[Wannier] kmesh_automatic problems

[Elio Physics]

Dear all,


I am trying to obtain the bands of a 42 atom supercell using Wannier functions. The cell vectors do not actually pertain to a well known symmetry. they are of the form:


Begin Unit_Cell_cart
bohr
42.82955445336 0.00008690366 0.000000
0.0000188921 9.28402379356  0.000000
0.00000000 0.000000000 28.32670894424
End Unit_Cell_cart



When I computed the bands in QE, I used the path created by XCrySDen. However this is causing problems in W90 code giving errors such as: (I tried several grids produced by XCrySDen)


 kmesh_shell_automatic: Singular Value Decomposition has found a very small singular value

                                                                                  and

Non symmetric k points!


On the other hand trying a 625 point grid produced by the kmesh utility solves the problem but the bands are kind of wiggly and do not correctly produce the bands. My question is : Is this due to the use of grid produced by kmesh  which is regular and the fact that my cell vectors do not belong to a given symmetry? Can this be solved by increasing number of points. In other words  can W90 code predict the bands for such a supercell  and what could be a possible reason for the wiggling behavior?


N.B : I have ibrav=0 and the option 'random' was used in projections


Regards


Elio

University of Rondonia

Brazil

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20170809/8f393568/attachment.html>