Marcos, Arnaud, I just happen to be looking at the denchar code right now.. It appears that it just takes the real part of the wave function and complains (probably a lot) if the complex part is non-zero. So if the real part is what you're looking for (unlikely, I know), then it'll do the right thing. Otherwise, it looks like it'll plot correctly either the gamma point or any k-point on the BZ boundary, like X, L, etc, since those should be real as well.
Best, Lucas On 7/12/07, Marcos Verissimo Alves <[EMAIL PROTECTED]> wrote:
Arnaud, There was a whole thread of discussions and attempts to overcome the problem... From what I remember, the whole thing ended with the conclusion that the current version of denchar would only plot wavefunctions at the Gamma point, even though supposedly there would be a version that could handle any k-point, but that version would not have been included in the present version of siesta. I remember Andrei Postnikov suggesting that a large supercell be used such that the desired k-point would fold on to the Gamma point, and then plot the appropriate state. As far as I know, the problem hasn't been really solved, in the sense that users that don't want to tackle coding apparently still can't plot the wavefunctions just by using denchar. If I'm not mistaken, wavefunctions at the Gamma point should be real. If the version of denchar included with siesta supports only gamma-point, then this could explain all the warnings on complex wavefunctions (I haven't gone into the code, though, so this is a sort of educated guess). So, for the moment, perhaps Postnikov's suggestion is the best, although a bit awkward: to plot the appropriate state found from the folding of the band structure in a large supercell. Hope this helps. Marcos > Like I have written in my first mail, I have already checked the > archives. > > The answer was (on 23 Aug 2006) that it was possible to plot > wavefunction at k-point other than Gamma with Denchar. Denchar should > reads the WFS file produced by Siesta and generates two CUBE files > -modulus and phase-. > > Now, there are 2 problems with this : > > 1 - If there is multiple k-points in the WFS file, only this message is > displayed : > Wavefunctions file contains more then 1 k-point > DENCHAR can only handle the Gamma point!! > > 2 - If ther is only one k-point in the WFS file, Denchar generates only > 1 CUBE file for each state instead of the 2 CUBE files it should > (-modulus and -phase) and this message gets printed a lot of times : > Warning: complex wavefunctions in file! > > The first problem is not important because I can generate WFS file with > only 1 k-point at the time > > The second problem is more important because I don't know what is in the > CUBE file produced since only 1 file is generated and not 2. > > Is it the modulus, the phase or something else ? > > > Thank you in advance > > Arnaud Maillard > > > On Wed, 2007-07-11 at 17:11 +0200, Marcos Verissimo Alves wrote: >> Arnaud, >> This subject has come up some time ago in the list. Check the archives. >> Marcos >> > Dear Siesta users, >> > I am trying to plot wavefunction at specific k-point. I have tried to >> > use DENCHAR but I get this message : >> > >> > Wavefunctions file contains more then 1 k-point >> > DENCHAR can only handle the Gamma point!! >> > >> > I am confused because I found in the mailing list archive this message >> > saying that it was possible to plot wavefunction at selected k-point >> > >> >> Wed, 23 Aug 2006 >> >> >> >> I am looking for the simplest way to plot the wavefunction obtained by >> >> siesta calculation. Is there some program which allows to plot the >> >> wavefunction >> >> at a selected k-point? >> > >> >> Yes, there is. It is called Denchar, it's >> >> distributed with Siesta and it's located in the >> >> Util directory. Follow the instructions in the >> >> Denchar manual. >> >> >> >> It reads the WFS file produced by Siesta and >> >> generates two CUBE files -modulus and phase- for >> >> each state you asked for. >> > >> > I'm using DENCHAR 1.3 , is there a newer version that could handle >> k-point >> > other than Gamma or is there an other to plot these wavefunctions ? >> > >> > ps. >> > The .fdf file used to generate the .WFS file contain these lines : >> > >> > WriteWaveFunctions .true. >> > WaveFuncKPointsScale ReciprocalLatticeVectors >> > %block WaveFuncKPoints >> > 0.000 0.000 0.000 from 238 to 242 # Gamma >> > 0.000 0.500 0.000 from 238 to 242 # >> > 0.000 0.000 0.500 from 238 to 242 # >> > %endblock WaveFuncKPoints >> > >> > Thank you >> > >> > Arnaud Maillard >> > > -- Dr. Marcos Verissimo Alves Post-Doctoral Fellow Condensed Matter and Statistical Physics Sector International Centre for Theoretical Physics Trieste, Italy -------- I have become so addicted to vi that I try to exit OpenOffice by typing :wq!
