[gpaw-users] lcao basis set generation

[Michael F. Peintinger]

I generated basis sets for every setup using:
gpaw-basis -f PW91 -t tzp -n PW91 H C N O F Al P S Cl Mo

Then I chose the basis like this:
calculator = GPAW(h=0.2, xc='PW1PW', mode='lcao', basis='PW91.tzp')

This allows me to manage the different basis sets in the setups path.

Just if anyone is interested...

I currently am comparing geometries of a molecular crystal with LD and LCAO 
for these methods:
PBE, revPBE, PBE0, PW1PW

I'll let you know if I find any suprises.

I would like to use TPSS as well, but I do not get it running yet.


The only thing i have not quite figured out is how to optimize the lattice 
constants consistent with the atom positions.

There is no option like FullOpG in Crystal or the Cell relaxation in VASP, 
right?
So I will have to perform atom optimizations for different lattice parameters. 
Or are there gradients available?


Thank you again


Michael


On Thursday, 22. April 2010 15:33:42 you wrote:
> On Thu, 2010-04-22 at 14:46 +0200, Michael F. Peintinger wrote:
> > Dear all,
> > 
> > must a basis set be generated for each functional?
> > 
> > or can I use a basis that was created with PBE also with PW91?
> 
> That should work fine.
> 
> Actually, there is a new tar-file with setups here:
> 
>   https://wiki.fysik.dtu.dk/gpaw/setups/setups.html
> 
> which now contains dzp basis functions for all elements.  So you could
> try to use those with basis='dzp'.  You can also use just the szp part
> with basis='szp(dzp)'.  This is almost completely untested, so be
> careful!
> 
> JJ

-- 
Dipl. Chem. Michael F. Peintinger

Institut für Physikalische
und Theoretische Chemie

Wegelerstr.12
D-53115 Bonn

Telefon: 0228/73-2998
E-Mail: mpei at thch.uni-bonn.de