[xcrysden] charge density and electrostatic potential

[Gabriele Sclauzero]

Dear Ariadna,

    I just want to add a few comments on top of what Tone said, hoping 
they will be useful to solve your problem.

On 09/07/2010 03:03 PM, Tone Kokalj wrote:
> On Fri, 2010-09-03 at 16:36 +0000, Ariadna Blanca Romero wrote:
>    
>> Hi all
>>
>> I have a problem to understand what a negative isovalue means when I
>> want to visualize the charge density and the electrostatic potential.
>> I am using crystal06 to get these.
>>      
> As for the charge density, the negative value typically appears only for
> charge density difference plots, and it indicates charge deficit
> regions.
>    

I don't know much about Crystal code, but I believe this is the same for 
every electronic structure code employing some broadening technique: if 
you use smearing functions like Methfessel and Paxton's there may be 
some regions with (hopefully very small) negative charge densities. If 
you don't like them, you'll need to switch to another smearing function 
which is not affected by this drawback.

>> "when displaying the isosurface of a molecular orbital, then it is a
>> good idea to display both positive and negative valued isosurface (for
>> example the -0.02/0.02 sqrt(BOHR^3/electrons)). By default the
>> positive isosurface is rendered in red, while negative in blue."
>>
>> The problem comes when I wanto to interpret it for electronic density,
>> which is given  in electrons/BHOR^3 by crystal06 and this quantity
>> should be POSITIVE.
>> First, the isovalue is the square root of the inverse of the
>> electronic density (at least by the units it seems), as is shown in
>> the xcrysden site, so what a negative isovalue means????? Does it mean
>> less quantity of electrons????.
>>      
> Don't confuse molecular orbitals and charge density!
>
> Molecular orbitals (say for molecules) can be chosen as being real; not
> in general true for crystals. For the latter reason, I believe (not
> sure) you can only plot the square moduli of molecular/crystal orbitals
> in CRYSTAL06 (i.e. \Phi^2, associated charge density), and in this case
> you only plot the "positive" isovalue. However if you would plot the
> real molecular orbital than the "+/-" isovalue button is useful.
>
>    

In crystals the wavefunctions at the Gamma point in the BZ can be chosen 
to be real valued, so that you can think about plotting something like:

sign(\psi)*|psi|^2

where \psi is the K-S eigenstate at Gamma. In Quantum-ESPRESSO this is 
done by setting a flag in the postprocessing step, maybe in Crystal this 
is done automatically?

>
> Regards, Tone
>
>    

Hope this helps,


GS

-- 

Gabriele Sclauzero, EPFL SB ITP CSEA
PH H2 462, Station 3, CH-1015 Lausanne