|
|
1.INTRODUCTIONSpace-based Earth observation missions, using satellites to monitor Earth’s physical and chemical environment, have proven to be a fast, accurate and cost-efficient way to gather numerous and various data in order to enhance meteorological knowledge, weather forecast and climate change understanding. Each specific mission requests a unique payload where infrared sensors are embedded in panchromatic (black and white imagery), multispectral (acquisition of multiple images of a scene simultaneously at specific spectral bands), hyperspectral (acquisition of multiple images of a scene simultaneously in hundreds of continuous narrow spectral bands) or spectrometer instruments. These infrared sensors can operate in SWIR, MWIR, LWIR or VLWIR spectral bands, the former band being usually the preferred choice to provide the strong contrast needed for high resolution imaging solution or to measure pollution levels, greenhouse gases and aerosols in the atmosphere. To address this SWIR Earth observation market segment, LYNRED developed in the early 2000 a first Infrared Focal Plan Array (IRFPA) detector named SATURN (1000x256, 30μm pixel pitch) that has been successfully used in numerous space missions like TROPOMI, PRISMA or HYSIS missions. A mid-format detector named NEPTUNE (500x256, 30μm pixel pitch) has also been derived and proposed in missions like Hayabusa, Chandrayaan or Spirale missions. Technical requirements for these detectors evolved over the years and LYNRED started in 2011 the development of a new IRFPA named NGP (1024x1024, 15μm pixel pitch). This detector, validated in harsh space environmental conditions, is still in production and has been selected up to now for the Copernicus Sentinel-5, MicroCarb and CO2M missions. A detailed presentation of this component, including measured electro-optical performances, is proposed in this paper. As the trend towards higher performances detectors continues, a new IRFPA development has been initiated in 2020. This large format detector named COBRA comes in two versions (COBRA-L = 1840x1112, 20μm pixel pitch and COBRA-S = 1380x640, 20μm pixel pitch) and perfectly matches the future needs of components featuring better spatial and spectral resolution as well as improved radiometric performances. More information about this detector are detailed in the next paragraphs. Finally, different packages adapted to NGP and COBRA detectors are presented in the last paragraph. These products are proposed in passive version, an architecture composed of an IRFPA integrated on a baseplate with a detection cold wiring, or in active version, an architecture where the IRFPA is housed in a Dewar combined with a long-life low microvibrations pulse-tube cooler. 2.FROM SWIR EARTH OBSERVATION APPLICATIONS TO NGP AND COBRA DETECTORS2.1Two main kinds of applicationsEarth observation applications can be splitted in two different categories: atmosphere and ground observation. The main objective of atmosphere observation is to identify the structure of the chemical and aerosol composition of Earth’s atmosphere. Most of the optical instruments operating in the SWIR spectral band and answering theses needs are based on a spectroscopy concept in order to analyze the atmosphere composition through infrared radiations that are absorbed by the different species. This kind of applications requests detectors that are large enough to cover the swath of the instrument and with radiometric performances compatible with a low level of infrared signals, an intrinsic characteristic of this type of applications. The TROPOMI instrument on-board Sentinel 5 precursor satellite as well as Sentinel 5 or more recently CO2M mission in the frame of the European Copernicus program are good examples of atmosphere observation systems. Regarding ground observation, the main objective is to provide imaging services for a range of applications in agriculture, forestry or assessment of coastal zones evolution to name just a few. Two sub-categories can be identified according to the number of spectral bands that are used in the optical instrument: multispectral or superspectral applications and hyperspectral ones. While multispectral or superspectral applications correspond to space observation of Earth in a few number of spectral bands like in Sentinel 2 mission, hyperspectral applications correspond to Earth observation from space in hundreds of continuous spectral bands inside the overall SWIR spectral range. PRISMA or HYSIS missions are typical examples of SWIR hyperspectral missions. 2.2Key requirements for SWIR Earth observation missionsHyperspectral instruments provide images of the observed scene with a high number of continuous spectral channels and with a high spectral resolution (typically 10 to 15 nm) in the considered waveband. Thus, detectors used in these instruments are two-dimensional arrays with an adapted spectral response to fit the instrument waveband. Specific functions concerning the control of the detector are required. Among them, one can mention:
LYNRED developed its SATURN detector to satisfy these requirements and has acquired thanks to it a large experience in the field of SWIR detectors for hyperspectral applications. Spectro-imagery instruments used for Earth atmosphere observation provide spectrograms of the observed scene with a typical spatial resolution of some kilometers. As an example, the CO2M mission contains an infrared spectrometer enabling to measure the quantity of CO2 in the atmosphere with a ground resolution of 4 km2. Here also, the detectors that are used in these instruments are two-dimensional arrays with an adapted spectral response to fit the waveband of interest depending on the species that are observed. Some specific characteristics and performances are required at detector level to satisfy these applications:
For this second category of applications, the SATURN detector has also been widely used like in TROPOMI instrument on Sentinel 5 precursor satellite or in 3MI instrument on Sentinel 5 satellite. 2.3Next generation SWIR staring arrays for Earth observationAnticipating the evolution of future SWIR Earth observation missions, LYNRED started the development of a new generation of SWIR staring arrays around 2011 with the NGP detector and more recently with COBRA. Figure 1 presents a chronology of these different developments with some examples of missions using each of these detectors. Figure 1.SWIR staring arrays development roadmap, including examples of mission using these detectors.  For these two next-generation detectors, the goal is to develop SWIR staring arrays answering most of the requirements for both hyperspectral and spectro-imagery applications described above. The main ones are highlighted hereafter:
3.NGP AND COBRA SWIR IRFPA DESCRIPTION3.1NGP and COBRA IRFPA basic architecture and manufacturing processAs mentioned previously, two different NGP and COBRA IRFPAs are proposed by LYNRED for SWIR Earth Observation missions. Even if they differ in size and performances as we will see later in this paragraph, they both rely on the same basic hybrid structural architecture composed of a Detector Circuit connected with a Read-Out Integrated Circuit (ROIC) thanks to indium bumps as illustrated in Figure 2. The Detection Circuit is manufactured by using LYNRED well-mastered wavelength tunable HgCdTe (Mercury Cadmium Telluride or MCT) technology: n-type photodiodes are formed thanks to a ion implantation process on a thin p-type MCT epitaxial layer that has been grown on a CdZnTe substrate, the IR sensitivity range being tailored to the desired wavelength range by adjusting the MCT layer material composition. The ROIC on its side is manufactured by using a specific silicon CMOS foundry process and additional under bump metallization (UBM) post-processing steps to facilitate the flip-chip hybridization. An anti-reflective coating in either single-layer (SLARC) or multi-layer (MLARC) version is finally added on top of the Detection Circuit to minimize the average reflectivity over a selected wavelength range. 3.2NGP and COBRA Detection Circuits and related performancesThe NGP and COBRA-L IRFPAs presented in Figure 3 are respectively 1024x1024 and 1840x1112 detection arrays with square-shaped pixel pitch of 15μm and 20μm. Designed for a targeted operational temperature around 150K, they both rely on the Detection Circuit and associated performances detailed hereafter. LYNRED has a strong experience with SWIR MCT technology and production that was gained through numerous flight models already delivered (more than 80 flights models have been shipped and ~50% of them have already been launched into space). Most of them use a 2.5μm cut-off material. NGP and COBRA detectors benefit advantageously from this highly mature technology building block and its associated heritage including space qualification. Figure 4 presents the typical spectral response of NGP and COBRA detectors with SLARC @150K. Even if the dark current is generally less critical for SWIR applications compared to LWIR/VLWIR ones, its contribution to the global signal can be close to the minimum photonic flux and a monitoring of this performance remains necessary. The next figure illustrates the dark current density values as a function of the IRFPA temperature: we have ~2.4 10-3 fA/μm2 at 150K. Spectral Detection Efficiency (SDE) is defined as the ratio of the amount of collected electrons over the amount of incident photons considering an ideal fully sensitive photodiode area (15x15 μm2 for NGP and 20x20 μm2 for COBRA). The figure hereafter illustrates the typical measured SDE evolution vs wavelength of IRFPAs including a ZnS quarter- wavelength SLARC added to minimize the average reflectivity and optimize the Quantum Efficiency. The typical PRNU is 1-3%. Figure 6.Typical measured Spectral Detection Efficiency at different positions on the IRFPA with SLARC (@150K).  Given the difference in pixel size between NGP and COBRA IRFPAs, the MTF which is a detector geometrical parameter must be analyzed specifically for each of them. The next figure illustrates typical MTF performances at the Nyquist frequency with the potential dispersion: 3.3NGP’s ROICThe other key IRFPA subpart is the ROIC, the brain of the component that is needed to integrate the electrons generated by the detector, amplify the different signals and multiplex them. As explained in paragraph 2, a specific attention was paid to the architecture, functionalities and performances of the NGP ROIC in order to adapt the detector design to the targeted applications. In particular, the ROIC enables to provide the user with essential functions like integration while readout operation, line selection, non-destructive multi-reading… Its architecture, based on an advanced CMOS technology process combined with a radiation hardening strategy inherited from previous space programs, was defined to obtain the following characteristics and performances.
3.4COBRA’s ROICAnticipating the evolutions of future Earth observation missions, Lynred started the development of the COBRA detector beginning of 2020. This detector will provide another improvement step in order to fit new hyperspectral and spectro-imagery space missions. In particular, this detector will provide the following main features:
As explained above, the COBRA detector is proposed in two versions in order to offer two different formats depending on the mission’s needs. Of course, all functionalities and performances for both versions are strictly identical. This ROIC architecture, based on an advanced CMOS technology process offering stitching capabilities needed to obtain a ROIC bigger than the reticule format, has been defined to obtain the followings characteristics and performances.
On a programmatic point of view, the ROIC functionalities and performances at ambient and operational temperature have been successfully verified. The final validation of the IRFPA is now on-going in order to confirm its overall characteristics. 3.5Overview of Key Parameters for NGP and COBRA IRFPAsThe Table 1 hereafter summarizes the key features, parameters and performances of both IRFPAs intended to SWIR Earth observation missions. Table 1.Summary of key parameters for NGP and COBRA IRFPAs.
4.DETECTOR CONFIGURATIONS FOR SWIR APPLICATIONSDifferent package architectures can be considered to accommodate NGP or COBRA IRFPAs and facilitate their integration at instrument system level:
5.CONCLUSIONLYNRED has developed and manufactured for more than 25 years IRFPAs and Detector Packages dedicated to space applications. Based on a well-mastered wavelength tunable MCT technology and appropriate CMOS processes, two staring arrays named NGP and COBRA have been presented in this paper: they both propose state of art characteristics, high electro-optical performances and superior reliability that are fully aligned with the most recent and future space-based Earth observation missions requiring improved spectral and spatial resolution. Examples of passive and active Detector Packages proposed in our product portfolio have also been detailed: their versatile architectures can be modified upon request to better fulfil any specific mission needs. ACKNOWLEDGEMENTSThe authors thank all LYNRED teams dedicated to IRFPAs and packages development as well as the European Space Agency (ESA), Centre National d’Etudes Spatiales (CNES) and Absolut Systems for their support and fruitful collaboration. REFERENCESNowicki-Bringuier Y., Chorier P.,
“Sofradir SWIR hyperspectral detectors for space applications,”
SPIE, Berlin
(2009). https://doi.org/10.1117/12.830721 Google Scholar
Leroy C., Fieque B., Jamin N., Chorier P.,
“SWIR space detectors and future developments at SOFRADIR,”
SPIE, Dresden
(2013). Google Scholar
Delannoy A., Fieque B., Chorier P., Ruine C.,
“NGP, a new large format infrared detector for observation, hyperspectral and spectroscopic space mission in VISIR, SWIR and MWIR wavebands,”
SPIE, Toulouse
(2015). Google Scholar
Jamin N.,
“New developments of multilinear and multispectral infrared sensors for space applications at LYNRED,”
ICSO2022). Google Scholar
Gaucel, J-M.,
“Remanence characterization of NGP detector in SWIR bands,”
ICSO2022). Google Scholar
|
