Bulk Silicon Tddft

Laurens

unread,

Mar 29, 2021, 9:10:37 AM3/29/21

to cp2k

Dear all,

I am trying to do tddft simulations of bulk silicon (please see attached band structure and in/output files).

Since tddft is not implemented for k-points, I have tried using supercells of sizes up to and including 5x5x5 times the primitive unit cell.

This changes the lowest excitation energy, but unfortunately the lowest excitation energy does not seem to converge to a value close to the calculated bandgap of 0.7 eV: the lowest excitation energy decreases from an energy of 2.2 eV for the single unit cell, to an energy between 0.28 eV (4x4x4 cell) and 0.85 eV (3x3x3 cell).

Does anybody have any suggestions on tddft simulations of bulk solids?

Best regards,

Laurens

Silicon_TD4_MOLOPT_out.txt

Silicon_TD2_MOLOPT_out.txt

Silicon_BS_MOLOPT_inp.txt

Silicon_TD3_MOLOPT_out.txt

Silicon_TD_MOLOPT_inp.txt

Silicon_BS.png

Silicon_TD5_MOLOPT_inp.txt

Silicon_TD2_MOLOPT_inp.txt

Silicon_TD4_MOLOPT_inp.txt

Silicon_TD5_MOLOPT_out.txt

Silicon_TD3_MOLOPT_inp.txt

Silicon_TD_MOLOPT_out.txt

Lucas Lodeiro

unread,

Mar 29, 2021, 11:33:38 AM3/29/21

to cp...@googlegroups.com

Hi,

I am not an experto on TDDFT, but seeing your output there is something suspicious. For TD1 you have 4 states occupied (thinking in the 0K electronic structure), and correctly the TDDFT transitions occur from these 4 occupied states to the virtual states in the "Gamma Point". But when you double (or higher) the cell, the excitations occur from states over the occupied states. For example in TD2: 32 occupied states, and the lower citations occur from 38... the same occur for TD3, TD4, TD5.

It seems that the smearing pumps fractions of electrons to virtual states, and the program assumes that virtual-low populated states are occupied states.

If this is the problem, you can solve it by removing the smearing... Si is a semiconductor, and without smearing is possible to compute it electronic structure.

Regards

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/7c817d8a-23c9-4f6c-95af-48cdb4bdaef2n%40googlegroups.com.

Matt W

unread,

Mar 29, 2021, 11:35:00 AM3/29/21

to cp2k

Your tddft states seem to coincide with the HOMO-LUMO gaps of the different systems - as they should for delocalised bulk states. That the HOMO-LUMO values are small seems could suggest the LUMO calc has not converged (maybe) or that your geometry is not replicating correctly?

Matt

Matt W

unread,

Mar 29, 2021, 11:42:04 AM3/29/21

to cp2k

Yes, this is likely the problem. I don't think TDDFT with smearing has been tested / considered.

Laurens

unread,

Mar 31, 2021, 5:45:41 AM3/31/21

to cp2k

Many thanks for the suggestions.

The lowest excitation energies indeed look more reasonable when not using smearing.

Laurens

Reply all

Reply to author

Forward