Study of second and third harmonic generation from an indium tin oxide nanolayer: influence of nonlocal effects and hot electrons

TL;DR

This study investigates second and third harmonic generation in indium tin oxide nanolayers near the epsilon-near-zero point, combining experiments and a comprehensive hydrodynamic model to understand the physical mechanisms involved.

Contribution

The paper introduces a detailed first-principles hydrodynamic model that captures nonlocal effects, hot electron dynamics, and nonlinear interactions, aligning well with experimental observations.

Findings

Enhanced second harmonic efficiency near epsilon-near-zero wavelength

Unusual third harmonic behavior at oblique incidence

Model accurately predicts experimental results

Abstract

We report comparative experimental and theoretical studies of second and third harmonic generation from a 20nm-thick indium tin oxide layer in proximity of the epsilon-near-zero condition. Using a tunable OPA laser we record both spectral and angular dependence of the generated harmonic signals close to this particular point. In addition to the enhancement of the second harmonic efficiency close to the epsilon-near-zero wavelength, at oblique incidence third harmonic generation displays unusual behavior, predicted but not observed before. We implement a comprehensive, first-principles hydrodynamic approach able to simulate our experimental conditions. The model is unique, flexible, and able to capture all major physical mechanisms that drive the electrodynamic behavior of conductive oxide layers: nonlocal effects, which blueshift the epsilon-near-zero resonance by tens of nanometers; plasma frequency redshift due to variations of the effective mass of hot carriers; charge density distribution inside the layer, which determines nonlinear surface and magnetic interactions; and the nonlinearity of the background medium triggered by bound electrons. We show that by taking these contributions into account our theoretical predictions are in very good qualitative and quantitative agreement with our experimental results. We show that by taking these contributions into account our theoretical predictions are in very good qualitative and quantitative agreement with our experimental results. We expect that our results can be extended to other geometries where ENZ nonlinearity plays an important role.