Mathematical analysis of plasmonic nanoparticles: the scalar case (Q524297)

The Helmholtz type equation \[ \nabla^{\top}\mu^{-1}\nabla u+\omega^{2}\epsilon u=0\text{ in }\mathbb{R}^{3}, \] where the difference \(u-u_{i}\), with \(u_{i}\left(x\right):=\exp\left(\mathrm{i} \omega\sqrt{\epsilon_{m}\mu_{m}}\,d^{\top}x\right)\) a prescribed incident plane wave in direction \(d\in\mathbb{R}^{3}\), \(\left|d\right|=1\), is assumed to satisfy a Sommerfeld radiation condition, is under consideration. Here \(\left(\epsilon,\mu\right)\) is a pair of piece-wise constant parameters with generically two pairs of complex scalar values \(\left(\epsilon_{c},\mu_{c}\right), \left(\epsilon_{m},\mu_{m}\right)\). The functions \(\epsilon,\mu\) assume the values with index \(c\) in a bounded region \(D\) composed of \(L\in\mathbb{N}\) shifted copies of a bounded, open, connected set \(\tilde{D}\). Main focus of the paper is on the low frequency asymptotics approached via a boundary integral equations re-formulation in terms of a single layer and an adjoint double layer potential ansatz. The problem is suggested as a mathematical model for plasmonic nanoparticles represented by the region \(D\) in the presence of a time-harmonic electro-magnetic field with, for simplicity, \(E=\left(0,0,E_{3}\right)\), i.e. in the transverse magnetic case (or, by relabeling, in the transverse electric case). Slightly simplified we have as assumptions for the parameters describing the material properties \(\left(\epsilon_{m},\mu_{m}\right)\in \;]0,\infty[\), \(\mu_{c}<0\) and \(\epsilon_{c}\in\mathbb{C}\) such that \(\Im\sqrt{-\epsilon_{c}}>0.\) Plasmonic resonances are here defined in mathematical terms as asymptotic phenomena in the low-frequency (by scaling: small particle size) regime of eigenvalues of a boundary integral operator. By analyzing the imaginary part of the associated Green function \(\Gamma\left(x,x_{0}\right)\) as a perturbation of the free-space Green function \(G\left(x,x_{0}\right)=-\frac{\exp\left(\mathrm{i} \omega\sqrt{\epsilon_{m}\mu_{m}}\left|x-x_{0}\right|\right)}{4\pi\left|x-x_{0}\right|}\) the possibility to achieve so-called super-resolution and super-focusing is shown. In conclusion, the generalization to the full Maxwell equations as well as a discussion of the optimization of the allocation of nano-particles is announced as future work.