[Wien] Bandstructure calculation

David Pankhurst david.pankhurst at materials.oxford.ac.uk
Thu Feb 19 12:29:30 CET 2004


Dear ChiYung,

I don't think the weights have any meaning for a band structure plot and
therefore it should not matter what values you put in (authors?).

I think the weights are only important when carrying out Brillouin zone
integrals on a uniform k-mesh, as you do during total energy calculations.  
In such cases, to save time, we use symmetry and only do calculations at
k-points within the irreducible wedge of the 1st BZ.  The weight of a
particular point tells the program the total number of points of that same
symmetry existing in the full BZ.  Hence, the weight of Gamma is always 1
and the weight of any point x 0 0 (where x<0.5) is 6 in the simple cubic
case.

In a band-structure plot you are just plotting the eigenvalues of the
secular equation along particular directions in reciprocal space and thus
the weights are unimportant.  I hope this helps.

best regards,

	Dave Pankhurst

On Thu, 19 Feb 2004, ChiYung, Yam wrote:

> Thank you for your help.
> 
> In the case of band structure calculation from GAMMA to X for a simple cubic
> unit cell. From the templates,
> 
> GAMMA         0    0    0   40 12.0
>               1    0    0   20 12.0
>               2    0    0   20  6.0
>               3    0    0   20  8.0
>               4    0    0   20 24.0
> DELTA         5    0    0   20 12.0
>               6    0    0   20 24.0
>               7    0    0   20 12.0
>               8    0    0   20 12.0
>               9    0    0   20  6.0
> X            10    0    0   20  8.0
> 
> if I want to change to 60 points, how should I change the weight of each point?
> Would you give me some examples? Thank you.
> 
> Best regards,
> ChiYung, Yam
> 
> On Thu, 19 Feb 2004, Stefaan Cottenier wrote:
> 
> > NOTE FOR THE AUTHORS: a description of case.klist seems to be absent in the
> > UG. Maybe something to add to section 6.5 ?
> > 
> > And now to your question:
> > 
> > > I use the following klist in job.in1 for calculating bandstructure. what
> > > is the meaning for each column? If I want to change from 20 points to 60
> > > points for the same range, how should I change the klist ?
> > >
> > > R             0    0   20   40  8.0 -0.5  1.5      simple cubic  template
> > >               0    0   19   40 24.0
> > >               0    0   18   40 24.0
> > >               0    0   17   40 12.0
> > 
> > first column: proportional to the k_x component (must be integer)
> > second column: proportional to the k_y component (integer)
> > third column: proportional to the k_z component (integer)
> > fourth column: the proportionality constant (integer, divide the preceding
> > three numbers by this one)
> > 
> > Hence, your second line '0 0 19 40' means a k-vector with components (0/40,
> > 0/40, 19/40).
> > 
> > fourth column: the weight of this point (depends on the point symmetry at
> > that point in the BZ).
> > 
> > To insert more points for the same range, you can
> > 
> > 1) either use Xcrysden (most straightforward)
> > 2) add additional points according to the recipy given above [(0/80, 0/80,
> > 39/80) etc.], but you have to care about the weights.
> > 
> > Stefaan
> > 
> > 
> > 
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > 
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> 










More information about the Wien mailing list