IR Sensing

Metabolites, gases, and many chemicals are structurally unique and exhibit distinctive absorptive fingerprints in the infrared (IR) portion of the spectrum. The accurate detection and tracking of such spectrum using portable devices carry enormous potential in the field of sensing and would enable many applications, including real-time health monitoring, gas detection, spectroscopy, imaging, and the identification of hazardous materials. IR detection can be performed through photon detection, exploiting the interaction of photons and electrons in semiconductors such as HgCdTe, InGaAs, GaAs/AlGaAs and more recently, in 2D materials; or thermal detection, which relies on the temperate-induced change of the material properties upon exposure to IR light. Photodetectors are usually faster and low noise but may require cryogenic temperatures to operate. Thermal detectors operate at room temperature but exhibit slower response and higher noise. Both approaches are inherently broadband and are usually incorporated into Fourier Transform IR (FTIR) spectrometers to achieve spectral selectivity through interferometry techniques. Unfortunately, this requires moving mirrors to create different optical paths, which makes FTIR bulky, incompatible with CMOS, and unsuitable for real-time monitoring. Despite recent efforts to miniaturize FTIR technology , state-of-the-art on-chip devices still face important challenges in terms of reproducibility, noise, and size.

In this research line, we explore different strategies to achieve spectrally-selective infrared sensing at room temperature. In the main thrust, we developed a building-block of a miniaturized IR detector that was fabriated at the CNM2 at Davis and the Berkeley cleanroom. The device relies on nanopatterning ultrathin and high-Q metasurfaces on top of free-space standing microelectromechanical systems (MEMSs) to efficiently absorb light with desired properties and spectral distribution. The MEMS, designed to achieve a high mechanical quality factor and is excited at resonance by a RF signal whose phase changes proportionally with the absorbed IR power. State of the art NEP ~50pW/Hz^(1/2) have been reported on a bench-top readout. Importantly, each device can target a different IR wavelength (full-width half maximum FWHM ~ 0.15um) by simply adjusting the nanoresonator dimensions. We have fabricated thousands of IR detectors and explore their properties using a home-made automated characterization system. Currently, we are modifying the system to incorporate an OPO tunable thus obtaining wavelength dependent responsivities. 

Additionally, we are exploring the use of this technology is various sensing scenarios, including the detection of gases such as CO2 and NO2. We are also appliying this sensors to detect head and neck cancer of biofluids like blood, saliva, sweat, and urine. To this purpose, we are merging our sensors with an artificial intelligence. Preliminary results, using Raman signals instead of IR data, show very promising performance.