[Pw_forum] Problems with electrostatic corrections (Makov-Payne or density counter charge) for aperiodic systems in 4.2 version

[Oliviero Andreussi]

1) Roughly speaking, eq. 15 of the original Makov-Payne paper express
the energy of periodic system (E) in terms of the energy of the isolated
system (E0) plus the energy of interaction between the periodic replicas
(E11 and E12). Neglecting the fact that the E12 term has the wrong sing
in the original equation (as it is stressed in a comment in the code),
to obtain the energy of the isolated system you want to subtract E11 and
E12 from your periodic total energy, as it is done in the code, i.e.
E0=E-E11-E12.

2) The DCC correction was not bugged or bit-rotting, and, as far as I know,
it was also working for 2D cases. It was removed from the code in order to clean it
and make pw's compilation easier: this is because DCC required an external
multigrid solver an a somehow cumbersome implementation. One of the
developers of the original DCC code (Ismaila Dabo) is now working on a
different version of the DCC that do not require multigrid solvers.

Best,

Oliviero Andreussi

Postdoctoral Associate MIT-DMSE,
77 Massachusetts Ave, Office 13-4084,
Cambridge, MA, 02139 USA

Visiting Researcher, Oxford University,
Departments of Materials, Rex Richards Building,
Park Road, Oxford, Oxon, OX1 3PH, UK

-------- Original Message --------
>  Subject: 	[Pw_forum] Problems with electrostatic corrections
>  (Makov-Payne or density counter charge) for aperiodic systems in 4.2 version
>  Date: 	Tue, 5 Apr 2011 13:33:22 +0100
>  From: 	ilyes hamdi<iiysidi at gmail.com>
>  Reply-To: 	PWSCF Forum<pw_forum at pwscf.org>
>  To: 	pw_forum at pwscf.org<pw_forum at pwscf.org>
>
>
>
>  Dear espresso users,
>
>  1) I'm a bit confused with the implementation of the Makov-Payne
>  electrostatic corrections in PW/makov-payne.f90 in Version 4.2:
>  - In their original paper, Makov and Payne define the first order
>  correction E11 =  - \alpha * q^2/(2\epsilon* L). This quantity should be
>  negative for cubic systems, however
>  the writing statement in PW/makov-payne.f90 subroutine defines the
>  correction to be -E11 (leading to a positive value).
>  - The second order correction E12 is propertional to q Q/L^3 and is also
>  defined as -E12 in the PW/makov-payne.f90 subroutine.
>  - The total energy should be Etot = E0 + E11 + E12 = E0 - \alpha *
>  q^2/(2\epsilon* L) + 2 \pi * q * Q / ( 3 \epsilon L^3), however it is
>  implemented as E0-E11-E12.
>
>  2) Why the DCC correction is disabled in version 4.2? Indeed when
>  setting asume_isolated = 'dcc' and&EE input parameters to their
>  default  values, I obtain an error message "DCC correction is disabled".
>  Looking in the PW/input.f90 subroutine, one found that the dcc
>  correction is disabled by an immediate call to errore subroutine.
>
>  Please can any one verify if I'm correct or did I miss something.
>
>  Best regards
>
>
>