The ABINIT Users Mailing List ( CLOSED )

[Nicola Marzari <marzari@mit.edu>]

Shultz, Jack wrote:
> Nicola,
>
> I looked at some of the research described on your website. Is the
> Hubbard approach to augment DFT the same thing as DFT+U? I'm also trying
> to find out what is meant by a nodal surface is that how you describe
> the environment of an enzyme?
>
> Jack


Hi Jack,

let me reply to the mailing list - apologies if others are not 
interested. Hubbard / DFT+U are the same thing, in this context.

For a review, I personally like the Cococcioni/de Gironcoli PRB of 2005,
for the insight on U being a linear-response property - but
you should really look at the early papers by Anisimov, Andersen,
Lichtenstein that introduced the method.

The nodal surface is the locus of points where the many body
wavefunction goes to zero - QMC often uses the one from DFT,
but if the one from DFT is not good (e.g. the wrong symmetry)
than there is little that QMC can do to improve such bad start.
This is much more of a problem in transition-metal complexes, where
the ordering of states that have significant d (or f) orbital
components can be incorrect in LDA/GGA/B3LYP, and then QMC would
struggle to get a good energy starting from the locus of zero
amplitude for the many body wavefunction constructed from a set of
single particles orbitals that are not really the good ones.

As you can see, it can be very deeply technical - just wanted to
highlight that it's not a topic where DFT (or even QMC) gives us a
turnkey solution.

                        nicola



---------------------------------------------------------------------
Prof Nicola Marzari   Department of Materials Science and Engineering
13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
tel 617.4522758 fax 2586534 marzari@mit.edu http://quasiamore.mit.edu