[Pw_forum] question about bands calculation

[lukethulin at netscape.net]

hai-ping,

>From a 2003 thesis by Giovanni Cangiani, but obviously there are other sources:

Gamma (0,0,0)
N (0, 1/2, 0)
X (0,0,1/2)
P (1/4,1/4,1/4)
Z (1/2,1/2,-1/2)

I don't recall getting any errors or strange results when choosing paths between these points.  You should search for this thesis as a good reference for titania.

Luke
  ----- Original Message ----- 
  From: lan haiping 
  To: pw_forum at pwscf.org 
  Sent: Monday, October 23, 2006 12:38 AM
  Subject: Re: [Pw_forum] question about bands calculation


  Dear Luke,
  thanks for your reply. my structure is anatase titania.
  Actually, i set the k-point paths according to high symmetry lines such as \Gamma[0,0,0] to Z [0,0,0.5] , Z to Q[0,0.5,0.5]  etc.
  Is it right ?

  Regards,
  hai-ping

   
  On 10/22/06, lukethulin at netscape.net <lukethulin at netscape.net> wrote: 
    Hai-ping,

    I don't think you're going about a band structure calculation correctly.  The scf calculation should only use automatic k-point selection of (4,4,4) or something like that.  The following nscf calculation needs you to input k-paths along high symmetry lines that vary by crystal structure.  What form of titania are you modeling, anatase, rutile? 

    Luke
      ----- Original Message ----- 
      From: lan haiping 
      To: pw_forum 
      Sent: Saturday, October 21, 2006 2:31 PM
      Subject: [Pw_forum] question about bands calculation

       
      Hi!

        i have a question about bands calculation .
       Just after finished a 'scf' calculation for tio2, i performed a series (about 5)  'nscf' calculations with  different k-grid paths. Except k-grid path, all settings are the same for these series ' nscf '  calculation.   
      In some nscf-output,  the values of "Highest occupied level and lowest unoccuped level"  are  found. But  these values are not found in the other nscf-output . I donot know what the reasons are related to . Morever, i also came to a calculation failed with complains "Too many eigenvalues are not converged".  

      Regards,
      hai-ping


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20061023/87316a02/attachment.html>