[Pw_forum] high symmetric Points

Ari P Seitsonen Ari.P.Seitsonen at iki.fi
Sat May 3 22:17:00 CEST 2014


Dear Nh Ben,

   You can choose in the input of QE which units should be used for the k 
points, so it does not matter which ones you obtain from different codes 
_so_long_as_ you use the ones that correspond to the coordinates that are 
given. So I don't see why the values given by XCrysDen would be wrong.

   Personally I conclude and/or calculate the k points for band structures 
myself using an editor, not using any tool. It takes more time, is more 
error-prone, but I hope I know what I am doing there [quoted from Sledge 
Hammer], or at least I know when to suspect an error (the bands come to 
high-symmetry points with a non-zero derivative or the degenaries do not 
match etc). I recommend to do it once oneself, in a "simple" case where 
you can verify your results comparing to elder, that is, previous 
calculations. In this particular case of a hexagonal system the 
coordinates in Cartesian units include the sqrt(3), that's why I suspected 
that the second set of coordinates was given in units of 2pi/a, whereas 
in the units of the reci-procal lattice vectors the coordinates are more 
like (1/2,0), (1/3,1/3) for the famous point K in graphene, ... And to 
conclude symmetries the reci-procal units ("crystal" in the input for QE) 
is usually easier in the case of non-trivial cells. Like here, the 
(-1/3,2/3) = (-1/3,-1/3) = (1/3,1/3) if the time-reversal symmetry holds 
(so that you can invert k => -k).

   Hopefully this helps something, or at least didn't waste too much of 
your time. :)

     Greetings,

        apsi

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
   Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
   Institut für Chemie der Universität Zürich, CH-8057 Zürich
   Tel: +41 44 63 54 497  /  Mobile: +41 79 71 90 935


On Sat, 3 May 2014, BENYAHIA NEZHA wrote:

> Dear Ari Paavo Seitsonen
> 
> First, thank you for your response. You're so helpfull.
> If i understood, Xcrysden generates a crystal coordinates(Reciprocal coordinates) but it gave me a wrong dispersion curve when plotted. By your response you mean that the
> coordinates which were pasted below (and were generated by plothon) are in the tpiba(2pi/a) unit, without forgetting that these ones gave me the right dispersion curve. Can
> xcrysden give me a tpiba coordinates inspite of reciprocal ones? and how?
> In this case how to generate tpiba coordinates(which give me the right dispersion curve) without using Plotphon?????
> 
> I will be very happy if you could enlightened me more on this "as the second point looks like (-1/3,2/3) above but contains apparently the 1/sqrt(3) ~= 0.57735... in the generated
> k points. What is 1/sqrt(3) ~= 0.57735???? 
> I'am sorry if i have asked a lot of questions but i really need all of them!!!!!!
> 
> 
> Regards.
> Nh Ben
> 
> 
>


More information about the users mailing list