Momentum dependent Independent-particle-approximation

[Fadil]

Dear all,

I am calculating q-dependent dielectric function for a metallic system.
But I realized that the program does not calculate all the q-points I gave as input and after calculating some q points it stopped without error. (r-IP_optics_chi_dipoles_q34)

I encountered another problem when I was changing the electric field direction with 'LongDrXd' to calculate the dielectric function in different directions.
Except for q1, even if I set the electric field direction in different directions, it calculates the same dielectric function.(see the epsilon files in the attachment)

Best

[Daniele Varsano]

Dear Fadil,

I encountered another problem when I was changing the electric field direction with 'LongDrXd' to calculate the dielectric function in different directions.
Except for q1, even if I set the electric field direction in different directions, it calculates the same dielectric function.(see the epsilon files in the attachment)

This is correct, the 'LongDrXd' set the direction in the q->0 limit and it is meaningful for iq=1. For the other q the direction is given by the finite momentum of the selected q point.

after calculating some q points it stopped without error.

This should be investigated, anyway before trying to spot the problem, let me ask you to update to a more recent versione of the code as many improvements and change happened in the newer releases:

Code: Select all

git clone git@github.com:yambo-code/yambo.git yambo

If the problem persists post here again providing the report again, input file and eventually any useful info (error message if any) in one of the log files.

Best,
Daniele

[Fadil]

Dear Daniele,

Thank you for the response.

This is correct, the 'LongDrXd' set the direction in the q->0 limit and it is meaningful for iq=1. For the other q the direction is given by the finite momentum of the selected q point.

So, since a two-dimensional material is not periodic in the z direction (no q in the z direction), how can we calculate \eps_{z}?

[Daniele Varsano]

Dear Fadil,

you can calculate the lim q-->0 in the z direction using LongDrXd=0,0,1
you cannot calculate the dispersion wrt q, but this is not very meaningful as a 2D system has not dispersion in that direction (flat bands).

Best,

Daniele

[Fadil]

Thank you!

Best

[Fadil]

Dear all

Is it possible to define kpoint mesh externally and calculate the momentum dependent GW-BSE spectrums?
When I define the kpoint mesh irregularly, it does not recognize my kpoint mesh and it is not possible to generate the input files with created SAVE files.

Best

[Davide Sangalli]

Dear Fadil,
yambo can compute finite-q optical properties selecting the q-points q=k-k' where both k and k' belong to a regular grid.
It is not possible to use random grids or any other grid.

Best,
D.

[Fadil]

I see.
Thank you!

[Fadil]

Dear all,

From previous posts I know for iq=1, in the IPA, using the 'LongDrXd' we can extract the dielectric tensor components (bulk material) along the xx, yy, zz-directions.
Is it possible to obtain dielectric tensor components for the finite-q calculations (iq/=1)?
What I mean is that can we extract z-component of dielectric function in a specific point like K-point (for q in the plane) in the Brillouin Zone?

Best

[Daniele Varsano]

Dear Fadil,
when looking at finite q, you are simulating an eels spectrum, so the perturbation is longitudinal in the direction of the incoming electron.

Best,
Daniele