Implementation of different RF-chains to drive acousto-optical tunable filters in the framework of an ESA space mission

Abstract

Wide aperture acousto-optical tunable filters (AOTF) appeared in the 70’s and found applications in various areas such as agriculture (crop stress monitoring), food industry (product quality), biology (fluorescence spectroscopy), etc. Their main advantages are the robustness, compactness, low power consumption, high filtering efficiency, tunability, potentially high spectral resolution (<1nm) and good image quality. In general, they offer interesting features to meet the needs in hyperspectral imaging applications. The physical process behind the wide aperture AOTF is the interaction of light and sound inside a birefringent crystal. For a given acoustic frequency, only photons of particular energy (i.e. wavelength) will couple with the acoustic wave, then leave the crystal in a slightly different direction than the rest of the light. By focusing the diverted beam onto a detector, one effectively performs a spectral image of the scene. Selecting another window of the light spectrum only requires the tuning of the sound frequency. A piezoelectric transducer bonded to the crystal is responsible for converting the electrical RF-signal into an acoustic wave. As most AOTFs operating in the visible and near-infrared domains are driven with frequencies ranging a few MHz to several hundreds of MHz, they do not necessitate particular electronics equipment. This statement does not hold when it comes to operating AOTFs in space environment: the RF-driving chain must be made from the limited catalogue of space-qualified parts. The problem gets even bigger when frequencies of several tens of MHz or higher must be generated. The work reported in this paper addresses different solutions to the problem of building a space-qualified RF-chain for an AOTF. This research was undertaken as part of the development of the ALTIUS space mission (atmospheric limb tracker for the investigation of the upcoming stratosphere) which aims at the measurement of atmospheric trace species (ozone, nitrogen dioxide, methane, water vapor,…) concentration profiles with a high spatial resolution. The measurement concept relies on the acquisition of spectral images of the bright atmospheric limb at well-chosen wavelengths. The imager concept allows to discope the need for scanning the atmosphere, as it was done by previous remote sensing missions. The instrument will be mounted onboard a PROBA-satellite (Project for On-Board Autonomy). The PROBA-satellite is a platform containing all the essential subsystems like a GPS, an attitude and orbit control system etc. to facilitate the payload which is placed on top of the platform. The complete project (platform and payload) is developed under the supervision of ESA (European Space Agency) and with funding of the Belgian Science Policy Office (BELSPO). The original ALTIUS concept made use of three independent spectral imagers (channels), each of them relying on an AOTF capable of isolating narrow passbands across the channel spectral range (ultraviolet (UV): from 250 to 400nm, Visible: from 440 to 800nm, near-infrared (NIR): from 900 to 1800nm). For the Visible and NIR channels, the two AOTFs will be made of a paratellurite crystal. For the UV channel, the AOTF would have been made from a KDP crystal (potassium dihydrogen phosphate). Unfortunately, the latter did not reach the necessary level of maturity for a space mission, and the KDP-based AOTF was replaced by a stack of Fabry-Pérot interferometers (FPI). Nevertheless, for technological interest, the AOTF approach will also be developed and matured for the UV-channel up to flight level. The focus of this paper is to design for each of the different channels a dedicated RF-chain, containing a RF-generator and a RF-amplifier. The output of the RF-amplifier is tunable to a specific frequency and a specific power level and is injected in the AOTF’s transducer via an impedance matching network. Different architectures are possible to generate the frequency range needed for the different wavelength domains. While the Hilbert Transform solution will be only summarized here and not further investigated, this paper focuses on the Phase Locked Loop (PLL) solution. PLL-solutions are breadboarded for the three wavelength domains. A detailed study is done on the achievable power levels in the Infrared, Visible and the Ultraviolet.