The anisotropy in BSE calculation

[Guo_BIT]

Dear developers:
Currently, I would like to perform direction-specific BSE calculations in a two-dimensional system, just as done in Fig. 3d of the article
https://www.nature.com/articles/s41563-022-01285-3
However, I have a few questions below：
（1）In my GW calculations, I set

Code: Select all

% LongDrXp
 1.000000 | 1.000000 | 0.000000 |

to calculate static screening in plane since my system is 2D. Then, in my BSE calculations,

Code: Select all

% BLongDir
 0.000000 | 1.000000 | 0.000000 |

was set to get the absorption spectra along the y direction. Is my process correct?

（2）When performing BSE calculations without considering anisotropy, I will set

Code: Select all

% BLongDir
 1.000000 | 1.000000 | 0.000000 |

which means calculating along the 45° direction in the xy plane. So, is it correct to set

Code: Select all

% BLongDir
 0.866 | 0.5 | 0.000000 |

if I want to perform BSE calculations when deviating 30 degrees along the x direction?

Best wishes,
Guo

[Daniele Varsano]

Dear Jingda,

Yes, both 1) and 2) are correct!

Best wishes,

Daniele

[Davide Sangalli]

Let me add that, in case your system is not isotropic in plane, you can use the following input for GW

Code: Select all

% LongDrXp
 1.000000 | 1.000000 | 0.000000 |
%
OptDipAverage

The logical `OptDipAverage` triggers an averaging procedure for the q=0 term of the screening.
The average is done over the Cartesian directions selected in the variable `LongDrXp` , e.g. in this case x=(1,0,0) and y=(0,1,0).
Without the logical, the screening at q=0 is instead computed along the direction (1,1,0)

This feature is available from version 5.1

Best,
D.