Low intensity saturation of an ISB transition by a mid-IR quantum cascade laser

TL;DR

We address low-intensity saturation of mid-infrared ISB transitions by engineering a metal–semiconductor–metal patch-cavity array containing GaAs/AlGaAs QWs, operating at the onset of strong light–matter coupling to reduce the saturation intensity. A coupled-mode theory model links incident intensity to the ISB population difference and Rabi frequency, capturing the transition from two polaritons to a single absorbed feature at high power. The experiment demonstrates saturation at 10-20 kW cm$^{-2}$ at room temperature, with reflectivity behavior that depends on whether the pump is at the cavity or polariton frequencies, consistent with SESAM-like operation in the mid-IR. This work establishes a viable path toward mid-IR SESAMs and highlights scalable, fast saturable absorption suitable for future laser technologies.

Abstract

We demonstrate that absorption saturation of a mid-infrared intersubband transition can be engineered to occur at moderate light intensities of the order of 10-20 kW$.$cm$^{-2}$ and at room temperature. The structure consists of an array of metal-semiconductor-metal patches hosting a judiciously designed 253 nm thick GaAs/AlGaAs semiconductor heterostructure. At low incident intensity the structure operates in the strong light-matter coupling regime and exhibits two absorption peaks at wavelengths close to 8.9 $μ$m. Saturation appears as a transition to the weak coupling regime - and therefore to a single-peaked absorption - when increasing the incident intensity. Comparison with a coupled mode theory model explains the data and permits to infer the relevant system parameters. When the pump laser is tuned at the cavity frequency, the reflectivity decreases with increasing incident intensity. When instead the laser is tuned at the polariton frequencies, the reflectivity non-linearly increases with increasing incident intensity. At those wavelengths the system therefore mimics the behavior of a saturable absorption mirror (SESAM) in the mid-IR range, a technology that is currently missing.

Figures (5)

• Figure 1: (a) Colorized SEM image of a typical patch antenna array defining the dimensions $s$ and $p$. (b) Self-consistent Schrödinger-Poisson simulation of a period of the structure showing the two confined electronic states. (c) Sketch of a single patch antenna with the electric field amplitude of the fundamental TM$_{00}$ mode.
• Figure 2: (a) Reflectivity of the bare cavity arrays (purple dashed line) and of the doped cavity arrays (blue solid lines) as a function of the cavity size. The spectra are offset by 0.5. Inset: zoom on the $s=1.35$ µm spectrum around the polariton splitting. (b) Cavity and polaritons dispersion relation as a function of the patch size. The symbols are the experimental data (circles: undoped cavities, squares: doped cavities) extracted from Lorentzian fits of the spectra. The solid lines are fit to eq. \ref{['eq:cavdisprel']} and eq. \ref{['eq:poldisprel']} respectively.
• Figure 3: (a) Reflectivity spectra of the doped $s=1.35$ µm cavity array under low (purple circles) and high (blue circles) intensity excitation. The shaded areas indicate the tuning ranges of the QCL chips. (b) Nonlinear reflectivity as a function of intensity for three different wavelengths (arrows in (a)). Open symbols correspond to experimental data, and solid lines to the CMT prediction. The vertical dashed lines indicate the saturation condition. The horizontal error bars represent uncertainty on the evaluation of the spot size.
• Figure 4: Experimental setup used for the measurement of non-linear reflectivity. $\lambda$/2: halfwave plate. Pol.: KRS5 wire grid polarizer. B.S.: 3mm thick 90/10 ZnSe beam splitter. MCT: Mercury Cadmium Telluride detector.
• Figure 5: (a) Band structure of a single QW with the two relevant conduction band states. (b) Occupation (carrier densities in $10^{11}$ cm$^{-2}$) of each level for three selected temperatures. (c) Normalized population difference as a function of the temperature. (d) Reflectivity as a function of the temperature obtained from the CMT model (main text) using the thermal variation of the population in (c) as an input.