Bridging mid and near infrared by combining optomechanics and self mixing

This work describes a self-mixing-assisted optomechanical platform for transferring information between near- and mid-infrared radiation. In particular, the self-mixing signal of a mid-infrared quantum cascade laser is used to detect the oscillation of a membrane driven by light-induced forces exerted by a near-infrared excitation beam, which is amplitude-modulated at the membrane resonance frequency. This technique benefits from spectral broadness and, therefore, can link different spectral regions from both the excitation and probe sides. This versatility can pave the way for future applications of this self-mixing-assisted optomechanical platform in communication and advanced sensing systems.

💡 Research Summary

This paper introduces a self‑mixing‑assisted optomechanical platform that bridges near‑infrared (NIR) and mid‑infrared (MIR) spectral regions, enabling information transfer between two widely separated wavelengths. The core of the system is a quantum cascade laser (QCL) operating at 4.5 µm, which serves simultaneously as light source and detector through self‑mixing (SM) interferometry. A thin silicon‑nitride trampoline membrane, coated with a Cr/Au metal layer, is placed inside a vacuum chamber (≈10⁻³ mbar) and mounted on a piezoelectric actuator (PZT) for calibration.

Two experimental configurations are explored. In the first (Conf. 1), the membrane is driven mechanically by the PZT while the QCL probe power is varied. The authors observe a linear red‑shift of the membrane’s resonance frequency (≈90 kHz) with increasing QCL power, at a rate of 14 Hz mW⁻¹, attributed to photothermal heating of the metal coating. When a 1064 nm Nd:YAG laser is added without modulation, the resonance shifts more strongly (≈99 Hz mW⁻¹) because the NIR beam is absorbed more efficiently, further relaxing the membrane tension.

In the second configuration (Conf. 2), the NIR beam is amplitude‑modulated (AM) via an acousto‑optic modulator (AOM) and directly drives the membrane through light‑induced forces (radiation pressure and photothermal effects). The QCL probe remains at a fixed power of 6.2 mW. By sweeping the AM frequency across the mechanical resonance and varying the modulation depth, the authors demonstrate that the membrane’s resonance frequency shifts linearly with the average NIR power at a rate of 102 Hz mW⁻¹. The SM signal extracted from the QCL shows a clear resonance peak, confirming that the optomechanical motion induced solely by the NIR beam is faithfully transduced into the MIR channel.

A supplementary experiment applies a slow AM modulation (≤40 Hz) to the NIR beam while the membrane is still PZT‑driven. Two distinct peaks appear in the SM spectrum: one corresponding to the pure mechanical resonance and a second, slightly lower‑frequency peak arising from the photothermal contribution. The two peaks merge above ≈40 Hz, indicating that the thermal relaxation bandwidth is well below the mechanical resonance frequency, and that the photothermal force is heavily damped at the ≈90 kHz timescale.

The work highlights several key advantages. First, the QCL’s intrinsic sensitivity to optical feedback enables high‑speed, label‑free detection without external photodiodes; voltage variations at the laser terminals can directly encode mechanical motion. Second, the excitation wavelength is unrestricted—any NIR source can be used—making the platform spectrally agnostic on the pump side while the probe remains in the MIR. Third, the coexistence of radiation pressure and photothermal forces provides flexibility: depending on material choice and beam parameters, either mechanism can dominate, allowing tailored actuation schemes. Fourth, operation in vacuum reduces environmental noise, improving measurement stability.

Potential applications include broadband infrared communication links (using the membrane as a wavelength‑conversion gate), multi‑spectral sensing (e.g., simultaneous detection of gases that absorb in different IR windows), non‑contact metrology of temperature or stress via the photothermal shift, and integration into quantum photonic circuits where fast feedback control is required. The authors suggest future work to quantitatively separate radiation‑pressure and photothermal contributions, to improve signal‑to‑noise by employing low‑noise QCL drivers and high‑bandwidth voltage readout, and to explore other spectral extensions such as THz or visible regimes using appropriate self‑mixing lasers. Overall, the study demonstrates a versatile, scalable route to link disparate infrared bands through a compact, self‑mixing‑enabled optomechanical interface.