What K lattice points are the optical matrix elements based on?

[lyzhao]

Dear all,
I use the PW to do the gound state calculation, then use the yambo 4.5.3 to run the BSE calculation.
As a test, I run the jobs using the 2D-hBN monolayer with the kgrid of 3x3x1.
PW prints three reduced k points and correspoinding weight:

Code: Select all

0.0000000   0.0000000   0.0000000  0.11111
0.0000000   0.3333333   0.0000000  0.66666
0.3333333   0.3333333   0.0000000  0.22222

I modify the yambo code, yambo-4.5.3\src\dipoles\DIPOLE_driver.F, to print the k points which the optical momentum elements are calculated based on.
at line 252, I add the following code

Code: Select all

........
    do i_sp_pol=1,n_sp_pol
      write (*,*) "nibz",Xk%nibz
      write (*,*) "pt1", Xk%pt(1,:)
      write (*,*) "pt2", Xk%pt(2,:)
      write (*,*) "pt3", Xk%pt(3,:)
      write (*,*) "ptbz",shape(Xk%ptbz)
      write (*,*) "ptbz1", Xk%ptbz(1,:)
      write (*,*) "w",Xk%weights
      do ik=1,Xk%nibz
        !
        if (.not.PAR_IND_DIPk_ibz%element_1D(ik)) cycle
  .......

I obtain the following results,

Code: Select all

 
 nibz           3
 pt1   0.00000000       0.00000000       0.00000000    
 pt2   0.00000000      0.333333313       0.00000000    
 pt3  0.333333343      0.499999970       0.00000000    
 ptbz           9           3
 ptbz1
 w  0.111111112      0.666666687      0.222222224    

I find that the third k point is different from that given by PW, though they have the same weigth (i.e., 0.22222).

So, my questions are
1. what K lattice points the optical momentum elements are based on in yambo?
2. what is the Xk%ptbz and how to output it? I think it is the unreduced k-grids becasue its shape is (9,3) as output above.
However, if I try to output it by write (*,*) "ptbz1", Xk%ptbz(1,:) for a test. Yambo prints nothing for it.

Any help will be appreciated.

[lyzhao]

I also try to use "ypp -k k" to print k grid based on the BSE work directory.
ypp -F ypp.in -J bse
I obtain the same k-grids as PW, as follow

Code: Select all

 
 ....................
 <---> Q-points  (BZ):  9
 <---> K-points  (BZ):  9
 <---> [06.01] == Built-in IBZ K-grid [rlu] ==
       0.000000000  0.000000000  0.000000000  1.000000000
       0.000000000  0.333333313  0.000000000  6.000000000
       0.333333373  0.333333284  0.000000000  2.000000000
 <---> [06.02] == GW K-grid analyzer ==
 ...................
 

Best regards.

[Daniele Varsano]

Dear Youzhao,

it is just a definition of k point units rlu (reciprocal lattice unit)-->iku (internal k units)
You can find the correspondence in the report e.g. the setup, by typing:

Code: Select all

yambo --fatlog 

or use

Code: Select all

ypp -grid k 

uncommenting ListPts and indicating Input/Output coordinates unit.

Best,

Daniele

[lyzhao]

Dear Daniele,
Great thanks for your quick help.
Can I print the non-reduced k-points in yambo?
In my test, it is 9 points.

Best regards.

[Daniele Varsano]

Dear Youzhao,

using ypp:

Code: Select all

ypp -grid k 

uncomment the keyword

Code: Select all

ExpandPts

then run ypp.
You will have the k-point grid in the entire BZ.

Best,
Daniele

[lyzhao]

Dear all,
I'm writing a program to do other optical calculations based on the output of a BSE calculation.
Now, I need the optical matrix elements (OME) based on the non-reduced k-points.
I have obtained the OME based on the reduced k-points with helps of Daniele and Davide.
To obtain the OME based on the non-reduced k-points,
I need "the list of non-reduced k-points" used in the BSE calcualtion,
and "the symmetry operator table" which may be used to unfold the reduced to non-reduced k points in the BSE or BSE ypp calculation.

Where can I find these two things?
With Daniele's help, I have found these two things from ypp -k k and r_setup , but I'm not sure if they are for BSE or BSE ypp calculations.
If yes, which unit should I use? rlu, iku, cc, or alat?

Best regards.

[lyzhao]

Dear all,
I am happy to say that I have obtained what I want from Xk%sstar.
For my efforts, any other comments are welcome.

Best regards.