Absolute value of the imaginary part of the dielectric function

[Quxiao]

Dear Deniel:
First of all thank you for your continued help !Recently,I have a doubt. The calculated imaginary part of the dielectric function is used for drawing. Is the imaginary part of the obtained dielectric function a relative value or an absolute value? How should the absolute value of the imaginary part of the dielectric function be obtained? My intention is to compare the size of the imaginary part of the dielectric function of the two systems .Please give me some help again.
Thanks
Quxiao
BIT

[Daniele Varsano]

Dear Quixiao,

The dielectric function is a dimensionless number and you can compare it. Note that this is true for a bulk system while for a reduced dimensionality system it is ill-defined and you have a dependence on the supercell volume. In that case, you should look at the polarizability alpha.

Best,
Daniele

[Quxiao]

Dear Deniel:
I have a question about the light direction.When calculating BSE, the setting of the light direction, or the setting of the electric field direction. My image here is not very clear. For example, q refers to the incident direction of light or the incident direction of electric field. Then I have seen in Learn that the direction is controlled by LongDrX. I guess this is the parameter for calculating the response function. There is a parameter BLongDir that also controls the direction of the electric field. For example, I am studying a two-dimensional material and calculate the incidence in the plane, such as the incidence in the x direction. Do you need to adjust these two parameters to 1.000000 | 0.000000 | 0.000000 | ?Arbitrary incidence in the plane, change to 1.000000 | 1.000000 | 0.000000 |?
Please give me some advice?
Thanks a lot!
Quxiao
BIT

[Daniele Varsano]

Dear Quixiao,

it is the direction of the electric field.

Then I have seen in Learn that the direction is controlled by LongDrX. I guess this is the parameter for calculating the response function

Yes

There is a parameter BLongDir that also controls the direction of the electric field

Same meaning but for the BSE absorption.

Do you need to adjust these two parameters to 1.000000 | 0.000000 | 0.000000 | ?Arbitrary incidence in the plane, change to 1.000000 | 1.000000 | 0.000000 |?

If the system is isotropic this does not make any difference. If there is anisotropy I would set LongDrX as (1|1|0) for the calculation of the screening and then I would look at the absorption for the two directions (1,0) and (0,1) separately.

Best,
Daniele

[Quxiao]

Dear Deniel：
First of all, I wish you a happy new year！Thank you for your answer. I seem to understand what you mean. If it is an anisotropic material, I will set longdrxs = 110 to calculate the accurate screening I also have a question, is it necessary to set 111 for each anisotropic block, so as to ensure the calculation in three directions，is that right？
Quxiao
BIT

[Quxiao]

Dear Deniel:
Hello, I also found a problem. After calculating the BSE, I analyzed the weight file calculated by yambo, which is o-yambo out. exc_weights file combined with r_setup file. According to the abscissa for calculating exciton weight obtained from the setup file, it is found that many of the listed points are not on the high symmetry point and high symmetry line of the simple Brillouin region I selected, which makes me wonder what are those points and can they be abandoned?I'm sure I can get a consistent conclusion based on the mutual conversion between cc and iku and the two types of coordinates in a file.
However,When I further want to judge the method of scattering points according to the scattering points in Brillouin zone, I find that some points are not scattered on the high symmetry points and high symmetry lines, but in other deviated positions. Therefore, I have some doubts. I don't know why such problems occur. Is it not just scattering points on the high symmetry points and high symmetry lines during calculation?
Best wishes
Quxiao
BIT

[Daniele Varsano]

Dear Quixiao,
sorry but I have not understood your problem, can you explain try to explain your doubts maybe with some examples?

Best,
Daniele

[Quxiao]

Dear Daniele：
You mentioned that when calculating various anisotropic materials, such as two-dimensional materials, and the vacuum layer is built on the Z axis, when calculating the screening, LongDrXs takes the direction of 1|1|0, and then if I calculate the absorption in the plane, I want to take the directions of BLongDir 1|0|0 and 0|1|0 respectively, right? I don't know if there is a problem with my understanding. If we calculate isotropic materials, there is no difference between LongDrXs taking 1|1|0 and taking 1|0|0, right? In addition, if it is a three-dimensional anisotropic material, do you want LongDrXs to take 1|1|1 when calculating the shielding, and then BLongDir respectively.
Please give me some advice!
Thanks！
Quxiao
BIT

[Daniele Varsano]

Dear Quxiao,

LongDrXs gives you the q->0 limit for the calculation of the screening (not that this is just one element of the screening matrix and its contribution will be smaller and smaller as sampling increases). BLongDir is the polarization of the light in the absorption.

If the system is isotropic different directions of the electric field will provide the same outcome.
In a three-dimensional anisotropic material, 1|1|1 essentially provides a good estimation of a proper average on the three cartesian directions so you have a good estimation of the head of the dielectric matrix.

About BLongDir usually, it is instructive to look at the three directions separately, the three spectra provide you information about the anisotropy and they can also be compared with experiments of absorption of linearly polarized light in oriented samples. Of course you can also look at the absorption for light in the (1,1,1) direction.

Best,
Daniele

[Quxiao]

Dear Daniele：
Thanks for your previous help ,I need your help again
I am calculating the GW of a two-dimensional system. Do I need to add the Coulomb truncation parameter when calculating, such as yambo -r ？My current input is yambo -g n -p p -V all, I see that the input command you replied to me before was yambo -X p -g n -V all, where X p stands for the inverse dielectric matrix with ppa,the -p p is the correlation part of the self energy.
And I found it seems that I typed yambo -p p -g n -V all,by default, I use PPA to approximate the inverse dielectric even though I didn't type the yambo -X p command ，is it?
Another question is, do I need to set the Coulomb truncation (yambo -r) to calculate GW in my two-dimensional system? Does it speed up the calculation(Monte Carlo integral) or must it be set?
The correct command to calculate the GW band of the two dimensional materials is the yambo -r -g n -p p -V all?or yambo -g n -p p -X p -V all,or yambo -r -X p -g n -p p -V all?I'm a little confused .
Thanks
Quxiao
BIT