[Pw_forum] Calculation of affinity energy and ionization

[Nicholas E. Singh-Miller]

Hi Trinh Vo,

Because you are working within periodic boundary condtions (whereas 
Melnikov and Chelikowsky were not).  You have some errors in the energy to 
consider when charging the system.

Assuming that your input file is as you want it (without other sources of 
error).

A charged system (with a compensating background) will have an 
error in the energy that is dependent on the size of the unit cell.  It is 
likely that you need to correct for this, either as post processing 
through a scheme like Makov and Payne suggest (PRB 51, 4014) or self 
consistently as in PRB 77, 115139 (or references therein).  I believe both 
are to some extent implemented in QE.

Hope that helps,

Nick

On Thu, 18 Jun 2009, Vo, Trinh wrote:

> Hello,
>
> As a test, I performed the calculations of the ionization and affinity energy of P-doped Si nanocrystal.  The steps I did as follows:
>
>
> 1.  Relax the nanostructure
> 2.  Remove one electron form the above relaxed structure and run scf calculation for this system to obtain ionization energy (I).
> 3.  Add one electron the relaxed structure (in #1) , and run scf calculation to obtain the affinity energy (A).
>
> The ionization energy and the affinity energy are calculated by:
>
>                        A = E(n)-E(n+1)
>                        I = E(n-1)-E(n)
> where E is the total state total energy of the n-, (n+1)-, and (n-1)-electron systems.
>
> When I compared the values of A and I that I obtained with those by a previous work, PRL 92, 0468021 (2004), I saw a large deviation.  The difference is ~0.8eV for I and ~0.4eV for A.  (I used the same structure of one the structures studied in the PRL. By the courtesy, Prof. Cheliskowki sent it to me for the test purpose).
>
> The calculation in the PRL is performed with real-space ab initio pseudopotential method.  Their calculations of hyperfine splitting agreed very well with experimental data.
>
> Since my results have large differences from theirs, I am not sure if the procedures that I did above are incorrect.  Could you give some comments if there is anything wrong from what I did.
>
> Thank you in advance your help,
>
> Trinh Vo
>
> JPL/CalTech
>
>
>
>
>
>
>
>

*****************************************
Nicholas E. Singh-Miller
Prof. Marzari Group (quasiamore.mit.edu)
Materials Science and Engineering
Massachusetts Institute of Technology
13-4066
(617)324-0372
*****************************************