[DFTB-Plus-User] R: electron phonon interaction and matrix derivative

Mon Jun 4 09:16:10 CEST 2018

Dear Dirk,

I can also add an answer to this question.

At the time when we worked on electron-vibron scattering in molecular junctions we implemented all calculations in a very old version of DFTB. 

In the current version of dftb+ all that machinery is still missing. But the approaching strategy should be changed a lot. In that old code I changed the core routines to perform the numerical differences:

[H_ab (dq) – H_ab(0)] / dq   and  overlap.

In practice I ended with an executable that was able to do only this calculation.

In dftb+ the strategy should be quite different and has not been developed yet. 

This is not the end of the story, because libNEGF currently has only part of the code needed to compute inelastic scattering. For sure it is not equipped to load and use the gamma^q_ab couplings. [The reason is that in libNEGF we wanted to do something much better and efficient than in the old code but things remained behind].

So I am sorry to say that with the current version of dftb+ it is not possible to do real electron-phonon scattering or IETS through molecular junctions. But there is ongoing work to complete the implementations.

Regards,

Alessandro Pecchia 

Da: DFTB-Plus-User [mailto:dftb-plus-user-bounces at mailman.zfn.uni-bremen.de] Per conto di Gabriele Penazzi
Inviato: venerdì 25 maggio 2018 15:31
A: User list for DFTB+ related questions
Oggetto: Re: [DFTB-Plus-User] electron phonon interaction and matrix derivative

Hi Dirk,

I first calculate all the modes, and then dH/dR_ix, dH/dR_ix where R_ix is the 'x' component of the normale displacement for the mode 'i'. It can surely be done better, but it was good enough for me as I was looking at few modes and performance was a non-issue.

2018-05-25 14:46 GMT+02:00 Dirk Ziemann <dirk.ziemann at physik.hu-berlin.de>:

thank you for the quick and helpful answer. 

this goes in the right direction and I will have a deeper look into your program. At a first look, I do not see, how you get the derivative of the overlap S and hamilton matrix H. Is it possible to get it from dftb+ (SCC or non-SCC) or did you calculated it by yourself?  I think this is the most crucial part of the el-ph coupling. 
from the papers on the negf transport calculations , I assume that these matrices are calculated in the dftb+ mpi-negf code and maybe also the el-ph coupling (gamma) on the level of atomic orbitals: 

gamma_ab^q   \approx  d H_ab / d_q  -  \sum_mn  d S_am / d_q  S^-1_mn H_nb - ...
(atomic orbitals a,b; vibration q)

it would be really nice to get these matrices or coupling out of the program. 

thank you again and best wishes  dirk

On 25/05/18 13:14, Gabriele Penazzi wrote:

Hi Dirk,

I assume you are speaking about the formulation as used for example for IETS spectra in molecular junctions.

I don't think that is is supported in the current version, even though there may be ongoing work on that.

Some time ago I wanted to calculate the coupling matrices for some molecules, and I did some scripting to help on that.

It worked only on molecules (or in Gamma, for bulk), and I would guess it won't run without hassle 

with the new versions of DFTB+ because it has been written 3 or 4 years ago and never used again, but maybe

if that's what you need, you can use it as inspiration. It's public domain: https://bitbucket.org/gpenazzi/dftbtools (lok in elcoupl,py).

It has also the ASE package (Atomic Simulation Environment) as a requirement. 

2018-05-25 11:27 GMT+02:00 Dirk Ziemann <dirk.ziemann at physik.hu-berlin.de>:

Dear all,

I am interested in the electron phonon interaction like it is performed 
in the negf version and described in the corresponding papers.

Is it possible to get the coupling matrix in dependence on the single 
particle states and phonon modes out of the negf version? Or is it 
possible to get the matrix derivative of the overlap/hamiltonian with 
respect to x/y/z direction? or something similar to simply create the 
coupling by "hand"?

Thank you very much and best wishes

dirk

p.s. I posted this question some time ago, but it was never answered. because I am still interested in this topic I want to ask this again. 

-- 
Dirk Ziemann
Humboldt Universität zu Berlin, Institut für Physik 

_______________________________________________
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

-- 

-- 
Gabriele Penazzi
mobile: +45 51894907
skype: gabriele.penazzi

_______________________________________________
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

-- 
Dirk Ziemann
Humboldt Universität zu Berlin, Institut für Physik 
AG Photobiophysik 
Newtonstr. 15, D-12489 Berlin
Room     1'406
Phone    +49 30 2093 4997
E-mail   ziemann at physik.hu-berlin.de

_______________________________________________
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

-- 

-- 
Gabriele Penazzi
mobile: +45 51894907
skype: gabriele.penazzi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20180604/d0c02451/attachment.htm>