Plasmonic lattice Kerker effect in ultraviolet-visible spectral range

Mostly forsaken, but revived after the emergence of all-dielectric nanophotonics, the Kerker effect can be observed in a variety of nanostructures from high-index constituents with strong electric and magnetic Mie resonances. A necessary requirement for the existence of a magnetic response limits the use of generally nonmagnetic conventional plasmonic nanostructures for the Kerker effect. In spite of this, we demonstrate here the emergence of the lattice Kerker effect in regular plasmonic Al nanostructures. Collective lattice oscillations emerging from the delicate interplay between Rayleigh anomalies and localized surface plasmon resonances both of electric and magnetic dipoles, and electric and magnetic quadrupoles result in suppression of the backscattering in a broad spectral range. Variation of geometrical parameters of Al arrays allows for tailoring the lattice Kerker effect throughout UV and visible wavelength ranges, which is close to impossible to achieve using other plasmonic or all-dielectric materials. It is argued that our results set the ground for wide ramifications in the plasmonics and further application of the Kerker effect.

Figure 1

(a),(f) Reflectance spectra, and (b)–(e), (g)–(j) multipole (ED, MQ, MD, EQ) decomposition of extinction efficiency for arrays with fixed $h_x=240$ nm and different $h_y$ (top), and with fixed $h_y=280$ nm and different $h_x$ (bottom). Al NPs with radius $R=60$ nm have been considered in all cases. Notice a suppression of reflection which follows [0;1] RA, ${\lambda }_{\mathrm{RA}}=\sqrt{{\varepsilon }_h}{h}_y$ [see Fig. 2 for details]. The order of the multipoles appearance is chosen for the sake of consistency with Eq. (5). Tabulated values of Al permittivity from Ref. [48] are used in simulations.

Figure 2

(a) Reflectance for arrays with different geometrical parameters $(R,h_x,h_y)$ as marked in the legend. Vertical dashed lines show respective spectral positions of $[1;0]$ and $[0;1]$ RAs for each array. (b) Multipole decomposition of the phase of the reflection amplitude $\varphi =\arg \left(E_{\mathrm{ref}}\right)$ Eq. (2)] for NPs array with $(60,240,280)$ nm [cf. orange line in (a)]. Notice the $\mathrm{\Delta }\varphi =\pi$ phase difference between EQ and MD contributions on one hand, and ED and MQ counterparts on the other hand, at $\lambda =\sqrt{{\varepsilon }_h}{h}_y$, the wavelength of the lattice Kerker effect. (c) Reflectance for arrays with different NP radii $R$ and fixed $h_x=240$ nm, $h_y=280$ nm. (d) Electric field distribution in the $ZY$ plane for $(60,240,280)$ nm array at $\lambda =417.6$ nm (top, maximum reflectance) and $\lambda =420.5$ nm (bottom, zero reflectance). Notice rapidly vanishing reflected and transmitted fields at $\lambda =420.5$ nm in the far zone due to strong localization of the electric field in the vicinity of NP, and high absorption shown in Fig. 3. The dynamics of the electromagnetic field propagation at both wavelengths is demonstrated in the Supplemental Material [66].

Figure 3

Absorption spectra for arrays with the same parameters as in Fig. 1, i.e., for varying (a) $h_y$ and (b) $h_x$, and (c) for array with $h_x=240$ nm and $h_y=280$ nm. Dashed horizontal lines indicate the lattice periods at which the absorption spectral line $\lambda =420.5$ nm (c) is observed.