Phonon dispersion curve of SLG

GULP
1

Hi Prof. Julian,

I am using GULP in order to produce phonon dispersion curve for Single Layer Graphene(SLG). I am using two types of unit cells:

Case (a) SLG with 2 basis atoms with basis vectors at 60 degree to each other.
Case (b) SLG with 96 basis atoms with orthogonal basis vectors.

For Case (a) I am getting somewhat close dispersion and phonon density of states(PDOS) curves to what research articles have been producing (max. freq of ~50 THz).
But in Case (b) the dispersion curve has its maximum frequency increased by ~10 THz. Also the PDOS is slightly changed.

–
Questions:
(1) Should there have such a change of frequencies. If yes then why? (Is it happening because of bigger dynamical matrix has now been created with 96 basis atoms?)

(2) A slight change in the x-direction basis vectors length without changing the coordinates of the atoms for Case (b) changes the dispersion curve. What should be an ideal basis vector in case of a rectangular unit cell. Should it be the maximum of x-coordinate of all the atoms of the structure?

–
I am also attaching the Input GULP file for your reference.

(I am using REBO potential with C-C bond length of 1.3968418 to produce a 60 X 4.19 Angs rectangular unit cell of SLG.)

disp.gin (5.2 KB)

Thanks and Regards.

2

The short answer is that the frequencies you get depend on the cell you are using. There is a direct relationship between supercells and points in the Brillouin zone for the original (smaller) cell. If you use a larger cell therefore you are sampling multiple k points of the small cell even at gamma. Therefore if you are trying to reproduce something from the literature you need to use the same cell and same direction (or do some fun maths to deconvolute the data back into the representation you need!)

1 Like