Hybrid cavity-antenna architecture for strong and tunable sideband-selective molecular Raman scattering enhancement

Abstract

Plasmon resonances at the surface of metallic antennas allow for extreme enhancement of Raman scattering. Intrinsic to plasmonics, however, is that extreme field confinement lacks precise spectral control, which would hold great promise in shaping the optomechanical interaction between light and molecular vibrations. We demonstrate an experimental platform composed of a plasmonic nanocube-on-mirror antenna coupled to an open, tunable Fabry-Perot microcavity for selective addressing of individual vibrational lines of molecules with strong Raman scattering enhancement. Multiple narrow and intense optical resonances arising from the hybridization of the cavity modes and the plasmonic broad resonance are used to simultaneously enhance the laser pump and the local density of optical states, and are characterized using rigorous modal analysis. The versatile bottom-up fabrication approach permits quantitative comparison with the bare nanocube-on-mirror system, both theoretically and experimentally. This shows that the hybrid system allows for similar SERS enhancement ratios with narrow optical modes, paving the way for dynamical backaction effects in molecular optomechanics.

Fig. 1.. The nanocube in an FP hybrid resonator.

(A) We realize a bottom-up hybrid resonator for SERS that leverages the self-assembled NCoM system for confinement, embedding it in a tunable FP cavity for tunability and sideband resolution. The nanocube is coated with BPT molecules, whose Raman emission enhancement can be tuned by changing the cavity mirror spacing. (B) Schematic representation of the tunable dual SERS enhancement obtained with the hybrid resonator, offering both pump and LDOS enhancement with sideband resolution.

Fig. 2.. Simulation results obtained with COMSOL and quasi-normal modal analysis, comparing the NCoM antenna, the empty FP cavity, and the hybrid resonator metrics.

Analysis performed with the MAN package (32). (A) Magnetic mode (11) profile of the NCoM tuned into resonance by a 6-nm alumina gap, featuring (B) a strong LDOS enhancement with respect to a homogeneous alumina environment (ρ/ρ₀) for a vertical dipole placed below the corner and in the middle of the gap. (C) Normalized radiation pattern of the magnetic (11) mode (red) with a vertical radiation direction as opposed to the usual dipolar (10) mode radiation pattern (light brown) with a mostly tangential radiation. (D) Field enhancement in the empty FP cavity obtained by placing a curved mirror on top of a flat gold mirror [30-nm thickness as in (A) to (C)], illuminated by a 1-μm-waist Gaussian beam. (E) Transmission spectrum of the bare cavity showcasing Q factors greater than 200 for 30-nm-thick gold mirrors. (F) Mode profiles of the NCoM in an FP cavity hybrid. The individual constituents are coupled through their strong H fields on the mirror surface. (G and H) Transmission spectra of the hybridized cavity structure for varying mirror distances. The linewidth and resonance frequency change as a function of the detuning with the NCoM. (I) Total LDOS enhancement for a vertical dipole placed in the middle of the gap below the cube’s corner reaching values up to those obtained with the bare NCoM, but with an additional control on the quality factor and resonance frequency. (J) Crosscut of (I) for the FP cavity and the NCoM at resonance: The total LDOS (dashed) is a sum of contributions of two hybridized quasi-normal modes, whose asymmetric shape arises from complex valued mode volumes (38).

Fig. 3.. Optical characterization of the bare constituents.

(A) Homebuilt confocal Raman and white-light spectroscopy setup working in both reflection and transmission. The sample position is controlled with two sets of piezos allowing cavity and NCoM alignment. The SERS excitation laser is tunable to match the cavity, while the cavity opening tunes the cavity resonance and FSR. (B) Transmission electron microscopy image of a typical nanocube and dark-field image of NCoM structures. (C) Extinction spectrum of a bare NCoM whose resonance frequency is tuned to 820 nm using a 2-nm alumina gap. The LDOS enhancement spectrum simulated for a dipole under the cube corner is shown for comparison (dashed). (D) SERS spectrum (black line) taken on the same NCoM displays a broad background from electronic Raman of the gold with a spectral shape similar to the extinction spectrum (red). The distinct peaks above it are characteristic Raman lines of BPT molecules that were previously coated on the nanocube. (E) Atomic force micrograph of the curved mirror fabricated using CO₂ laser ablation, which results in curved and smooth ⁓300-nm-deep concave features with a radius of curvature of about 9 μm. The inset shows a camera image of the bare cavity in transmission (scale bar, 10 μm). (F) Measured transmission spectra of a bare cavity for a saw-tooth displacement of the z-piezo that controls the cavity opening, which allows tuning the cavity modes resonance. (G) Transmission spectrum for a fixed mirror distance of 1.5 μm. A cavity resonance with Q = 130 at the m = 3 longitudinal mode order is extracted from a Lorentzian fit.

Fig. 4.. Signature of hybridization and selective LDOS enhancement.

(A) Cavity transmission spectra when laterally displacing a nanoparticle from the outside into the cavity space. (B) Crosscuts comparing the bare cavity transmission (blue) with the hybridized cavity transmission [orange curve, corresponding to the dashed line in (A)]. (C) Variations in FSR and linewidth of the m = 4 cavity resonance [blue box in (A)] as a function of cube position, showing how the nanoparticle induces a dispersive and dissipative shift of the cavity resonance. (D) Raman spectra of the same hybrid for varying mirror distances. This allows selectively enhancing single Raman peaks of the BPT molecules. (E) Crosscuts corresponding to the selective enhancement of three main BPT Raman peaks indicated by circles in (D).

Fig. 5.. Comparing NCoM and hybrid performances.

(A) NCoM extinction spectrum and (B) corresponding SERS signal. (C) Hybrid resonator transmission spectra with three distinct cavity opening lengths. (D) Corresponding SERS spectra with a distinct Raman line enhanced in the three different panels. Both the pump field and the LDOS for Raman emission are enhanced in each of the cases, as we match the cavity FSR to the Raman shift by adjusting the cavity opening. The pump light matches a first resonance, while strong Raman scattering is only obtained for molecular vibrational lines that are specifically tuned to the second cavity resonance.