Leveraging Plasmonic Nanocavity Arrays Forming Metasurfaces to Boost Second Harmonic Generation due to Surface Effects

Plasmonic metasurfaces have emerged as a promising platform for enhancing a range of nonlinear optical processes, offering compact geometry and flexibility in light manipulation. Second order nonlinear processes, like second harmonic generation (SHG), typically require non-centrosymmetric crystals to be realized. Here, we experimentally demonstrate enhanced SHG response by using a gold nanocavity array forming a plasmonic metasurface absorber where titanium dioxide (TiO2), a centrosymmetric dielectric material, with subwavelength thickness is deposited in the realized nanogaps. While such dielectric material has an extremely low second order nonlinear susceptibility, we observe 105-fold boosting in the nonlinear SHG process mainly due to the surface nonlinear susceptibility of the gold metal aided by the significant electric field enhancement that occurs in the nanogaps due to the formed nanocavity resonance. The experimental results obtained are theoretically explained with extensive and rigorous nonlinear simulations that consider all the bulk and surface linear and nonlinear material properties. The presented robust harmonic generation from an ultrathin plasmonic metasurface can be used in nonlinear and quantum integrated photonic applications.

💡 Research Summary

In this work the authors demonstrate a dramatic enhancement of second‑harmonic generation (SHG) using a plasmonic metasurface composed of a gold (Au) nanostripe array separated from an Au mirror by an ultrathin (4 nm) TiO₂ layer. The TiO₂ film is centrosymmetric and possesses an almost negligible bulk χ^(2), yet the device exhibits a 10⁵‑fold increase in SHG intensity compared with a plain Au film. The key to this performance lies in the strong electric‑field confinement that occurs inside the nanogap when the structure is resonantly excited at the fundamental wavelength. The resonant plasmonic mode concentrates the field primarily along the out‑of‑plane (z) direction, which strongly drives the nonlinear surface polarization at the Au‑TiO₂ interfaces.

The metasurface geometry (nanostripe width, height, and period) was optimized via full‑wave electromagnetic simulations to maximize the local field enhancement while achieving near‑perfect absorption at the pump wavelength. The optimized design yields a reflected SHG signal that can be measured directly without the need for near‑field probes, a clear advantage over isolated nanoparticle configurations.

To interpret the experimental results, the authors performed rigorous nonlinear finite‑element simulations in COMSOL. They incorporated the surface nonlinear susceptibility of Au using experimentally measured Rudnick‑Stern parameters (γ, β, δ) and treated the bulk χ^(2) of both Au and TiO₂ as negligible because the materials are centrosymmetric and the electric‑field divergence inside them is essentially zero. The nonlinear wave equation derived from Maxwell’s equations was solved under the undepleted‑pump approximation, with the incident femtosecond pulse intensity set to the experimental peak power (~10⁸ W cm⁻²). The simulated SHG spectra and absolute conversion efficiencies match the measured data quantitatively, confirming that the dominant contribution originates from surface‑induced nonlinear currents rather than bulk effects.

Additional linear optical characterization shows that the metasurface absorbs almost all incident light at the fundamental frequency and also exhibits significant absorption at the second‑harmonic wavelength, which helps suppress re‑absorption of the generated SHG signal and further boosts the observable output.

Overall, the study provides three major insights: (1) Surface nonlinearities of noble metals, when strongly enhanced by nanogap field confinement, can dominate SHG even when the dielectric filling is centrosymmetric and intrinsically weakly nonlinear; (2) Periodic metasurface arrays enable robust, far‑field detectable SHG signals, overcoming the weak scattering limitations of single nanocavities or nanoparticles; and (3) The close agreement between theory and experiment validates a design workflow that combines electromagnetic optimization with accurate surface‑nonlinearity modeling, paving the way for ultrathin, efficient nonlinear photonic components. Potential applications include on‑chip frequency conversion, nonlinear sensing, and integration into quantum photonic circuits where compact, high‑efficiency nonlinear elements are essential.