The ABINIT Users Mailing List ( CLOSED )

["Horacio Wagner Leite Alves" <hwlalves@ufsj.edu.br>]

Dear Yunhee,

I think you have made some mistake about the concepts in
order to calculate the DDB file for the Si(100) surface. It is true that the
H is an artificial boundary condition for your slab, in order to avoid
fictitious dipoles between the two Si surfaces and to facilitate the
self-consistency in your total energy calculations. However, in your DDB
calculations, you have calculated the DDB contribution for all the Si atoms
and avoid the H ones. If you look inside your DDB file you will note that
there. That is the reason that the anaddb claims that your DDB is
incomplete. So, you have two choices: i) Include the H atoms in all
calculations and, at the the end of anaddb run, do not interprete the Si-H
vibrational modes(which have higher frequencies) - you can do that by
regarding the eigenvectors, and; ii) rerun the DDB calculations but, without
the H ones. In this case I hope you will have a lot of difficulties
in all SCF convergence cycles during the run.

I hope I had
understanding your problem.... good continuation!!!  

Horacio W.
LEITE ALVES
Departamento de Ciências Naturais
Universidade Federal de
São João del Rei
Caixa Postal: 110 CEP: 36.300-000
São João del Rei -
MG
Fone: (55)(32)33.79.24.89
Fax: (55)(32)33.79.24.83

--------- Mensagem Original --------
De: forum@abinit.org
Para: "forum@abinit.org" <forum@abinit.org>
Assunto: Re: [abinit-forum] The dynamical matrix was incomplete : phonon frequencies may be wrong ...
Data: 20/06/06 06:05


Dear Javier,

Thank you very much for your kind reply and explanations about my questions.
And I have two linked questions.

Q1)
I considered only 20 atoms among 24 atoms to calculate the phonon frequencies,
so this error message is occured ... from your explanations.
I calculate the Si(001) surface, so this system has 20 Si atoms and 4 H atoms.
To know the Si(001) surface phonon band, the 4 H atoms is not important,
so I removed the 4 H atoms to calculate the dynamical matrix.
According to this error messages, shoul I have to add the 4 H atoms as well to calculate dynamical matrix?
If I removed 4 H atoms, is their phonon band not correct?

Q2)
You said, "Anyway often these incomplete dynamical matrices can be merged (using
> MRGDDB) in some cases in order to get a bigger matrix that can be
> analyzed with ANADDB..."
that means the when I merged the dynamical matrices, do that probelms disappear?
What does that mean the bigger matrix that can be analyzed with ANADDB?

Thank you again and I'm waiting for your reply~!

Best regards,
Yunhee Chang


----- Original Message -----
From: "Javier Torres" <javier.torres@unito.it>
To: <forum@abinit.org>
Sent: Monday, June 19, 2006 6:58 PM
Subject: Re: [abinit-forum] The dynamical matrix was incomplete : phonon frequencies may be wrong ...


> Dear Yunhee
>
> The message that you obtain during the rf calculation is a consequence
> of a incomplete dynamical matrix... A complete dynamical matrix is
> generated considering the displacement of all the irreducible atoms in
> your unit cell in the three cell directions...
> Ex:
> If you have 6 atoms in your unit cell you have to define
>
> rfphon 1
> rfatpol 1 6 # it means from atom 1 to atom 6
> rfdir 1 1 1 # it means displacements along a, b and c directions
>
> If your have atoms related by symmetry the calculation is performed only
> for the irreducible set of atoms and directions, th ematrix then is
> completed considering symmetry...
>
> The message, that you observe, is displayed when you calculate the
> phonon frequencies considering only some atoms or one direction;
> therefore ABINIT tell you that the obtined phonons may be wrong...
> Anyway often these incomplete dynamical matrices can be merged (using
> MRGDDB) in some cases in order to get a bigger matrix that can be
> analyzed with ANADDB...
>
> Best Regards
>
> F. Torres
>
>
>

________________________________________________
Message sent using UebiMiau 2.7.2