Fig. 1. Characterization of single NCoM plasmonic nanocavities. (a) Scanning electron microscopy image of a single-crystal gold flake (covered with a 2.5-nm-thick alumina layer) with gold nanocubes deposited onto it (as shown in the inset) to form NCoM plasmonic nanocavities. (b) Cross-sectional TEM image of an NCoM plasmonic nanocavity with a gap consisting of a 1.8-nm-thick CTAB bilayer and a 2.5-nm-thick alumina layer. Note that the CTAB layer capping the gold nanocube cannot be observed because it was destroyed during the preparation of the cross-sectional lamella. The blue dashed lines indicate the outlines of the alumina layer. (c), (d) Dark-field scattering images of NCoM plasmonic nanocavities in air taken under TE (c) and TM (d) excitation conditions. (e) Experimentally measured scattering spectra of the nanocavity encircled in (c) and (d) under TE (red line) and TM (black line) excitation conditions. (f) Measured scattering spectra of 50 nanocavities formed on the same gold flake under unpolarized light illumination. The solid dot, triangle, star, and rhombus represent the resonance wavelengths of modes M1,M2,V1, and V2, respectively, marked in (e). Inor represents the normalized scattering intensity.

Fig. 2. Effect of PMMA layer thickness on plasmonic resonances of NCoM nanocavities. (a) Schematical illustration of precise control of the thickness of the PMMA layer surrounding an NCoM plasmonic nanocavity via O2 plasma etching. (b) Measured thickness-etching time correlation curve of the PMMA layer. (c), (d) Measured dark-field scattering spectra of an individual NCoM plasmonic nanocavity embedded in a PMMA layer with a varying thickness under TE (c) and TM (d) excitation conditions. Lower panels show the corresponding dark-field scattering images of the nanocavity for various thicknesses of the PMMA layer.

Fig. 3. Statistical analysis of plasmonic resonances of NCoM nanocavity embedded in PMMA layer with varied thickness. (a), (b) Numerically calculated scattering spectra of an NCoM plasmonic nanocavity embedded in a PMMA layer with a varied thickness under TE (a) and TM (b) excitation conditions. For better readability, the scattering intensities are presented in a logarithmic scale. The dashed lines indicate the resonance wavelengths of modes M1 (red lines), V1 (orange lines), V2 (blue lines), and M2 (purple lines). (c) Resonance wavelengths for the plasmonic modes obtained from experiments (solid dots) and simulations (dashed lines) for various thicknesses of the PMMA layer. (d), (e) Peak intensity ratio for modes M1 and M2 under TE excitation (d) and modes V1 and V2 under TM excitation (e) derived from the experimental measurements (solid dots) and simulations (dashed lines) for various thicknesses of the PMMA layer.

Fig. 4. Numerical simulations of excitation and radiation efficiencies of NCoM nanocavities. (a), (b) Numerically simulated near-field distributions of an electric field in the yz-plane corresponding to nanocavity modes M1 for the nanocavity without (a) and with the PMMA layer (b). (c) Numerically simulated radiation efficiency for modes M1 (red line), M2 (purple line), V1 (orange line), and V2 (blue line) obtained using an eigen-frequency solver, for nanocavities with varied thickness of the PMMA layer. (d)–(i) Numerically simulated near-field distributions of an electric field in the yz-plane corresponding to nanocavity modes M2 (d), (e), V1 (f), (g) and V2 (h), (i) for the nanocavity with varied thickness of the PMMA layer.

Fig. 5. Procedure for fabrication of dielectric-embedded NCoM plasmonic nanocavities. ALD, atomic layer deposition.

Fig. 6. (a) Dark-field spectroscopy setup. (b) Optical micrograph of spectral collection region on the object plane.

Fig. 7. Schematic illustration of the model used for numerical simulations of the scattering spectra, surface charge density distributions, and eigen-frequency of nanocube-on-mirror plasmonic nanocavites.

Fig. 8. (a) Numerically simulated scattering spectra under TM- or TE-polarized excitation (note that the TE signal was magnified by five times). (b)–(e) Normalized charge density distributions on the surfaces of the gold nanocube and flake corresponding to the plasmonic modes (b) M1, (c) M2, (d) V1, and (e) V2 labeled in (a).

Fig. 9. (a) Optical micrograph of a PMMA layer with an initial thickness of ∼96nm, which was gently scored by a surgical blade after spin-coating. The red dashed lines represent the scanning path for measuring the step height with a stylus profilometer. (b) Example of measured step height profile corresponding to the scan path in (a). The final step height is obtained by averaging the five measured heights.

Fig. 10. (a) Dark-field scattering images of NCoM nanocavities with or without PMMA layer taken under TM excitation condition for multi-round spin-coating and etching, showing consistent relative positions. (b) Scattering spectra from an NCoM nanocavity measured before spin-coating of PMMA layer (black line), after etching for 160 s (the PMMA layer has been completely removed, red line) and 320 s (blue line) under TM excitation.