Absolute value of the imaginary part of the dielectric function

[Daniele Varsano]

Dear Quxia,
this is the direction of the electric field, e.g. linear polarization of the light, orthogonal to the light porpagation.
Best,
Daniele

[Quxiao]

Dear Daniele：
First of all thank you very much for your patience！
I want to study the anisotropy of the system.
Now it seems that the direction of the incident light can be reflected by the electric field E in yambo.
In order to use the software better, I want to confirm whether my understanding is correct.
If I want to study the incident normal to the plane (incidence in the Z direction), then set Blongdir to 1|1|0?
This is equivalent to the vibration direction in the plane, and if you want to study the incidence parallel to the plane, set it to 0|0|1,
right?
Can the specific phase difference corresponding to the circularly polarized light be set? Do I need to use yambo_kerr?
Thanks for your help again!
Best wishes
Quxiao
BIT

[Quxiao]

Dear Daniele：
I re-read your previous reply and have some thoughts.
LongDrXs determines the direction of the incident electric field, which is used to calculate the dielectric tensor, while BlongDir is used to determine the direction of absorption, right? So if I calculate different incident directions, I need to change the LongDrXs，is it?
Thanks a lot for your help!
Best wishes!
Quxiao
BIT

[Daniele Varsano]

Dear Quixiao,

if you want to calculate absorption in different directions you need to change BlongDir as it affects the dipoles and hence the strength of the excitation.
The screening also can be affected in anysotropic materials and you can look at the impact of LongDrXs, but usually, this is not critical as it will affect only one element of the matrix W_oo(q=0). Usually considering an average direction (1,1,1) for bulk or (1,1,0) for 2D material is a good approximation for the average value.

Best,
Daniele

[Quxiao]

Dear Daniele：
Thanks for your reply!I have got that with your help!
Because I used the Coulomb truncation technique to calculate the BSE, and the eps I got was very small.
I read the link you gave, but I was still a little confused. It was because the Coulomb truncation technique was used to treat the unit cell as infinite, and then the volume was very large. is large, so the resulting eps is very small,is it ？
But the value I derived from the formula for deriving eps from the imaginary part in o-*.alpha does not match the value of the imaginary part of eps output from the o-*.eps file, because What is the reason, am I misunderstanding?
There is another question, I know that the unit of alpha is born, and the unit of eps is undefined, right? According to the dimension, it seems to be None.
Please give me some help!
Looking forward to your reply!
Best wishes
Quxiao
BIT

[Daniele Varsano]

Dear Quixiao,

Eps is very small and goes to zero for infinite volume if you see it in 3D (you are simulating an array of layers) and adding more and more vacuum you find the eps of vacuum (1,0). Obviously you are not using an infinite volume, but you are using a truncated potential the Coulomb interaction and it does not diverge anymore as 1/q^2.
eps=1+vX and X-->0 as q^2, for 2D Vcoul~1/q, for 1D ~ln(q).
So for dimensions lower than 3 you will get very small, and again the vacuum dielectric constant for q->0.

Yes, eps is a number, it has no dimensions.

Best,

Daniele

[Quxiao]

Dear Daniele：
Thank you for your previous help, I carefully read your reply and the course on YouTube.Now,I got that!
Thanks!
Best wishes!
BIT
Quxiao