[Pw_forum] converging V_xc in the vacuum

Giuseppe Mattioli giuseppe.mattioli at ism.cnr.it
Fri Feb 16 18:52:48 CET 2018


Dear Janos

Sorry, I've not an answer for your problem. But as a good practice you  
should not use an old version of the code. 5.2.0 is not only almost 3  
years old. It also belongs to the previous 5.x major release of QE (we  
are now using 6.2.1). Let us suppose that there is a bug in the code  
(I'm not saying that *there is* one, routines which calculate  
potentials might be older than my grandmother and just as robust!).  
Nobody is really interested in debugging old versions of the code. I  
suggest therefore that you reproduce the problem with a recent  
(possibly the latest) stable version of the code and post it again here!
HTH
Giuseppe

Quoting Janos Daru <janos.daru at gmail.com>:

> Dear experts,
>
> I would like to print out plane averaged exchange-correlation potential for
> a PBE Cu(111) slab along its normal vector (z-direction).
> To this end I have printed out plot_num:
>
> 1  = total potential V_bare + V_H + V_xc
> and
> 11 = the V_bare + V_H potential
>
> and subtracted the second from the first and averaged it.
> However, unlike in JCTC 2009,5,881-886 (figure 3.) or in Nano Lett.
> 2015,2448-2455 (figure 3.) V_xc does not converge to zero in the vacuum
> region (it has an approximate value of  -0.015 Hartree).
>
> The SI from the second paper authors explains that:
> "we note that it is well-known that in the vacuum the exchange-correlation
> (xc) part converges much more slowly than the electrostatic part of the
> total potential. Therefore, when computing the local potential of bare
> graphite(0001) and Au(111), we employed more strict settings in the
> band-structure calculations. Specifically, we increased the number of
> FFT-grid points and enhanced the representation of the charge density,
> potential and augmentation charges. Furthermore, we have employed an
> additional support grid for the augmentation charges, which further
> increases the accuracy of calculated quantities."
>
> So I tried to tighten ecutrho, the realspace grid, the convergence
> criteria, I have increased the vacuum, number of Cu layers etc. but V_xc
> does not converge to zero in the vacuum. Could some of the experts explain
> to me if it is due to a wrong setup, or eg. due to a different reference
> level?
>
> I have attached the full input, and V_xc plot. I have used QE 5.2.0 for my
> calculations.
>
> Thank you for the answer in anticipation,
>
> Janos Daru



GIUSEPPE MATTIOLI
CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
Via Salaria Km 29,300 - C.P. 10
I-00015 - Monterotondo Scalo (RM)
Mob (*preferred*) +39 373 7305625
Tel + 39 06 90672342 - Fax +39 06 90672316
E-mail: <giuseppe.mattioli at ism.cnr.it>




More information about the users mailing list