Resonant, broadband and highly efficient optical frequency conversion in semiconductor nanowire gratings at visible and UV wavelengths

TL;DR

This paper demonstrates that semiconductor nanowire gratings can achieve resonant, broadband, and highly efficient optical frequency conversion at visible and UV wavelengths, leveraging nonlinear resonances and field localization.

Contribution

It reveals that resonant nonlinearities in semiconductor nanowires enable efficient harmonic generation across a broad spectral range, surpassing traditional non-resonant limits.

Findings

Third harmonic conversion efficiency of ~1% in silicon nanowires.

Surface second harmonic conversion efficiency around 0.01%.

Potential to extend silicon photonics into UV and visible ranges.

Abstract

Using a hydrodynamic approach we examine bulk- and surface-induced second and third harmonic generation from semiconductor nanowire gratings having a resonant nonlinearity in the absorption region. We demonstrate resonant, broadband and highly efficient optical frequency conversion: contrary to conventional wisdom, we show that harmonic generation can take full advantage of resonant nonlinearities in a spectral range where nonlinear optical coefficients are boosted well beyond what is achievable in the transparent, long-wavelength, non-resonant regime. Using femtosecond pulses with approximately 500 MW/cm2 peak power density, we predict third harmonic conversion efficiencies of approximately 1% in a silicon nanowire array, at nearly any desired UV or visible wavelength, including the range of negative dielectric constant. We also predict surface second harmonic conversion efficiencies of order 0.01%, depending on the electronic effective mass, bistable behavior of the signals as a result of a reshaped resonance, and the onset fifth order nonlinear effects. These remarkable findings, arising from the combined effects of nonlinear resonance dispersion, field localization, and phase-locking, could significantly extend the operational spectral bandwidth of silicon photonics, and strongly suggest that neither linear absorption nor skin depth should be motivating factors to exclude either semiconductors or metals from the list of useful or practical nonlinear materials in any spectral range.