Automatic Characterization of High-Performance MEMS-Based IR Sensors

The next generation of infrared (IR) sensors may enable unprecedented applications in fields like spectroscopy, health monitoring, and communication systems. For instance, metasurface-enhanced micro-electromechanical system (MEMS)-based IR sensors have demonstrated excellent performance in terms of responsivity and spectral-selectivity. However, it is burdensome to experimentally determine the performance limits of this and other IR sensing technologies as it requires time-consuming and expensive systematic testing not always feasible in research settings. To address this challenge, an automated solution for characterizing miniaturized sensing devices is presented in this paper and applied to experimentally determine the performance limits of MEMS-based IR sensors. The system offers low-cost, rapid, and automated characterization of on-chip IR sensors, determining key performance metrics such as noise, responsivity, and noise-equivalent power. The platform is flexible, easily adapted to different types of devices, chip layouts, and light sources, and is designed to test a large number of sensors within the same wafer −spending ~20 seconds per device− using a combination of optical and radiofrequency interrogation techniques. The system has been applied to test over 1500 MEMS-based IR sensors. Collected data revealed hidden trade-offs between responsivity and noise with respect to the device thickness and allowed a statistical analysis of sensing response versus device geometrical dimensions. The best-performing devices exhibit a quality factor, responsivity, fluctuation induced noise, and noise-equivalent power of 2391, 164 Hz/nW, 0.257 Hz/ $\sqrt {\mathbf {Hz}}$ , and 5.01 pW/ $\sqrt {\mathbf {Hz}}$ respectively. The proposed automated platform provides an efficient and cost-effective solution for characterizing the next generation of IR sensing devices.