Specify (identify) the effect of phonons on carrier transportation

[weixiang]

Dear ATK experts,
I would like to specify or identify the effect of the phonons on the current of a  GNR TFET device at different temperatures.
Is it reasonable to compare the current calculated from a STD configuration with the current calculated from its corresponding pristine configuration at the same temperature? I.e. the I_STD denote the current with both electron ( the Fermi-Dirac distribution)  and the phonon effect, and the I_pristine denote the current with only the electron effect.
And by the way, can I say that the I_pristine = I_noninteracting? as shown in this figure from this paper Phys. Rev. B 96, 161404(R)


Thanks!

[Petr Khomyakov]

And by the way, can I say that the I_pristine = I_noninteracting? as shown in this figure from this paper Phys. Rev. B 96, 161404(R)

Noninteracting  means that no electron-phonon interaction is included in the current calculation. If this is what you call 'pristine', then I_pristine = I_noninteracting.

[weixiang]

Thanks for replying. Is there a way to know the contribution of each phonon branch (LA, LO, TA, TO, etc) or each mode to the current difference?
I found out this page https://docs.quantumwise.com/tutorials/inelastic_current_in_si_pn_junction/inelastic_current_in_si_pn_junction.html
that by calculating the inelastic transmission spectrum, it is possible to specify that say  the phonon that contributes the most to the current is that with index
66, and energy ℏω=63.01meV. But this calculation is based on LOE and XLOE methods, which are very expensive. Is there any other faster approach in ATK that can do the similar thing like this?
.

[Petr Khomyakov]

What I can say is that this cannot be done with the special thermal displacement method or similar, because phonon modes are not computed explicitly. So, you have to use LOE or XLOE for doing that analysis.

[weixiang]

Thanks！

[weixiang]

I have the following two questions for using the XLOE method to calculate the Inelastic transmission spectrum:
(1) How to specify properly the energy range [E₀,E₁], and the grid range and sampling of k_A, k_B and q_A, q_B?
(2) How does the time complexity of calculating the  Hamiltonian derivatives and Inelastic transmission spectrum (using the Phonon energy intervals method) roughly scale with the device atom number N? ( say, for eaxmple NlogN, N², etc)

Thanks!

[weixiang]

Can anyone answer my question? Please help!