how to plot graphene phonon band structure having 18 atoms in unit cell

[Gagandeep]

I want to ask do you have any idea about Graphene Phonon bandstructures.I have plotted for graphene with two atoms in basis cell..I want to plot for 3*3 sheet of graphene..How to plot in ATK..unable to understand in latest manual too or how to choose density matrix..Kindly help..it is written in manual

The dynamical matrix is calculated using the finite difference method in a repeated cell, which is sometimes also referred to as frozen phonon- or super-cell method.

How to pursue further

[zh]

By using the Brillouin zone folding, you can  draw the phonon dispersion of a 3x3 supercell from that of a primitive unit cell. This may save much more computing time.

Alternatively, you may treat the 9x9 superell sheet of graphene by using the customized UnitCell() rather than the Hexagonal(), and then calculate the phonon by following the usual steps.  This may be time-consuming calculation.

[ams_nanolab]

You may use ATK classical for phonons, it may save time
Please see these tutorials
http://quantumwise.com/publications/tutorials/itemlist/category/94-phonons-and-thermal-properties

[Gagandeep]

How to use  the Brillouin zone folding in ATK...Tutorials are for 2 atoms in unit cell of graphene not for 18 atom or 32 atom sheet..

[Gagandeep]

I created 3*3 supercell from database using 7*7*1  k-points using ATK-DFT and hexagonal unit cell....i opimized the sheet first..then dragging the minimized sheet in script,i added phonon bandstructures from analysis..then changed k-sampling to 111 in each direction and 3*3*1 in dynamical matrix..and 200 segment in phonon bandstructures and brillioun zone to Gamma, M, K, G....is it ok?or should i change it to ATK-classical and use Brenner or Tersoff

[zh]

How to use  the Brillouin zone folding in ATK...Tutorials are for 2 atoms in unit cell of graphene not for 18 atom or 32 atom sheet..

This is not implemented in ATK. You have to write a script to implement it by yourself.

[Anders Blom]

It would be a good idea to consider why you really need the dispersion for the supercell, since physically it's equivalent to that of the primitive cell... And the states in the supercell can always be obtained just by mapping the k-point in the large cell to a corresponding point in the small one (this is the zone folding procedure).

[Gagandeep]

#reply 6 Thanks for the reply but can you please elaborate it

[Gagandeep]

Dear Dr Blom
I want to calculate Phonon band structure for 18/32 atom graphene sheet and then adsorb it by some metal (1/18 or 1/32 )and see the changes

[Petr Khomyakov]

To see the change of phonon frequencies of graphene, you may first do a supercell calculation for a single graphene sheet on the metal surface of your interest at the actual equilibrium separation distance between the graphene sheet and metal surface.

You then increase the separation distance to a larger value to significantly reduce the interaction between the graphene sheet and metal surface, and do a supercell calculation once again. This should give you an idea on how the phonon frequencies change upon decoupling graphene from the metal surface, i.e., upon increasing the separation distance.

For the sake of comparison, you may also want doing two additional supercell calculations for a free-standing graphene sheet and a bare metal surface without graphene. For both calculations, one should adopt the same supercell as that in the graphene-metal system.

All that information should allow you making some conclusions on the effect of the metal substrate on the phonon frequencies of supported graphene.       

[Gagandeep]

Thanks for the reply sir..But my problem is whether i am going accurately or not as i earlier written Suppose i want to prepare 50 atoms graphene sheet and study its phonons...I took 2 atoms from data base , repeated hexagonal unit cell in 4*4.. and for phonons i set 3*3*1 dynamical matrix....It will give me density matrix for 144 atoms..




alternatively i choose 8 atoms in unit cell(i dont know whether i should customize unit cell() instead of hexagonal for graphene or remain it hexa only which is default), and then set dynamical matrix for 5*5*1 which gives me matrix for 200 atoms...

my question is whether iam getting phonons for 144/200 atoms of graphene sheet or 32 atom/ 8 atoms in both cases respectively..


Thanks in anticipation
 

[Anders Blom]

The repetitions for the dynamical matrix are just a numerical detail. If you compute 8 atoms with whatever repetition, you get the phonon spectrum for 8 atoms. The number of repetitions will however influence the accuracy of the results. It's not certain 3x3 is enough - cf. the tutorial on phonon spectrum of Si, http://quantumwise.com/publications/tutorials/item/836-silicon-phonon-bandstructure