Meaning of the tag LongDrXp

[Longjun Xiang]

Dear developers:

I have a technical question related to the tag LongDrXp: In your tutorials, the tag means direction of the electric field for the calculation of the q=0 component of the dielectric constant. However, in terms of your implementation paper:yambo: An ab initio tool for excited state calculations, the formula (6)-(9) (see https://www.sciencedirect.com/science/a ... 5509000472) are devoted to calculating the dielectric constant with PPA approximation. But where is the direction of the electrical field in these formulae? Can I just set (1,1,1) for bulk system and (0,0,1) for two-dimensional systems？ But what's the physical meaning of this direction?

[Daniele Varsano]

Daer Longjun,
the electric field appears in the definition of dipoles, ie the rho_nn appearing in EQ.8.
You can have a look to the new yambo paper where this issue is explained in details:
https://iopscience.iop.org/article/10.1 ... 48X/ab15d0
See sec.3.1

Anyway, this is the direction of the electric field, or if you want the polarization of the photon perturbing your system.
Setting (1,1,1) for a bulk system is a good idea (you are somehow averaging in the 3 directions). For a 2D system (e.g. lying in xy direction) a good way to set the LongDrXp variable could be (1,1,0), i.e. light polarized in the plane of the system.

Best,
Daniele

[Reza_Reza]

Daer Daniele,

I want to change the direction of the electrical field to compute optical absorption. In that case, I should change BLongDir in the BSE input file. I want to know if it is essential to make the same change in GW (LongDrXp) and screening calculation too, since I perform BSE on top of quasiparticle energies.

Best,

Reza

[Daniele Varsano]

Dear Reza,

I want to know if it is essential to make the same change in GW (LongDrXp) and screening calculation too

No, it is not essential, if you set an average direction e.g. (1,1,1) for a bulk or (1,1,0) for a 2D system, the screening will not depend much on the direction.

Best,

Daniele