[CPMD-list] Question about molecular systems in CPMD

[Christian Tuma]

Dear CPMD community,

I have a general question concerning the calculation of adsorption energies
of molecules in periodic systems, e.g. for the adsorption of a single small
molecule in a zeolite. There is no problem to get the energy of the zeolite
with and without the adsorbed small molecule. The unit cell is triclinic and
the same in both cases, so the plane wave basis sets are the same and one
gets two energy terms so far. The question I have is how to calculate the 
energy of the small molecule as the third energy term in order to get the 
adsorption energy. One could think of two different ways "A" and "B":

A) The basis set must not change in calculations of reaction energies. So
one has to take the same unit cell as in the other two calculations. In this
case the CPMD code cannot handle the small molecule as an "ISOLATED SYSTEM"
since "SYMMETRY" is "14" instead of "0" as required for isolated systems. 
That means a molecular crystal is calculated.

B) A molecular crystal is somehow strange, the treatment as a gasphase 
molecule ("SYMMETRY 0") is preferred. But here one cannot use the triclinic
unit cell in CPMD as in the other two calculations. So an orthorhombic unit 
cell (big enough to include most of the wavefunction) needs to be defined. 
That means a new basis set is introduced compared to the other two 
calculations.


Questions:

1.) Can I use "A" in cases with large unit cells so that no interaction 
between periodic images of the small molecule can be assumed? Does the
total energy of the molecular crystal correspond to the energy of the 
gasphase molecule in this case?

2.) The same questions as 1.), but now for a charged molecule.

3.) Can I use "B" assuming that the shape and the size of the unit cell
do not change the "quality" of the plane wave basis set as long as the
cutoff value is not too small? In other words: Which is the expected order
of magnitude of the change in the total energy when the cell vectors change
(from a triclinic to an orthorhombic system) for an isolated system? 
(How does this depend on the cutoff value?)


I'm glad for any suggestions or explanations to my questions. Thanks!

Christian.
-- 
Christian Tuma             Humboldt-Universitaet Berlin
ct at chemie.hu-berlin.de     Arbeitsgruppe Quantenchemie (Prof. Sauer)
phone: +49-30-20937140     Brook-Taylor-Str. 2, 12489 Berlin, GERMANY
fax:   +49-30-20937136     http://www.chemie.hu-berlin.de/ag_sauer