Im$\{\chi^{(3)}\}$ spectra of 110-cut GaAs, GaP, and Si near the two-photon absorption band edge

TL;DR

This study measures the third-order nonlinear optical spectra of GaAs, GaP, and Si near their two-photon absorption edges using femtosecond spectroscopy, revealing detailed spectral features and symmetry breaking effects.

Contribution

It provides the first comprehensive characterization of the imaginary part of the third-order nonlinear susceptibility tensor for these materials near the two-photon absorption edge.

Findings

Peak of $eta()$ in GaAs near 0.78 $E_g$

Dispersion in GaP and Si increases with photon energy

Kleinman symmetry is broken in all three materials

Abstract

Spectra of the degenerate two-photon absorption coefficient $\beta(\omega)$, anisotropy parameter $\sigma(\omega)$, and dichroism parameter $\delta(\omega) = \big[\sigma(\omega) + 2\eta(\omega)\big]/2$ of crystalline 110-cut GaAs, GaP, and Si at 300 K were measured using femtosecond pump-probe modulation spectroscopy over an excitation range in the vicinity of each material's half-band gap $E_g/2$ (overall $0.62 < \hbar \omega < 1.91$ eV, or $2000 > \lambda > 650$ nm). Together these three parameters completely characterize the three independent components of the imaginary part of the degenerate third-order nonlinear optical susceptibility tensor $\mathrm{Im}\{\chi^{(3)}_{abcd}(\omega)\}$. In direct-gap GaAs, these components peak at $\hbar \omega \approx 0.78 E_g$ which is close to the peak at $\hbar \omega = 0.71 E_g$ predicted by the Jones-Reiss phenomenological model. The dispersion is comparable to ab initio calculations. In indirect-gap GaP and Si, these components tend to increase with $\hbar \omega$ over our tuning range. In Si, the dispersion differs significantly from predictions of semi-empirical models and ab initio calculations do not account for transitions below the two-photon direct band gap, motivating further investigation. Kleinman symmetry was observed to be broken in all three materials. We also note anomalies observed and their possible origins, emphasizing the advantages of a 2-beam experiment in identifying the contribution of various nonlinear effects.