[Phonopy-users] RV: Nanoribbon primitive axis (--pa)

[Pablo Á. Z. <pab...@ho...>]

Hello everyone,


I am trying to compute the thermal conductivity using pwscf interface of a graphene nanoribbon. Thus, i have calculated the forces using phono3py propperly, replicating periodicaly the system in the x direction, while setting a vaccum space along y and z directions. My input system is a 80 atoms slab with the width I want in the y direction, while as i said, it is replicated along the x axis in the calculation in QE, so i selected the 1x1x1 supercell, since the system is a supercell itself.


The problem I would like to ask is when calculating the thermal conductivity using this command:


phono3py --pwscf --dim="1 1 1" -c input.in --pa="..." \
   --mesh="3 1 1" --fc3 --fc2 --br

which --pa option (primitive axis) should i set? As i understand from this  debate: https://sourceforge.net/p/phonopy/mailman/message/32790661/ , and from every example in <https://github.com/atztogo/phono3py/> https://github.com/atztogo/phono3py/, this numbers represent the matrix (Mp) which link the vectors  used in the unit cell  (set by POSCAR), with those of the primitive cell:  u_primitive=u_poscar·Mp (https://atztogo.github.io/phonopy/setting-tags.html). Since  the primitive cell of graphene is already know : a( sqrt(3)/2, 1/2 ,0 );  a( sqrt(3)/2, -1/2,0 ), a=2.46 Angs, I can easily calculate Mp values. However, if i use       phonopy -c  --pwscf input.in --symmetry        the primitive axis in the pposcar output file which phonopy seems to find for my system are completely different!: (4.2614 0 0 ; 0 10.0033 0 ; 0 0 35.01172)
Thus, the question is, which primitive axis (--pa option) should i use?

These are the cell parameters used in input.in which provide a vaccum space between y and z replicated slabs:

 CELL_PARAMETERS {bohr}
  16.10586        0.000        0.000
  0.000       66.16257        0.000
  0.000        0.000       18.90359


Thank you for your patience!