[CPMD-list] highest levels of theory

[Axel Kohlmeyer]

On Fri, 26 Nov 2004, Robert Williams wrote:

RW> Dear list subscribers,

hi bob!

RW> What might be regarded as the highest
RW> levels of theory available in cpmd,
RW> double numerical with polarization or better,
RW> particularly that include functionals
RW> appropriate for estimating dispersion forces -
RW> such as pbe?

hmm, i am not sure, if i have fully understood 
what you mean by 'level of theory'. but if you
see this in a similar way of programs like gaussian,
then you have to realize, that the basis set in cpmd
are plane waves which are in so far different from 
local basis sets, that they have the same 'quality'
everywhere and that your basis set size solely
determined by the plane wave energy cutoff 
(CUTOFF keyword). 

as far as the 'method' is concerned, there are
different LDA and GGA functionals available
(see FUNCTIONALS keyword or the sourcecode).
all of them share the general problem of DFT,
that their representation of dispersion interactions
is not very good. there is a feature implemented,
which can correct for it, by adding an empirical VDW
term and, if i understood correctly, the inclusion
of hartree-fock exchange should be an alternative,
though computational very expensive solution.

so far i have used neither of them and thus cannot
comment on how useful they are for your system and 
under which conditions they are at all usable in the 
current cpmd release. perhaps somebody else (juerg?)
can comment on it.

RW> 
RW> My initial calculations on a unit cell
RW> of N-methylacetamide using the default
RW> lda correlation (PZ) with pbe (from
RW> contrib/pseudo/[H,C,N]_MT_PBE.psp)

hmmm. please correct me, if i am mistaking you,
but it seems to me, that you have confused something here.
you have to specify the desired functional in
the &DFT section and then select a _matching_
set of pseudopotential files, i.e. if you want
to use PBE or LDA, then you have to use pseudopotentials,
that were created from all-electron calculations with PBE 
or LDA, respectively.


RW> yielded very accurate frequencies for
RW> internal modes, with fairly accurate
RW> displacements (INS intensities),
RW> but fairly poor frequencies for interactions.
RW> My calculations using dnp/gga/pbe with DMol3
RW> have done better on all accounts except for INS intensities.
RW> 
RW> Could someone recommend a good
RW> combination for correlation and
RW> gradient correction?
RW> GRADIENT CORRECTION [GGAX?,PBEX?, REVPBEX?] [GGAC?,PBEC?,REVPBEC?]?

it is generally recommended to use the keyword FUNCTIONAL instead of
explicitely selecting the exchange and correlation 
and their gradient corrections unless you know what you are doing.

regards,
	axel kohlmeyer.


-- 

=======================================================================
Dr. Axel Kohlmeyer   e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-26673
Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 (0)234/32-14045
D-44780 Bochum  http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.