A new snapshot interferometric imaging spectrometer: a first comparison with a classical grating spectrometer.

Atmospheric gas monitoring is of major importance for climate change and air quality. Indeed, emissions regulations and control rely on the detection and quantification of the concentration of gases such as CO₂, CH₄, NO₂, O₃, etc. Good control on emissions is key to reduce those gases impacts on climate change and people’s health.

The accuracy and relevance of such measurements depend on higher spatial, spectral and temporal resolutions. To this end, conventional dispersive hyperspectral imaging systems are typically used. However, these sensors are submitted to compromises in terms of price, spectral and spatial resolutions and temporal acquisition frequency. To overcome these compromises, a new ground-breaking device is currently developed under the name Imaging Spectrometer on Chip (ImSPOC). It is based on an interferometric imaging system that allows real time acquisition with significant spatial and spectral resolutions. The device, which takes the volume of a matches’ box, could in the future be a building block for Nano-satellites, drone, or ground based measurements platforms. The particularity of this device is the snapshot acquisition of an interferometer by pixel of the imaged scene instead of a spectrum. This is obtained by using a matrix of Fabry-Perot interferometers with different thicknesses placed in front of a photodetector. ImSPOC is then of great interest for real-time acquisitions. However, the acquisition of interferometers requires signal processing developments to reconstruct the corresponding spectra. This reconstruction relies typically on the resolution of an inverse problem. Some models of the device have been proposed to this end.

To validate the efficiency of this new device and to test the developed algorithms, acquisitions were conducted tracking the sun during a whole day using our device and a conventional diffraction grating based spectrometer. In this way, the reconstructed spectra from our device can be compared to the classical spectrometer ones. Particularly, the absorption peaks are compared (their central wavelengths, amplitude, etc.). To go further, the gas characterization from both devices will be compared (gas detection, evolution over time of the vertical concentration profiles, etc.). These results allow the validation of our device to this application and highlight the signal processing improvements that could be done in the future to have more accurate measurements.