[QE-users] force constant

[Lorenzo Paulatto]

> But the problem is this gives espresso.ifc2.xml which cannot use with 
> ShengBTE. If there is a way to convert *.xml to normal format it will work.


I found in my old files one input that should work:

  &input
    fildyn  =  'BOGUS.xml'
    zasr    =  'crystal',
    flfrc   =  'fc'
  /
4 4 4
13
dyn1
dyn2
dyn3
dyn4
dyn5
dyn6
dyn7
dyn8
dyn9
dyn10
dyn11
dyn12
dyn13

So, the value of fildyn is nonsense, but it must end with "xml" to tell 
q2r to look for xml files, than the list of files do not have the xml 
postfix. Of course you will get a useless XML dynamical matrix file, 
unless you modify slightly PH/PHonon/q2r.f90 (or PH/PHonon/d0_q2r.f90, 
depending on the version) changing
  IF (xmldyn) THEN
     IF (lrigid) THEN
        CALL write_dyn_mat_header( flfrc, ntyp, nat, ibrav, nspin_mag,  &


to

  IF (.FALSE.) THEN
     IF (lrigid) THEN
        CALL write_dyn_mat_header( flfrc, ntyp, nat, ibrav, nspin_mag,  &


By the way, newer versions of the phonon code do not produce xml 
dynamical matrix files, unless you specifically ask for it by ending the 
fildyn variable with "xml".

Another way to get normal dynamical matrix files from xml ones is do 
q2r, then matdyn with the input variable "fldyn" specified for all the 
q-points in the dynmat0 file to produce the dynamical matrix in ascii 
form (missing the header, which you can do by hand starting from a 
similar system) Finally apply q2qstar.x to each of these files to obtain 
all the q-points in the star.

Finally, as you are using a competing code, I'll plug my own, it does 
three-body force constants from perturbation theory, phonon lifetime, 
thermal conductivity and all that:
  http://sf.net/p/d3q

kind regards


-- 
Lorenzo Paulatto - Paris