How to use BSE double grid method with diagonalization solver

[jasonhan0710]

Dear Yambo developers,

I have tried to calculate the excitonic structure. I want to use dense grid in BSE calculation in order to converge exciton properties. however I found that the double grid method only works with inversion solver, which cannot give the wavefunction of excitons. Is it possible to implement the double grid method within diagonalization solver? Or are there any feasible ways to get the wavefunctions with inversion solver?

Best,
Jason

[Daniele Varsano]

Dear Jason,
the solution could be to consider large single-grid and solve the BSE using iterative algorithms, the eigenfunction of the lower excitons can be then calculated using the slepc solver.
Best,
Daniele

[jasonhan0710]

Hi Daniele,

Thank you for your reply.
Can I use a small k grid in GW calculation and use a dense k grid in BSE calculation with slepc solver? Will yambo interpolate the QP energy automatically when use different k grid database?

Best,
Jason

[Daniele Varsano]

Dear Jason,

yes, yambo will read the QP of the small grid and interpolate in the larger grid.

Best,
Daniele

[jasonhan0710]

Hi Daniele,

As you suggested before,

the solution could be to consider large single-grid and solve the BSE using iterative algorithms

. What's the difference between double gird method and large single grid method? If I use the double grid method, I may just consider the bands near Fermi surface in order to save memory in calculation.

Best,
Jason

[Daniele Varsano]

Dear jason,

If I use the double grid method, I may just consider the bands near Fermi surface in order to save memory in calculation.

This can be done also considering single grids. The point is that slepc solver can be used for a single grid only (afaik).

Daniele

[jasonhan0710]

Dear Daniele,

I know. Thank you!

Best,
Jason