Accuracy of the surface integral-equation formulations for large negative permittivity values

Computational solutions of plasmonic problems involving metals at optical frequencies formulated with surface integral equations are considered. Numerical inaccuracies arise when using the conventional formulations for penetrable bodies, especially as the negative real permittivity becomes very large at the lower frequencies of the optical spectrum. In order to close the gap between plasmonic and perfectly conducting simulations, it is required to extend the applicability of surface integral equations to increasingly large negative permittivity values, which is possible by scaling the matrix blocks correctly to capture the limiting process. We demonstrate an example for such an integral-equation formulation, as well as its accuracy, stability, and efficiency in comparison to the traditional formulations.