using empirical VDW correction with CP2K

[Axel]

hi everybody,

i've just uploaded an input example for adding empirical
dispersion correction to DFT by use of the input files alone.

due to use of the new input preprocessor facility it should be
comparatively easy to adapt this to a new system and to keep
the multiple force environment input parts consistent.

the tcl file demonstrates on top of that how one can
set up a potentical scan driven by VMD.

http://groups.google.com/group/cp2k/web/emp-vdw-scan-example.tar.gz

hope this is useful. comments are appreciated.
please post questions related to this upload here.

cheers,
axel.

[lmliu]

Dear Axel,

It is very useful.



Thanks a lot,

Best Regards,
Limin

2008/3/7, Axel <akoh...@gmail.com>:

[lmliu]

Dear Axel,

I compare your CH4-CH4 parameter with the Chemical Physics 327 (2006) 54–62 and there is greatly difference. Do this part use the other unit  or you take the parameter from other reference?

When I test water-water system, it shows the following errors. what does this mean?

*** 01:34:59 WARNING in force_env_methods:mixed_energy_forces err=-300   ***
 *** ASSERTION (cond) failed at line   1151 Error (0.129985E-11) in       ***
 *** computing numerical derivatives larger then(0.100000E-11) .          ***

Thanks,
Limin

[Axel]

On Mar 9, 8:36 pm, lmliu <lml...@gmail.com> wrote:
> Dear Axel,
>

> I compare your CH4-CH4 parameter with the Chemical Physics 327 (2006) 54-62

> and there is greatly difference. Do this part use the other unit or you
> take the parameter from other reference?

the parameters are from the input in the CPMD-test suite, not
from the chem phys paper. units are the same as in CPMD
(= atomic units).

>
> When I test water-water system, it shows the following errors. what does
> this mean?

what is the input you are using?

axel.

[Teodoro Laino]

Dear Limin,

Please read carefully the error/warning  messages:

On 10 Mar 2008, at 01:36, lmliu wrote:

*** 01:34:59 WARNING in force_env_methods:mixed_energy_forces err=-300   ***
 *** ASSERTION (cond) failed at line   1151 Error (0.129985E-11) in       ***
 *** computing numerical derivatives larger then(0.100000E-11) .          ***

First: this is not an error message but it is a warning (WARNING in [..]). 

Second: It clearly says that: Error (0.129985E-11) in computing numerical derivatives larger then(0.100000E-11) .

I guess you can cope with this error on the evaluation of the MIXED forces.. don't you (look at the magnitude of the numbers)?

If you are annoyed from seeing this warning message you can tune the precision of the evaluation of the numerical derivative

with the keyword:

http://cp2k.berlios.de/input/InputReference~__ROOT__~FORCE_EVAL~MIXED~GENERIC.html#ERROR_LIMIT

10^-12 is an extremely safe parameter. I guess 10^-7 will do your work in a very good way as well (please compare these numbers

with the precision achieved in computing the DFT forces).

Cheers,

Teo

[lmliu]

thanks axel and Teo,

I use the H-H, H-O and O-O parameters for PBE,  taking from the Chem. Phys. 327 (2006) 54.

          ATOMS        H O
          RCUT [angstrom]        10
          RMIN [angstrom]         0.2
          RMAX [angstrom]        11.0
          VARIABLES    R
          PARAMETERS   CSIX      RZERO    D
          VALUES      -2.675     3.947    3
          FUNCTION    (CSIX/(R**6))*((1.0 - EXP(-1.0*D*((R/RZERO)**7)))**4)

          ATOMS        H H
          PARAMETERS   CSIX      RZERO    D
          VALUES      -2.290     3.538    3

          ATOMS        O O
          PARAMETERS   CSIX      RZERO    D
          VALUES       -9.631    4.233    3

Best Regards,
Limin

2008/3/10, Teodoro Laino <teodor...@gmail.com>: