[Pw_forum] Problem w/ 1st BZ of orthorhombic structures in QE

Cameron Foss cjfoss at umass.edu
Sat Oct 4 00:52:24 CEST 2014


Hello all,

I am trying to getting phonon frequencies of biaxially strained silicon and
germanium. Particularly, the lattice constants in the three dimensions are
as such ax = ay != az in either case(!= refers to "not equal to").

To do so I chose ibrav=10 which is an orthorhombic fcc structure, and I
explicitly define A,B,C,cosAB,cosAC,cosBC. where A=B!=C and all cosines are
0 (simply because the angles remain at 90 degrees)

Now everything is OK, until I want to sample the 1BZ. For sampling in the 3
Cartesian directions, ie (100), (010), (001), the matdyn code seems to
assume that the lattice constant in all three directions is the same (eg it
assumes the value of "A"). Which in the case of the (001) direction is a
problem, since the true lattice constant in that direction is C.  (C!=A)

Unfortunately I have not looked through enough of the code to distinguish a
way around this without just compensating for this when I define each
q-vector that I pass on to matdyn. However, this seems like a hassle in the
case of a true orthorhombic structure where ax!=ay!=az.

Has anyone else experienced this and/or have a better alternative to simply
defining scaled q-vectors?

Best,
Cameron
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20141003/d8fa0245/attachment.html>


More information about the users mailing list