LCPQ
Salle de séminaire IRSAMC
Abstract :
I will present my work within the frameworks of time-dependent
current-density-functional theory (TDCDFT) and many-body perturbation
theory (MBPT).
TDCDFT is a generalization of time-dependent density-functional theory
(TDDFT) in which the key quantity is the electron current density [1].
Using a local-current-density approximation for the exchange-correlation
potential we were able to resolve several problems of standard TDDFT,
which uses a local-density approximation, at no extra computational cost.
For example, we were able to solve the longstanding problem of the
overestimation of the static polarizabilities of conjugated polymers.
TD(C)DFT can describe neutral excitations but not excitations in which the
total number of electrons change. Therefore, in order to calculate
electron addition and removal energies we use MBPT within the GW
approximation which is an accurate method that correctly describes the
screened electron interaction in many-electron systems [2]. Unfortunately
this method is computationally demanding which is mainly due to the large
number of empty states that have to be taken into account in the standard
spectral representations for the self-energy which is the quantity within
MBPT that contains all the many-body effects. I will demonstrate how one
can reformulate the GW method such that it depends on occupied states
only. Our method is general, easy to implement and leads to an immediate
speedup of calculations.
[1] G. Vignale & W. Kohn, Phys. Rev. Lett.77, 2037 (1996)
[2] L. Hedin, Phys. Rev. 139, A796 (1965)