

most likely the input parameters. how did you come up with them?
i will add comments to the appropriate places.

> &INFO
>   This is a test job for the coordination-complex CpYCH2SiH3_ethyl.
> $END
>
> &CPMD
>   MOLECULAR DYNAMICS
>   restart accumulators wavefunction coordinates velocities cell

this is a really you are not restarting from a converged wavefunction.
this is _essential_ unless you are doing a simulated annealing to
converge the wavefunction.

>   TIMESTEP
>    5.0

>   EMASS
>    400.0

using a time step of 5.0 with this fictitious mass
is probably too large. see the bochum tutorial
for a demonstration. note that, atomic masses,
fictitious mass, plane wave cutoff, pseudopotential
and time step all have to be balanced so that you
get meaningful results.

[...]

> &VDW
>   VDW CORRECTION
>    1
>    C6  1 1 -216.9850 6.57936 2.9688
> &END

from your description it looks as if you want to be
running an isolated molecule, why do you activate
the VDW correction feature?

> &ATOMS
> Carbon
> *C_MT_BLYP KLEINMAN-BYLANDER RAGGIO=1.0

ouch. you have selected the PBE functional above and
are using a pseudopotential created with BLYP. this is
inconsistent and will basically render your results useless.

> Yttrium
> *Y_MT_HCTH_SEMI.psp RAGGIO=1.0
>    LMAX=P LOC=P

this entry actually has _four_ problems.
1) it is using a pseudopotential for yet another functional
and one that significantly differs in design from PBE and BLYP.
2) you are excluding the d-channel (which is significant for this element),
by using LMAX=P. most likely LOC=P will not work. you should
try LOC=D on an isolated yttrium atom and check the eigenvalue spectrum
whethere there is a ghost state and if yes (which is likely) switch to
LOC=S
3) your plane wave cutoff from above is not suitable for this pseudopotential,
you will need at least 90ry, probably even something in the area of 100 or 120.
4) you are not using KLEINMAN-BYLANDER which will have CPMD use
GAUSS-HERMITE, but for this element, you need to crank up the number or
integration points (if that is what you want).

note, that if you are using an unknown pseudopotential with no recommendations
or documentation (e.g. papers using it), you _have_ to test it for the required
plane wave cutoff, ghost states and the best local potential etc. etc.
since this is all a complex problem and since you seem to have no experience
in running CPMD i highly recommend to get some practice with a simpler system
and then check out the parameters that you _have_ to adapt to your system
one by one.

> In addition, does anybody know the source of ref. 73 mentioned in the online

yes. please see the mailing list archives. ;-)

> manual? (related to VDW CORRECTION parameters). Also, I only found one kind

yes. ;-)

> of pseudopotential (Y_MT_HCTH_SEMI.psp) or Yttrium from CPMD webpage. Does
> anybody succceded in the MD calculation of yttrium organometallic complex?
>
> I would be very appreciated for your any comments on this matter.

pleople have been running calculations with yttrium. you will find a list
of first know references for many pseudopotentials in mauro boero's
pseudopotential library hosted on juerg hutter's homepage in zurich.

regards,
      axel.


>
> Regards
> Luo
>
>
> _______________________________________________
> CPMD-list mailing list
> CPMD-list at cpmd.org
> http://cpmd.org/mailman/listinfo/cpmd-list
>
>
>


-- 
=======================================================================
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.