[DFTB-Plus-User] R: Compute electrostatic potential with NEGF

Alessandro Pecchia alessandro.pecchia at ismn.cnr.it
Mon Oct 31 10:51:56 CET 2016


Hello Andrea,

 

DFTB never calculates the full charge density due to electrons+nuclei, but only deviations from reference atomic charge densities (neutral atoms). Furthermore such deviations are expanded in spherical harmonics and in the traditional dftb implementation only the s-waves are considered with a simple Mulliken analysis. 

In practice what is computed in dftb is not suitable for your calculation because the potential seen for a neutral CNT will be 0! This does not mean there are no electrons or nuclei around. Just all their interactions are included in the SK parameters.

If you are good in programming you could compute the charge density from the wavefunctions, extract our mudpack poisson solver and create a stand-alone interface for it, read-in the charge density using an exceptionally fine grid and finally solve the potential… 

if you do this, let me know that we will include it in the dftb+ tools !

 

Maybe it is easier for you to carry calculations using DFT packages, but bare in mind that most use pseudopotentials. 

If you really need the ions and the core electrons than you can use CP2K, GPAW, Gaussian…

 

 

Hope this helps,

 

Alessandro. 

 

 

Da: DFTB-Plus-User [mailto:dftb-plus-user-bounces at mailman.zfn.uni-bremen.de] Per conto di Andrea Pedrielli
Inviato: domenica 30 ottobre 2016 18:32
A: User list for DFTB+ related questions
Oggetto: Re: [DFTB-Plus-User] Compute electrostatic potential with NEGF

 

Thanks again,

essentially the aim of the work is to simulate the electron microscopy holography response that is proportional to the integrated coulomb potential along the beam direction. So, one electron goes through the sample (in this case the nanotube) and feels a phase shift due to the coulomb potential in each point along the path. The potential has to be computed in each point of the space, also inside the atomic orbitals. In the best case I would compute the potential from the charge density + nuclei, like in DFT. I understand that the orbitals in DFTB are not changed in shape. However, supposing a not so large change in the shape of the orbitals it will be useful also only compute the potential starting from the charge density of the non deformed orbitals. I tried to do this plotting the the charge density using waveplot to a cube file and using an Ewald summation over all the points. The problem is that the computational cost is very high because I need a fine grid to describe the charge density inside the atoms. If I understand well the Poisson solver doesn't take as input the charge density that one can plot with waveplot but only the difference of charge with respect to the reference configuration? 

 

Andrea 

Il domenica 30 ottobre 2016, Gabriele Penazzi <penazzi at uni-bremen.de> ha scritto:

Hi Andrea,

I am not sure what you want to do, but dftb as a method may not include the details you need. 

I see now the point about mulliken charges from your previous mail, but the use of mulliken charges is pretty much a direct consequence of the way dftb (at least in its current and most popular version) is formulated. Nuclei+electron density does not make too much sense, as you work in a monopole approximation based on differences with respect to a reference configuration. The s-like projection of the potential reflects this and is consistent with the usual scc-dftb based on gamma-functional.

For example, you have no potential in a system without self-consistent fluctuation. And even if you compare two system with different hybridization, if the scc component is zero the real space potential will be the same, i.e. zero. So my guess is no, you can not really have detailed information about he charge density between atoms and you probably have to go at DFT level. 

Best,

Gabriele

 

On 10/29/2016 11:55 PM, Andrea Pedrielli wrote:

 

Thank you Gabriele,  

If I could, I would ask you another thing. I read in the online tutorial on silicon nanowire that the charge density is expanded in s spherical orbitals weighted with Mullen charges. Part of my interest in compute charge density and potential is take into account the charge density between the atoms, due to the bonds. If the charge density is expanded in s-like orbitals the charge density between the atoms (for example in a nanotube) is still well described? 

 

Andrea


Il sabato 29 ottobre 2016, Gabriele Penazzi <penazzi at uni-bremen.de <javascript:_e(%7B%7D,'cvml','penazzi at uni-bremen.de');> > ha scritto:

Hi Andrea,

yes, you can do a non transport calculation with the real space Poisson solver, I did the same for the same reason. Whether you can simulate your system or not, however, may depend on the boundary conditions you want to impose. For example, all periodic does not work if I remember right, but for a 1D system you should be able to set up a well defined calculation. 

Gabriele

 

On 10/28/2016 10:51 PM, Andrea Pedrielli wrote:

Hi users,  

I need to compute the electrostatic potential produced by the charge density from a dftb+ calculation. I know that there is no possibility using dftb+, but I have seen that in the NEGF package there is a Poisson solver. In particular I have to compute the inner and the outer potential of a carbon nanotube, so a 1D object. Can I use a NEGF package for a non-transport calculation with the aim of extract the electrostatic potential? I underline that I need the potential due to the total charge density+nuclei and Mulliken charges are not suitable for my purpose.

 

Thank you in advance,

 

Andrea

 

_______________________________________________
DFTB-Plus-User mailing list
DFTB-Plus-User at mailman.zfn.uni-bremen.de
https://mailman.zfn.uni-bremen.de/cgi-bin/mailman/listinfo/dftb-plus-user

-- 
Dr. Gabriele Penazzi
BCCMS - University of Bremen
 
http://www.bccms.uni-bremen.de/
http://sites.google.com/site/gabrielepenazzi/
 
phone: +49 (0) 421 218 9328
fax: +49 (0) 421 218 4764
_______________________________________________
DFTB-Plus-User mailing list
DFTB-Plus-User at mailman.zfn.uni-bremen.de <javascript:_e(%7B%7D,'cvml','DFTB-Plus-User at mailman.zfn.uni-bremen.de');> 
https://mailman.zfn.uni-bremen.de/cgi-bin/mailman/listinfo/dftb-plus-user

-- 
Dr. Gabriele Penazzi
BCCMS - University of Bremen
 
http://www.bccms.uni-bremen.de/
http://sites.google.com/site/gabrielepenazzi/
 
phone: +49 (0) 421 218 9328
fax: +49 (0) 421 218 4764
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20161031/8a403b61/attachment-0001.html>


More information about the DFTB-Plus-User mailing list