QMM: CELL and PERIODIC group

254 views
Skip to first unread message

Luca

unread,
Sep 17, 2007, 9:52:42 AM9/17/07
to cp2k
Hi all,
In QMM section there are &CELL and &PERIODIC group.
Which is the relationship with the same parameters in SUBSYS and in MM
group ?
Do they have same function or they are used for internal scope (grid
for GEEP etc)?
How can I set these parameters?

Thanks

Luca

Teodoro Laino

unread,
Sep 17, 2007, 9:59:01 AM9/17/07
to cp...@googlegroups.com
Hi Luca,

&CELL contains information about the QM cell.. you've to set up it
according the size of your system and
whether you want to do isolated system (with a proper decoupling
technique) or periodic ones..

&PERIODIC in QMMM controls the FULL PERIODIC treatment of the QMMM
electrostatic.. depends what you want
to do.. most of the time you just do QMMM decoupling the QM from all
the other periodic images, keeping only the
zero-cell electrostatic between QM and MM.. Depends on your problem..


Teo

Luca

unread,
Sep 17, 2007, 5:07:43 PM9/17/07
to cp2k
Hi Teo,
Ok, CELL and PERIODIC in QMMM section have same behavior and same
meant
of CELL in SUBSYS and PERIODIC in MM section.
In an isolated system I understand different set of cell parameter.
I do not understand when I want to simulate (NVE or NPT) MD for one QM-
solute (ab-initio organic molecule) in MM-water in a periodic box. In
this case I should have same cell parameters and same rules for PBC in
both system QM and MM....

Thank you
Luca

Teodoro Laino

unread,
Sep 17, 2007, 5:20:44 PM9/17/07
to cp...@googlegroups.com
Sorry Luca I guess I didn't fully understand what you meant..

Anyway, let me just point few things:
NPT does not work with QMMM.. the QMMM stress tensor is not implemented.. anyway this is not a serious limitation..
Better equilibrate the system at MM level and than do an NVT or NVE..

If you want to go for an NVE (or NVT) QM and MM do not necessarily need to have the same setup for PBC..
You can have everything fully periodic (http://dx.doi.org/10.1021/ct6001169) or just QMMM minimum image interaction..

Within a QMMM calculation: the box in the QM section does not affect the PBC (if a decoupling is used) but only the use 
of the plane wave for the Hartree and XC terms (Have a look at the QS algorithm).. 
The CELL section in the MM affects electrostatic and PBC of the global system. The QMMM interaction is controlled through
the QMMM section..
 
Hope this helps,
Teo
Reply all
Reply to author
Forward
0 new messages