Giant non-linear susceptibility of hydrogenic donors in silicon and germanium

TL;DR

This paper develops an implicit summation technique for hydrogenic impurities in multi-valley semiconductors, predicting giant non-linear susceptibilities in silicon and germanium that surpass existing materials for third-harmonic generation at terahertz frequencies.

Contribution

It introduces a new method for calculating non-linear susceptibility in multi-valley semiconductors and predicts exceptionally high hyperpolarizability in silicon and germanium.

Findings

Predicted hyperpolarizability of Si:P varies significantly with frequency.

Third-order susceptibility exceeds that of bulk InSb.

Susceptibility times thickness surpasses that of graphene and quantum wells.

Abstract

Implicit summation is a technique for the conversion of sums over intermediate states in multiphoton absorption and the high-order susceptibility in hydrogen into simple integrals. Here, we derive the equivalent technique for hydrogenic impurities in multi-valley semiconductors. While the absorption has useful applications, it is primarily a loss process; conversely, the non-linear susceptibility is a crucial parameter for active photonic devices. For Si:P, we predict the hyperpolarizability ranges from $\chi^{(3)}/n_{\text{3D}}=2.9 $ to $580 \times 10^{-38}$ $\text{m}^5/\text{V}^2$ depending on the frequency, even while avoiding resonance. Using samples of a reasonable density, $n_{\text{3D}}$, and thickness, $L$, to produce third-harmonic generation at 9 THz, a frequency that is difficult to produce with existing solid-state sources, we predict that $\chi^{(3)}$ should exceed that of bulk InSb and $\chi^{(3)}L$ should exceed that of graphene and resonantly enhanced quantum wells.