ABINIT Discussion Forums

Hi all,

Does Abinit have the capability of running ab initio calculation of optical spectra of liquids using many body effects in the electronic excitations of liquids (BSE) ?
If not, does any one know of any other program which potentially is able to tackle this issue ???

Bests
Vito

There's no limitation in the code that prevents you from computing BSE spectra for liquids with Abinit.
Note, however, that Abinit is a planewave-based code and therefore it does not represent the most efficient
approach for the analysis of non-periodic systems. You have to simulate the liquid with a supercell
of appropriate size --> many G vectors in the basis set --> a *lot* of memory and a significant increase in the CPU time.

So the main questions are

* Do you have enough computing capabilities available? For example, can you run a "simple" GS calculation for your system?

* If the system is too expensive even at the GS level, I suggest looking for other approaches based on a different basis set.
bigdft is a interesting alternative for these kind of systems but, as far as I know, BSE is not coded.
Octopus (real space grid) can compute spectra within TDDFT that is much less CPU and memory demanding than BSE.
All the BSE codes I'm aware of are either all-electron or are based on pseudopotentials and planewaves.
Perhaps there are other alternatives but I didn't make any extensive search on the internet.

TDDFT gives unfortunately very poor result for absorption spectrum of a very simple system like water.
I have enough computing capabilities but not enough for BSE liquid calculations.
Any way i have to forget it.

Thanks