[CPMD-list] TDDFT question

[Philip Shemella]

Hi,

I think what Dong Xu is referring to is a hybrid functional where there 
is a fitting of coefficients for terms from Hartree-Fock exchange & DFT 
exchange-correlation terms.  The percentage of each term is found by 
fitting a set of molecular data (ionization potentials, etc).  The most 
popular example is B3LYP but actually you can use any mixture of XC you 
want.

 From what I've understood most of the hybrid functionals are used so 
they can include exact HF exchange.  With CPMD, HF exchange is not 
usually calculated, unless you have too much computer time or a really 
good reason to do so!

If you don't want to go into the code -- one way to create a hybrid 
functional manually would be to calculate the exchange-correlation 
energy with separate jobs and then find the plug in the coefficients. 
That is kind of a home-made hybrid functional.  Or... Perhaps the code 
prints the Becke88, HF, and Slater exchange as well as the correlation 
terms.  IF that is the case then you could just change the  exchange, 
the HF exchange, and whatever other terms in the mixture that you like. 
  If you look at the TOTAL ENERGY section after wavefunction convergence 
you'll see the contributions from each term. I would test it for my own 
sanity but I don't find any pseudopotentials for B3LYP.

Take care, Phil



> On Fri, 8 Dec 2006, Dong Xu wrote:
> 
> DX> Hi All,
> DX> 
> DX> I'm new to CPMD and I have 2 questions about DFT and TDDFT.
> DX> 
> DX> 1. For DFT calculations, can I customize the percent contributions
> DX> from exchange and correlation functionals in a way similar to
> DX> Gaussian's IOPs ?
> 
> in principle, you can select the functionals for exchange 
> and correlation, both for LDA and gradient correction individually.
> several more popular combinations have been implemented directly
> in a more efficient form (NEWCODE vs. OLDCODE). please
> see the manual and the file dftin.F for details.
>  
> it is not quite clear to me, to what kind of percentages you 
> are referring to. but to go beyond selecting the individual
> functionals, you might have to program it (should be rather
> straightforward in this case, though).
> 
> axel.
> 
> DX> 2. Does TDDFT support analytical gradients for geometry optimization?
> DX> 
> DX> Thanks,
> DX> 
> DX> -DX