How to treat low dimensional systems: Difference between revisions

From The Yambo Project
Jump to navigation Jump to search
No edit summary
Line 1: Line 1:
In this tutorial you will learn for a low-dimensional (2D) material how to:
In this tutorial you will learn for a low-dimensional (2D) material how to:


* use the Random Integration Method (RIM) to avoid numerical divergence of the self-energy when low-dimensional dense k-sampling is used
* avoid numerical divergence problems when low-dimensional dense k-grids are used using the Random Integration Method
* generate a truncated coulomb potential with a box-like cutoff  
* generate a truncated coulomb potential with a box-like cutoff to eliminate the image-image interactions
* use this truncated coulomb potential in the GW calculation
* use this truncated coulomb potential in the GW calculation
* use this truncated coulomb potential in the BSE calculation
* use this truncated coulomb potential in the BSE calculation
Line 15: Line 15:




==Use the RIM to avoid the numerical divergence of the self-energy when low-dimensional dense k-sampling is used==
==avoid numerical divergence problems when low-dimensional dense k-grids are used using the Random Integration Method==
In a low-dimensional material like 2D-hBN less than 3D k-grids are used: like n x n x 1.  
In a low-dimensional material, like 2D-hBN, less than 3D k-grids are generally used like in this tutorial where a 6x6x1 k-grid is used.
This can create a numerical instability of the q-space integration due to the presence of the FT of the Coulomb Potential in all the main equations
This can create a numerical instability of the q-space integration due to the presence of the FT Coulomb Potential in all the main equations
(see i.e. the exchange self-energy equation)
(see i.e. the exchange self-energy equation)


To eliminate this problem YAMBO uses the so-called Random Integration Method
which means to use a Monte Carlo integration using random Q-points whose number RandQpts is given in input.
Create the input to generate the ndb.RIM database
$ yambo -F yambo_RIM.in -r
  RandQpts= 1000000                  # [RIM] Number of random q-points in the BZ
  RandGvec= 1            RL    # [RIM] Coulomb interaction RS components
Close input and Run yambo
$ yambo -F yambo_RIM.in -J 2D
At the end in the 2D directory you will have a new database ndb.RIM
We underline that the evidence of the divergence problem is not evident with the 6x6x1 k-grid used in this tutorial.
It becomes evident using  denser k-grids. (try exchange self-energy runs using the SAVE directories generated with
15x15x1 and 30x030x1 k-grids)
Here we report the HF gap calculated with the 3 k-grids without (noRIM) and with Random Inetgration Method (RIM)
            noRIM  RIM
  6x6x1 
  15x15x1
  30x30x1


The calculation of the self-energy requires a BZ integration performed numerically on a suitable k-grid.
The use of a low-dimensional k-sampling (i.e. in a 2D material a 2D k-grid like n x n x 1)
implies a numerical divergence due to the q=0 term of the Coulomb potential (see i.e. the exchange self-energy espression)
when a dense BZ sampling is used.
To eliminate the numerical divergence YAMBO use the so-called Random Integration Method
performing a Monte Carlo integration near q=0 using a set of random Q-points whose number RandQpts is given in the input


Create the input to generate the ndb.RIM database
Create the input to generate the ndb.RIM database

Revision as of 15:15, 27 March 2017

In this tutorial you will learn for a low-dimensional (2D) material how to:

  • avoid numerical divergence problems when low-dimensional dense k-grids are used using the Random Integration Method
  • generate a truncated coulomb potential with a box-like cutoff to eliminate the image-image interactions
  • use this truncated coulomb potential in the GW calculation
  • use this truncated coulomb potential in the BSE calculation
  • analyze the difference with corresponding calculations without the use of a truncated potetnial

Prerequisites


avoid numerical divergence problems when low-dimensional dense k-grids are used using the Random Integration Method

In a low-dimensional material, like 2D-hBN, less than 3D k-grids are generally used like in this tutorial where a 6x6x1 k-grid is used. This can create a numerical instability of the q-space integration due to the presence of the FT Coulomb Potential in all the main equations (see i.e. the exchange self-energy equation)

To eliminate this problem YAMBO uses the so-called Random Integration Method which means to use a Monte Carlo integration using random Q-points whose number RandQpts is given in input.

Create the input to generate the ndb.RIM database

$ yambo -F yambo_RIM.in -r 
 RandQpts= 1000000                   # [RIM] Number of random q-points in the BZ
 RandGvec= 1            RL    # [RIM] Coulomb interaction RS components 


Close input and Run yambo

$ yambo -F yambo_RIM.in -J 2D

At the end in the 2D directory you will have a new database ndb.RIM


We underline that the evidence of the divergence problem is not evident with the 6x6x1 k-grid used in this tutorial. It becomes evident using denser k-grids. (try exchange self-energy runs using the SAVE directories generated with 15x15x1 and 30x030x1 k-grids)

Here we report the HF gap calculated with the 3 k-grids without (noRIM) and with Random Inetgration Method (RIM)

            noRIM   RIM
 6x6x1  
 15x15x1
 30x30x1


Create the input to generate the ndb.RIM database

$ yambo -F yambo_RIM.in -r 
 RandQpts= 1000000                   # [RIM] Number of random q-points in the BZ
 RandGvec= 1            RL    # [RIM] Coulomb interaction RS components 


Close input and Run yambo

$ yambo -F yambo_RIM.in -J 2D

At the end in the 2D directory you will have a new database ndb.RIM

Generate a truncated coulomb potential/ndb.cutoff database (yambo -r)

To simulate an isolated nano-material a convergence with cell vacuum size is in principle required, like in the DFT runs. The use of a truncated Coulomb potential allows to achieve faster convergence eliminating the interaction between the repeated images along the non-periodic direction (see i.e. D. Varsano et al Phys. Rev. B and .. ) In this tutorial we learn how to generate a box-like cutoff for a 2D system with the non-periodic direction along z.

In YAMBO you can use :

spherical   cutoff (for 0D systems)  
cylindrical cutoff (for 1D systems) 
box-like    cutoff (for 0D, 1D and 2D systems)

The Coulomb potential with a box-like cutoff is defined as

Vc1.png

Then the FT component is

Vc2.png

where

Vc3.png

For a 2D-system with non period direction along z-axis we have

Vc4.png

Important remarks:

  • the Random Integration Method (RIM) is required to perform the Q-space integration
  • for sufficiently large supercells a choose L_i slightly smaller than the cell size in the i-direction ensures to avoid interaction between replicas


Creation of the input file:

$ yambo -F yambo_cut2D.in  -r

Open the input file yambo_cut2D.in

Change the variables inside as:

RandQpts= 1000000          # [RIM] Number of random q-points in the BZ
RandGvec= 100        RL    # [RIM] Coulomb interaction RS components
CUTGeo= "box z"            # [CUT] Coulomb Cutoff geometry: box/cylinder/sphere X/Y/Z/XY..
% CUTBox
 0.00     | 0.00     | 32.0    |        # [CUT] [au] Box sides

Close the input file

Run yambo:

$ yambo -F  yambo_cut2D.in  -J 2D

in the directory 2D you will find the two new databases

ndb.RIM		ndb.cutoff