Effective optical constants: A fundamental discrepancy

How effective optical constants of randomly oriented polydomain media are determined by the dielectric tensor functions of the domains is still an open question. On this matter, we show that the basis of the often-applied effective medium approximation (EMA) is not compatible with the linear dichroism theory. Despite the similarity of the approaches there exists a striking discrepancy, which is obvious especially in the large anisotropy limit, while only small differences in the vicinity of resonances emerge if the anisotropy is negligible. The application of several different effective optical constants approaches taking into account either local field corrections or non-principal orientations lead to rather similar results except for very strong anisotropies. Since all of the approaches have a seemingly valid theoretical basis and due to the lack of experimental data it is difficult to decide which approach should be favoured.