Modeling nano-optics: towards a more physical formulation of the coupled dipole method

Summary form only given. It is often argued that the local-field correction inferred from the Clausius-Mossotti polarizability (which is the long wavelength limit of any of the usual forms of polarizability used in the CDM) is exact in the static limit. This only true in the bulk. In the CDM one should not neglect the local-field correction due to the geometry of the environment of the dipole. This neglect leads to an incorrect definition of the local-field factor and hence a wrong macroscopic field. We show how to obtain for some geometries a form of the CDM that accounts for the environment of each dipole and leads to a formulation of the CDM that is exact in the long wavelength limit.