[Q-e-developers] Erratum: Re: Re: Integral of total potentail and charge density

[Jiqiang Li]

Sorry,
The previous email has an error in question b), which is labelled by underline.
it should be:
a) I am sorry that a similar integration seemed not working. My result of \int[rho(r)dr] was not exactly the total number of electrons, which differed by about 0.1 for a supercell of GaN of 72 atoms, with relative error 0.0001. The number of significant digits is more than 14 in QE. So, I am not sure whether the difference arised from the numerical noise or I made some mistake. I used the data in real space from pp.x.
 
b)Actually, I want to evaluate the electron-phonon coupling constants by explicitly finite displacements of all the atoms along three directons. Now I use the rho(r) = <f|i> (where |i> and |f> are initial and final states respectively, we only consider the case of Gamma point only calculation, wavefunctions is real array, thus <f|i> is also real. I output it by add some sentences in the code of QE) and v1(r) belong to the equilibrum atomic configuration, and v2(r) corresponding to the displaced one, then \int[(v2-v1)×rho/dR×dr] is the local contribution to el-ph constants, am I right?  Before carrying out this calculation, the made a benmark test for the code outputing |i> and |f>.  When I sum over all the bands, i.e., let |i> = |f>,  \sum_i { <i|i>} is exactly equal to the rho output by pp.x.

 
c)Now I want to revise some part of code to add function for QE, are there any DETAILED tutorials about the code of QE? The  lecture slides online seem too short for practically coding.


 
From: jqli14
Date: 2015-06-09 01:37
To: General discussion list for Quantum ESPRESSO developers
Subject: Re: [Q-e-developers] Integral of total potentail and charge density
Dear Paolo,
 
a) I am sorry that a similar integration seemed not working. My result of \int[rho(r)dr] was not exactly the total number of electrons, which differed by 0.001. The number of significant digits is more than 14 in QE. So, I am not sure whether the difference arised from the numerical noise or I made some mistake. I used the data in real space from pp.x 
 
b)Actually, I want to evaluate the electron-phonon coupling constants by explicitly finite displacements of all the atoms along three directons. Now I use the rho(r) and v1(r) belong to the equilibrum atomic configuration, and v2(r) corresponding to the displaced one, then \int[(v2-v1)×rho/dR×dr] is the local contribution to el-ph constants, am I right?
 
c)Now I want to revise some part of code to add function for QE, are there any DETAILED tutorials about the code of QE? The  lecture slides online seem too short for practically coding.
 
Thanks!
 
Jiqiang
 
in 2015-06-09 00:38:01,"Paolo Giannozzi" <p.giannozzi at gmail.com> Compose:
 
>
>
>On Sat, Jun 6, 2015 at 4:40 PM, Jiqiang Li <jqli14 at fudan.edu.cn> wrote:
> 
>a) I am wondering whether the integral can be simply handled as \sum_i {V(i) * rho(i)} * volume / N, where volume of unit cell and N the size of array for V and rho. If no, any other formula?
>
>
>yes: all integrals in QE are computed that way, on the Discrete Fourier Transform grid
>
>
>
>b) Is the integral above equal to the sum of 'hartree contribution" and "xc contribution"  in the stdout file?
>
>
>no: "xc contribution" =  variable etxc =  \int \epsilon_{xc}(r) * rho(r) d^3r
>while \int V_{xc}(r) * rho(r) d^3r is contained in variable vtxc (not printed in output).
> 
>
>Paolo
>
>
>Or where can I find some information in the stdout file as benchmark data?
>  
>     Any consideration will be gratefully appreciated.
>
>
>
>
>Best regards!
>
>
>Jiqiang Li
>Fudan University, China
>
>_______________________________________________
>Q-e-developers mailing list
>Q-e-developers at qe-forge.org
>http://qe-forge.org/mailman/listinfo/q-e-developers
>
>
>
>
>
>--
>
>Paolo Giannozzi, Dept. Chemistry&Physics&Environment,
>Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
>hone +39-0432-558216, fax +39-0432-558222
 
 
 
 
 
_______________________________________________
Q-e-developers mailing list
Q-e-developers at qe-forge.org
http://qe-forge.org/mailman/listinfo/q-e-developers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20150609/7904e45c/attachment.html>