[Pw_forum] about phonon and structural optimization

[Eyvaz Isaev]

Hi, 

First, I would like to answer questions about phonons.

1. No, I did nothing with cubic BN, but I remember a
paper of Parlinski.

2. You can calculate LO/TO splitting for
semiconductors and insulators  using both dynmat.x and
matdyn.x. The result will be the same. 
Input file for dynmat.x looks like

 phonons of Ti2SiC at 0.0000000,0.0000000,0.0000000
 &input
  amass(1)=47.87,
  amass(2)=28.09,
  amass(3)=12.01,
  asr=.true.,
  fildyn='Ti2SiC.dynG',
  filout='Ti2SiC.out',
  filmol='Ti2SiC.molden'
 &end
0.0000000,0.0000000,0.0000000

Command line as usual 
$BIN_DIR/dynmat.x < input_file > output_file

The manual says correct, of course, but in this
particular case you have a choice. 

3. For GGA in most cases  larger lattice parameter
will be obtained, so, phonon frequences reduced. 

4. 10cm^-1 of 1304cm^-1 is about 10*100/1300=0.8%. I
guess you have got good agreement with experimental
data.

Bests,
Eyvaz.

--- sun jason <jasonsun98 at hotmail.com> wrote:

> Hi, eyvaz,
> 
> Thank you to answer my questions:)
> 
> did you calculate the phonon mode of cubic BN
> before?
> 
> today, I recalculate the phonon using PBE
> potentials, the result is better 
> now, but the LO mode is 1294 cm-1, about 10 cm-1
> less than the experimental 
> value.
> 
> I calculated the LO/TO spliting using q2r.x + 
> matdyn.x.
>                                            
> ~~~~~~~~~~~~
> 
> but the user manual says, the dynmat.x should be use
> to calculate LO/TO 
> spliting and IR.
>                             ~~~~~~~~~~~
> 
> so which program do you use to get LO/TO spliting?
> 
> 
> if you use pwtools dynmat.x, how to use it?
> 
> Thank you in advance.
> 
> 
> 
> another question is about cell optimization.
> 
> 2) does pw.x can calculate from the optimized cell
> automatically? or I 
> should read the output file of vc-relax and write
> the coordinate into the 
> input file of pw.x manually? 
> 
> 3) I perform a vc-relaxtion using PBE USPP, the
> optimized lattice constants 
> is larger than the experimental value. why? it's
> reasonable?
> 
> 
> jason
> 
> Message: 6
> Date: Wed, 6 Jul 2005 09:51:30 -0700 (PDT)
> From: Eyvaz Isaev <eyvaz_isaev at yahoo.com>
> Subject: Re: [Pw_forum] about phonon calculation
> To: pw_forum at pwscf.org
> Reply-To: pw_forum at pwscf.org
> 
> Hi,
> 
> In order to get correct LO mode you should calculate
> phonons using dynmat.x from /pwtools directory.
> 
> Effective charges for B and N are not equal due to
> numerical errors in our computer calculations. 
> 
> Bests,
> Eyvaz.
> 
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
> 



		
__________________________________ 
Yahoo! Mail 
Stay connected, organized, and protected. Take the tour: 
http://tour.mail.yahoo.com/mailtour.html