The static screening plays an important role in the estimate of the
Coulomb pseudopotential mu*.
The screening was first calculated within the random phase
approximation (RPA) for a model with
a long-ranged Coulomb interaction. The screening was found to be
very efficient within the RPA, and it leads to a large reduction of mu*.
In view of the strong correlation effects in these systems, the accuracy
of RPA may seem very questionable. For this reason, the static screening
has been studied in a lattice Quantum Monte Carlo method for a
Hubbard-like model of A3C60
( Phys. Rev. Lett. , 83 , 620
(1999) ). RPA was found to
be surprisingly accurate for values of the Coulomb interaction U
which correspond to a metal. For values of U close to a Mott
transition the screening becomes less efficient than predicted by RPA and
for a Mott insulator RPA is qualitatively wrong.
Max-Planck Institut für Festkörperforschung
Postfach 800 665
D-70506 Stuttgart