The SuperCam Instrument Suite on the NASA Mars 2020 Rover: Body Unit and Combined System Tests

Space Sci Rev (2021) 217:4 https://doi.org/10.1007/s11214-020-00777-5 The SuperCam Instrument Suite on the NASA Mars 2020 Rover: Body Unit and Combined System Tests RogerC.Wiens 1·Sylvestre Maurice2·Scott H. Robinson1·Anthony E. Nelson1· Philippe Cais3·Pernelle Bernardi4·Raymond T. Newell1·Sam Clegg1· Shiv K. Sharma5·Steven Storms1·Jonathan Deming1·Darrel Beckman1· Ann M. Ollila1·Olivier Gasnault2·Ryan B. Anderson6·Yves André7· S. Michael Angel8·Gorka Arana9·Elizabeth Auden1·Pierre Beck10 · Joseph Becker1·Karim Benzerara11 ·Sylvain Bernard11 ·Olivier Beyssac11 · Louis Borges1·Bruno Bousquet12 ·Kerry Boyd1·Michael Caffrey1· Jeffrey Carlson13 ·Kepa Castro9·Jorden Celis1·Baptiste Chide2,14 ·Kevin Clark13 · Edward Cloutis15 ·Elizabeth C. Cordoba13 ·Agnes Cousin2·Magdalena Dale1· Lauren Deflores13 ·Dorothea Delapp1·Muriel Deleuze7·Matthew Dirmyer1· Christophe Donny7·Gilles Dromart16 ·M. George Duran1·Miles Egan5· Joan Ervin13 ·Cecile Fabre17 ·Amaury Fau11 ·Woodward Fischer18 ·Olivier Forni2· Thierry Fouchet4·Reuben Fresquez1·Jens Frydenvang19 ·Denine Gasway1· Ivair Gontijo13 ·John Grotzinger18 ·Xavier Jacob20 ·Sophie Jacquinod4· Jeffrey R. Johnson21 ·Roberta A. Klisiewicz1·James Lake1·Nina Lanza1· Javier Laserna22 ·Jeremie Lasue2·Stéphane Le Mouélic23 ·Carey Legett IV1· Richard Leveille24 ·Eric Lewin10 ·Guillermo Lopez-Reyes25 ·Ralph Lorenz21 · Eric Lorigny7·Steven P. Love1·Briana Lucero1·Juan Manuel Madariaga9· Morten Madsen19 ·Soren Madsen13 ·Nicolas Mangold23 ·Jose Antonio Manrique25 · J.P. Martinez1·Jesus Martinez-Frias26 ·Kevin P. McCabe1· Timothy H. McConnochie27 ·Justin M. McGlown1·Scott M. McLennan28 · Noureddine Melikechi29 ·Pierre-Yves Meslin2·John M. Michel1·David Mimoun14 · Anupam Misra5·Gilles Montagnac16 ·Franck Montmessin30 ·Valerie Mousset7· Naomi Murdoch14 ·Horton Newsom31 ·Logan A. Ott1·Zachary R. Ousnamer13 · Laurent Pares2·Yann Parot2·Rafal Pawluczyk32 ·C. Glen Peterson1·Paolo Pilleri2· Patrick Pinet2·Gabriel Pont7·Francois Poulet33 ·Cheryl Provost32 · Benjamin Quertier3·Heather Quinn1·William Rapin11 ·Jean-Michel Reess4· Amy H. Regan1·Adriana L. Reyes-Newell1·Philip J. Romano13 ·Clement Royer33 · Fernando Rull25 ·Benigno Sandoval1·Joseph H. Sarrao1·Violaine Sautter11 · Marcel J. Schoppers13 ·Susanne Schröder34 ·Daniel Seitz1·Terra Shepherd1· Pablo Sobron35 ·Bruno Dubois36 ·Vishnu Sridhar13 ·Michael J. Toplis2· Imanol Torre-Fdez9·Ian A. Trettel13 ·Mark Underwood13 ·Andres Valdez1· Jacob Valdez1·Dawn Venhaus1·Peter Willis13 Received: 16 May 2020 / Accepted: 27 November 2020 / Published online: 21 December 2020 © The Author(s) 2020 The Mars 2020 Mission Edited by Kenneth A. Farley, Kenneth H. Williford and Kathryn M. Stack BR.C. Wiens [email protected] 1Los Alamos National Laboratory, Los Alamos, NM, USA 4Page 2 of 87 R.C. Wiens et al. Abstract The SuperCam instrument suite provides the Mars 2020 rover, Perseverance, with a number of versatile remote-sensing techniques that can be used at long distance as well as within the robotic-arm workspace. These include laser-induced breakdown spectroscopy (LIBS), remote time-resolved Raman and luminescence spectroscopies, and visible and infrared (VISIR; separately referred to as VIS and IR) reflectance spectroscopy. A remote micro-imager (RMI) provides high-resolution color context imaging, and a microphone can be used as a stand-alone tool for environmental studies or to determine physical properties of rocks and soils from shock waves of laser-produced plasmas. SuperCam is built in three parts: The mast unit (MU), consisting of the laser, telescope, RMI, IR spectrometer, and associated electronics, is described in a companion paper. The on-board calibration targets are described in another companion paper. Here we describe SuperCam’s body unit (BU) and testing of the integrated instrument. The BU, mounted inside the rover body, receives light from the MU via a 5.8 m optical fiber. The light is split into three wavelength bands by a demultiplexer, and is routed via fiber bundles to three optical spectrometers, two of which (UV and violet; 245–340 and 385–465 nm) are crossed Czerny-Turner reflection spectrometers, nearly identical to their counterparts on ChemCam. The third is a high-efficiency transmission spectrometer containing an optical intensifier capable of gating exposures to 100 ns or longer, with variable delay times relative to the laser pulse. This spectrometer covers 535–853 nm (105–7070 cm−1Raman shift relative to the 532 nm green laser beam) with 12 cm−1full-width at half-maximum peak resolution in the Raman fingerprint region. The BU electronics boards interface with the rover and control the instrument, returning data to the rover. Thermal systems maintain a warm temperature during cruise to Mars to avoid contamination on the optics, and cool the detectors during operations on Mars. 2Institut de Recherche en Astrophysique et Planetologie (IRAP), Université de Toulouse, UPS, CNRS, Toulouse, France 3Laboratoire d’astrophysique de Bordeaux, Univ. Bordeaux, CNRS, Bordeaux, France 4Laboratoire d’Etudes Spatiales et d’Instrumentation en Astrophysique, Observatoire de Paris, Meudon, France 5University of Hawaii, Manoa, HI, USA 6U.S. Geological Survey Astrogeology Science Center, Flagstaff, AZ, USA 7Centre National d’Etudes Spatiales, Toulouse, France 8University of South Carolina, Columbia, SC, USA 9University of Basque Country, UPV/EHU, Bilbao, Spain 10 Institut de Planétologie et d’Astrophysique de Grenoble, Université Grenoble Alpes, Grenoble, France 11 Institut de Minéralogie, Physique des Matériaux et Cosmochimie, CNRS, Museum National d’Histoire Naturelle, Sorbonne Université, Paris, France 12 Centre Lasers Intenses et Applications, University of Bordeaux, Bordeaux, France 13 Jet Propulsion Laboratory/Caltech, Pasadena, CA, USA 14 Institut Supérieur de l’Aéronautique et de l’Espace (ISAE), Toulouse, France 15 University of Winnipeg, Winnipeg, Canada 16 Univ Lyon, ENSL, Univ Lyon 1, CNRS, LGL-TPE, 69364 Lyon, France 17 GeoRessources, Université de Lorraine, Nancy, France The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 3 of 87 4 Results obtained with the integrated instrument demonstrate its capabilities for LIBS, for which a library of 332 standards was developed. Examples of Raman and VISIR spectroscopy are shown, demonstrating clear mineral identification with both techniques. Luminescence spectra demonstrate the utility of having both spectral and temporal dimensions. Finally, RMI and microphone tests on the rover demonstrate the capabilities of these subsystems as well. Keywords Perseverance rover ·LIBS ·Raman spectroscopy ·Infrared spectroscopy · Microphone on Mars ·SuperCam ·Jezero crater ·Mars 1 Introduction NASA’s Mars rovers have used various remote-sensing instruments over the last two and a half decades. The Sojourner rover was outfitted for remote sensing with only imagers (Golombek et al. 1999). The Mars Exploration Rovers (MER) were equipped with Miniature Thermal Emission Spectrometers (Mini-TES; Christensen et al. 2003) in addition to stereo multispectral imagers (Pancam; Bell et al. 2003). Mini-TES provided the first compositional remote sensing from a Mars rover beyond imaging filter wheels. However, the surface of Mars is covered by dust, which limits the ability of passive remote-sensing devices to make observations of the mineralogy or chemistry of the underlying rocks. The Chemistry and Camera (ChemCam) instrument on the Curiosity rover overcomes this challenge by using a laser to ablate the dust and additionally enabling remote depth profiles to several hundred µm to understand the surface conditions of the rocks (Maurice et al. 2012; 18 California Institute of Technology, Pasadena, CA, USA 19 University of Copenhagen, Copenhagen, Denmark 20 Institut de mécanique des fluides de Toulouse (CNRS, INP, Univ. Toulouse), Toulouse, France 21 Johns Hopkins University Applied Physics Laboratory, Laurel, MD, USA 22 Universidad de Malaga, Malaga, Spain 23 Laboratoire de Planétologie et Géodynamique, Université de Nantes, Université d’Angers, CNRS UMR 6112, Nantes, France 24 McGill University, Montreal, Canada 25 University of Valladolid, UVA, Valladolid, Spain 26 Agencia Estatal Consejo Superior de Investigaciones Cientificas, Madrid, Spain 27 University of Maryland, College Park, MD, USA 28 State University of New York, Stony Brook, NY, USA 29 University of Massachusetts, Lowell, MA, USA 30 Laboratoire Atmosphères, Milieux, Observations Spatiales, Paris, France 31 University of New Mexico, Albuquerque, NM, USA 32 FiberTech Optica, Kitchener, ON, Canada 33 Institut d’Astrophysique Spatiale (IAS), Orsay, France 34 Deutsches Zentrum für Luftund Raumfahrt (DLR), Institute of Optical Sensor Systems, Berlin, Germany 4Page 4 of 87 R.C. Wiens et al. Wiens et al. 2012). ChemCam uses laser-induced breakdown spectroscopy (LIBS) to obtain semi-quantitative elemental abundances from rasters of small observation points 350– 550 µm in diameter (Maurice et al. 2012). While ChemCam is limited mostly to chemical compositions rather than mineralogy, its ability to detect and quantify hydrogen is important for understanding the hydration state of the soils and for identifying some hydrated minerals (Schröder et al. 2015; Rapin et al. 2016,2018,2019; Thomas et al. 2020). ChemCam’s chemistry is complemented by visible-range (“VIS”) reflectance spectroscopy to ∼850 nm that allowed Johnson et al. (2015,2017) to constrain the mineralogy of iron-bearing materials (e.g., hematite, olivine, and ferric sulfates). However, this passive spectral range is not diagnostic for phyllosilicates and carbonates, which are important for understanding the history of Mars’ hydration, climate, and habitability. Because of Curiosity’s relative lack of remote mineral-identification capabilities, the Mars 2020 Science Definition Team mandated that the next NASA rover should possess the ability to observe mineral compositions by remote sensing (Mustard et al. 2013). The SuperCam instrument is a response to this requirement for remote mineralogy while preserving the ability to remove dust prior to making observations of nearby targets, and providing the same or better chemistry and high-resolution imaging as ChemCam. This new instrument resulted from a happy collision of ideas from previous mission proposals. It was recognized years ago that laser-induced breakdown spectroscopy (LIBS) and remote Raman spectroscopy both required a laser, a telescope, and an optical spectrometer (e.g., Wiens et al. 2005), and members of the SuperCam team sought to make that a reality over the years. The first attempt was for the ExoMars rover (Courreges-Lacoste et al. 2007), but the LIBS was descoped early in the ExoMars development (Rull et al. 2017). A remote Raman-LIBS combination much closer in design to SuperCam was developed during the Venus Surface and Atmosphere Geochemical Explorer (SAGE) mission which only proceeded through PhaseA development (Clegg et al. 2014). It was well understood that the Martian surface would benefit very strongly from a combination of remote Raman and visible-to-infrared (VISIR) reflectance spectroscopy. These two mineralogy techniques are highly complementary, as Raman signals occur as a result of a change in polarizability of a molecule with atomic vibrations, whereas infrared spectroscopy is sensitive to a change in the dipole moments. Raman spectroscopy is sensitive to a symmetric stretch but infrared spectroscopy is not. On the other hand, Raman can be insensitive to asymmetric stretches to which infrared spectroscopy is sensitive. The power of Raman spectroscopy originates from the fact that the activity of Raman modes depends both on the form of a vibrational harmonic and the stereochemistry of the molecule in question. Given the unexpectedly high abundances of feldspars (not generally recognized by near infrared spectroscopy) in igneous float rocks, conglomerates, and sedimentary outcrops in Aeolis Palus at Gale crater (e.g., Sautter et al. 2015,2016), it is important to identify these minerals, a capability provided by Raman spectroscopy. Short-wave infrared spectroscopy is highly selective and diagnostic for phyllosilicates and other (polar/nonpolar) minerals, so SuperCam includes both spectroscopic techniques. In addition, two more techniques were added. The hardware used for pulsed-laser Raman spectroscopy also enables time-resolved luminescence (TRL) spectroscopy. Luminescence is a general term that encompasses the fast, spin-allowed transitions referred to as fluorescence, and the slower, spin-forbidden transitions called phosphorescence. SuperCam can 35 SETI Institute, Mountain View, CA, USA 36 Université de Toulouse; UPS-OMP, Toulouse, France The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 5 of 87 4 Table 1 List of techniques employed by SuperCam for remote sensing. See text for explanations Technique Purpose DistanceaFootprint LIBS Quantitative elemental abundances To 7 m 0.25–0.45 mm Raman Identification of Raman-bright minerals, organics To ∼7 m 0.74 mradb TRL Identification of organics, minerals, REEscTo ∼7 m 0.74 mrad VISIRdIdentification of minerals; atmospheric studies To km VIS: 0.74 mrad IR: 1.2 mrad RMI Rock textures, contexts To km 18.8 mrad Microphone Physical properties of rocks, atmospheric studies To ∼4m N/A aMinimum distance, all optical cases: 1.05 m from the MU, positioned ∼2 m above flat ground bmrad =milliradians cREEs =rare-earth elements dVIS spectral range is 0.40–0.85 µm; IR range is 1.3–2.6 µm; other ranges are given in the text detect, distinguish, and characterize both fluorescence and phosphorescence using TRL. Finally, acoustic spectral sensing was added to remotely determine the physical properties of rocks and to assess atmospheric properties (e.g., Murdoch et al. 2019; Chide et al. 2019). Description of the SuperCam instrument, shown in Fig. 1, is divided among several papers. The science goals, most of the technical requirements, and the description of the Mast Unit (MU) are in a companion paper (Maurice et al. 2020). Here we provide a description of the Body Unit (BU) and of the integrated testing prior to launch. An overview of the rover calibration targets is in another companion paper (Manrique et al. 2020). More detail on subsystems and calibrations are or will be provided in other papers published separately (Royer et al. 2020, and future papers). 2 Instrument Overview We begin with a brief overview of the entire instrument (Fig. 1) before focusing on the Body Unit. Table 1provides a short list of the techniques employed by SuperCam. LIBS provides atomic emission spectra of material ablated from small spots on rock or soil targets, leading to quantitative elemental compositions of major, minor, and trace elements. ChemCam is able to detect and quantify ∼25 elements (e.g., Maurice et al. 2016), and SuperCam is expected to achieve the same as, or slightly better performance than ChemCam, with the same distance capability (Table 1). The analytical footprint is necessarily small, as the optical power density must be maximized to create a plasma on the target. The laser used to achieve the plasmas provides up to 14 mJ and >10 MW/mm2of 1064 nm photons per pulse (Maurice et al. 2020). As described later, the use of the transmission spectrometer in the green to red spectral range allows time gating and intensification of the signal, which may be used for special LIBS studies, such as to amplify an otherwise weak emission line. SuperCam employs the first use of green-laser Raman spectroscopy in space, and shares the distinction of the first planetary Raman spectrometer with SHERLOC (Bhartia et al. 2020, this journal). SuperCam achieves Raman spectroscopy at remote distances to ∼7m by using a pulsed laser—the same one as for LIBS, frequency doubled to 532 nm—and an intensified, gated detector coupled to a transmission spectrometer. The green Raman laser 4Page 6 of 87 R.C. Wiens et al. Fig. 1 The SuperCam instrument, consisting of the Mast Unit (MU), shown in (a), the Body Unit (BU), shown in (b), and the SuperCam Calibration Target (SCCT), shown in (c). As part of the MU (a), the laser can be seen to the left, protruding from behind the electronics box. The telescope is at the far end, at the center of which the periscope mirror for the green laser beam is mounted. The corresponding periscope mirror can be seen just past the electronics box, facing away from the camera. At the near end of the electronics box, a heating pad is just to the left of the connectors. The Mast Unit is mounted on insulating feet, and is shown here resting on a handling fixture. The BU (b) shows a transmission spectrometer resting behind two identical reflection spectrometers, all mounted on top of the electronics box. Three optical fiber bundles can be seen with their protective shields near the upper left; these transfer light to the spectrometers from the demultiplexer. The only part of the demultiplexer that is visible is the fiber connector, protruding at the left center. This is where the light from the MU enters the BU. One of three sets of thermoelectric coolers is seen in the lower center, identified by two visible heat pipes that run under the spectrometers to cool their detectors. On the SCCT (c), twenty-nine circular targets and several other calibration targets are mounted. The titanium plate at the upper right is used for wavelength calibration via LIBS spectra. Imaging targets and a Mars meteorite sample line the left side. Dimensions of all three SuperCam units are given in Table 2 The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 7 of 87 4 beam is collimated rather than focused, and it overlaps with the 0.74 mrad field of view (FOV) of the spectrometer, defined by the telescope and optical fiber that transfers the light from the MU to the BU (Maurice et al. 2020). This footprint thus ranges between 1.5 mm diameter when looking straight down at the ground from the telescope’s height of 2 m, to ∼5 mm diameter when observing at a distance of 7 m. Raman spectra are always dim, and collecting these spectra remotely results in a limited number of photons (Sect. 7.3.3). SuperCam’s design is a compromise to enable all of the techniques, which limited the ability to optimize the photon throughput, with the result that SuperCam observes minerals and organic materials that are strong Raman emitters. Among the best are carbonates, sulfates, and phosphates, but SuperCam also expects to identify quartz and plagioclase feldspar, the latter of which are not, or are only poorly, detected by near-infrared reflectance spectroscopy. The pulsed laser and time-gated spectrometer also provide the capability for TRL. Prompt fluorescence is emitted by organic materials, being emitted and decaying within nanoseconds of stimulation (Lakowicz 2006). This prompt organic fluorescence is the bane of Raman spectroscopy on Earth, as even time-resolved Raman spectroscopy with nanosecond laser pulses and the fastest-gated detectors generally do not discriminate against it. However, on Mars, it may be a powerful tool for discovering concentrations of organic materials. SuperCam’s intensifier gate can be delayed up to milliseconds with a temporal resolution of 10 ns, providing a second dimension for characterizing mineral fluorescence. Using the time dimension, the presence of certain rare-earth elements (REEs) and transition metals can be identified (Gaft et al. 2015; Ollila et al. 2018). SuperCam’s VISIR reflectance spectroscopy is the first to cover the spectral range of 0.4 to 2.6 µm from the surface of Mars. It does so utilizing several spectrometers. The main phyllosilicate identification region, in the near infrared, has been successfully used by the orbiting spectrometers, specifically the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM; Murchie et al. 2007) and Observatoire pour la Mineralogie, l’Eau, les Glaces et l’Activité (OMEGA; Bibring et al. 2004). On SuperCam this spectral range is provided by a wavelength-scanning spectrometer (1.3–2.6 µm) in the MU, while observations in the VIS range (0.4–0.85 µm) take advantage of the BU spectrometers that are also used for LIBS, Raman, and TRL spectroscopies. There is a gap in the spectral coverage between 0.85 and 1.3 µm (Fig. 41 of Maurice et al. 2020). The VIS range (covered by the violet and transmission spectrometers) has a somewhat smaller target footprint than the IR spectrometer (Table 1). On Mars 2020, SuperCam’s point spectral observations are well complemented by Mastcam-Z, which can image at discrete wavelengths from 0.44 to 1.0 µm (Bell et al. 2020, this journal). The remote micro-imager (RMI) and microphone complete the list of techniques available with SuperCam (Table 1). High-resolution context images of the analysis areas are critical to their interpretation, especially given the small footprint of the analysis techniques. The RMI provides this context via Bayer-filter color images with a resolution ≤80 µrad (defined as a line pair with more than 20% contrast), allowing 160 µm grains to be resolved at 2 m distance from the instrument, e.g., directly in front of the rover. (The RMI resolution is not pixel limited, as its instantaneous field of view, or IFOV, is 9.2 µrad, not considering the Bayer filter.) The microphone provides acoustic signals from the LIBS shock wave. Studies (e.g., Chide et al. 2019) have shown that a combination of rock hardness and density can be obtained by the rate of decrease of acoustic energy with increasing number of laser pulses at the same location. The microphone will also be useful for atmospheric studies (Chide et al. 2020), perhaps including phenomena of which we are not yet aware. As subsystems of the MU, the RMI and microphone are described in Maurice et al. (2020). Figure 2shows a schematic diagram of the SuperCam instrument. SuperCam is divided into two major units, the BU and the MU. The MU resides at the top of the rover’s mast and 4Page 8 of 87 R.C. Wiens et al. Fig. 2 Schematic diagram showing the major units and subcomponents of the SuperCam instrument suite. The Mast Unit (MU) consists of the main laser which provides two wavelengths using two Galilean beam expanders, the telescope, a continuous-wave laser (CWL) for focusing, and a microphone. The optical box (OBOX) also includes the infrared (IR) spectrometer and the Remote Micro-Imager (RMI), the detector of which is a complementary metal oxide semiconductor (CMOS). An electronics box (EBOX) controls and powers the various subsystems in the MU. Acquisition of the target is provided by the rover mast azimuthal (AZ) and elevation (EL) motions. Electrical cables and an optical fiber connect the Body Unit (BU) to the MU. The fiber carries light in the 245–853 nm range to the demultiplexer (labeled Demux) in the BU, which distributes the light to three spectrometers covering ultraviolet (UV), violet (VIO), green, orange, and red spectral ranges. The latter are characterized by a transmission spectrometer, which uses an intensifier driven by a high-voltage power supply (HVPS). All three BU spectrometers collect light with charge-coupled devices (CCDs) cooled by thermoelectric coolers (TECs). The electronics box (EBOX) in the BU operates the instrument, provides power to the BU spectrometers and the MU, and communicates with the rover through the control and data handling (C&DH) board. A set of calibration targets is mounted on the back of the rover to facilitate calibration while on Mars contains the laser, telescope, Remote Micro-Imager, Infrared Spectrometer, Microphone, and associated electronics. The MU was designed, built, and tested in France under the support and direction of the Centre National d’Etudes Spatiales (CNES), with integration at the Institut de Recherche en Astrophysique et Planetologie (IRAP). The BU contains an optical demultiplexer, optical spectrometers for LIBS, Raman, and passive VIS spectroscopy, and associated electronics to control the BU and MU and interface with the rover. The BU was designed, built, and tested at Los Alamos National Laboratory, in the US. A third part of the instrument, onboard calibration targets were provided by an international working group within the SuperCam team and consist mostly of sintered pellets (e.g., Montagnac et al. 2018). Characterization of the targets was performed by a team of European scientists; the Universidad de Valladolid was responsible for integration and environmental testing of the target assembly (Manrique et al. 2020). The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 9 of 87 4 Fig. 3 Locations of the SuperCam units on the rover. The right side shows the rover body inverted, with the Body Unit circled. It is next to the RSM side of the rover to minimize the length of the fiber that transfers the optical signal from the mast unit. The rover’s instrument and electronics bay is 1181 ×1106 mm (length, left-right, x width) Table 2 SuperCam physical properties aIncludes operation of the thermoelectric coolers. BU power when idling is 12 W Mass (kg) Dimensions l×w×h(mm) Max power (W) Mast Unit 6.11 383 ×201 ×163 27 Body Unit 4.44 221 ×157 ×205 43a Cal. Target 0.25 110 ×100 ×17 0 TOTAL 10.80 70 Figure 1shows the completed SuperCam instrument, which was designed to conform to the space provided by the rover project (Fig. 3). The available volume was quite similar to that of ChemCam (Maurice et al., Wiens et al. 2012), and so from both heritage and mechanical/thermal-requirement standpoints, SuperCam looks very similar to its predecessor. SuperCam’s telescope differs from ChemCam’s, as the newer instrument features a periscope mirror that directs Raman laser light to the intended sample (Fig. 1). The BU’s optical demultiplexer and two reflection spectrometers—ultraviolet and violet (UV, VIO)—are nearly identical to ChemCam’s, but its third reflection spectrometer was replaced by a time-gated, intensified, high-throughput transmission spectrometer to enable Raman spectroscopy. The calibration target assembly (Fig. 1; Manrique et al. 2020) is vastly expanded and improved over ChemCam’s (Fabre et al. 2011; Vaniman et al. 2012;Wiens et al. 2012). Table 2presents some of SuperCam’s physical properties. Overall, SuperCam has almost exactly the same mass as ChemCam. The MU and calibration targets are slightly heavier, but the BU is lighter. The difference is largely due to ChemCam’s thermoelectric cooler (TEC) assembly, which—thanks to a JPL team–was added to that instrument within the final 18 months of its development, when mass was not a concern (Wiens et al. 2012). The TECs on SuperCam were planned from the beginning, and so the design of the associated cooling system is lighter. Another difference affecting the mass of SuperCam is the replacement of beryllium by titanium for the BU spectrometers. Thermal expansion of ChemCam’s spectrometers requires careful compensation for changes in the wavelength calibration with respect to the instantaneous temperature of the spectrometers on Mars, which change by 20 ◦C diurnally. Titanium has a lower coefficient of thermal expansion (CTE) than Be, so maintaining wavelength calibration will be easier. To fit Ti spectrometers into the tight mass 4Page 16 of 87 R.C. Wiens et al. Fig. 9 Typical inspection images of demultiplexer end (a) and spectrometer end (b) of fiber bundles, showing the one feeding the transmission spectrometer. The core of each fiber is 50 µm diameter. Fibers are backlit for the inspections. The relative intensity of each fiber may be a function of the position of the lamp, and so does not indicate relative throughput in this case. Inset shows a magnified image of the linear array of fibers at the spectrometer end before (c)andafter(d) bonding a 28.7 µm slit. Inset (e) shows the mapping of fibers in the array, with central fibers in the circle mapped to central fibers in the line Table 5 Modeled theoretical maximum transmission through the fiber bundle and slit assemblies, and measured transmissions for various wavelengths Spectrometer Slit width (µm) % Transmission @ Wavelength (nm) Theoretical maximum 300 405 565 660 780 UV 20.7 234242443651 VIO 21.0 224242423552 Transmission28.7 325153534669 at LANL for functionality and throughput, and to ensure proper terminations, circularity and linearity of the arrangements, and mapping from one end to the other. After characterization, slits procured from National Aperture were installed on the best bundles. The slits are 1.16 mm long, produced in 13 µm thick plates that were blackened. The slits were inspected and measured at LANL prior to installation, and performance was checked after installation. The slit widths used in each spectrometer are given in Table 5, along with the measured optical efficiency of the completed fiber assemblies with slits attached. Note that the reported transmission values include losses both from the slit (given under “Theoretical maximum”) and the losses from the packing fraction of the nineteen 50 µm fibers. As with any spectrometer, selecting a slit width involves a trade between spectral resolving power and optical efficiency. To ensure the optimal choice, we built and tested fiber bundle assemblies with several slit widths and on each we measured the optical throughput and resolving power, defined as the average full width at half maximum (FWHM) of several atomic emission lines from a neon lamp. The results, seen for the transmission spectrometer in Fig. 10, led us to select the 28.7 µm wide slit assembly for flight. The neon emission lines that were used do not cover the most challenging spectral region, from 535–555 nm, where spectral resolution is worse. With this slit, the transmission spectrometer just meets its The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 17 of 87 4 Fig. 10 Comparison of optical throughput efficiency (a) and spectral resolution (b) as functions of entrance slit width for transmission spectrometer. The theoretical maximum optical transmission is shown as a dotted line in (a). As expected, a wider slit allows more light but also increases the FWHM of spectral features. The 28.7 micron slit was selected for flight. Although not shown in the figure (the lowest-wavelength Ne line used was at 576 nm), this slit just meets the 12 cm−1resolution requirement at the short-wavenumber end of the spectrum FWHM resolution requirement of 12 cm−1everywhere. Similar testing led to the selection of the 20.7 and 21.0 µm slits for the UV and VIO spectrometers, respectively (Table 5). 3.1.3 Reflection Spectrometers As shown in Fig. 5, two of the three spectrometers are crossed Czerny-Turner reflection spectrometers. Table 3gives specifications in terms of wavelength range and resolution of the spectrometers. The f/4 optical design is essentially identical to that of ChemCam (Wiens et al. 2012). The two spectrometers are identical to each other with the exception of the mirrors, gratings, and grating angles. Figure 11 shows a rendering of the interior of the reflection spectrometers. The gratings, manufactured by Richardson, are 2400 lines per mm (lpmm) on Schott N-SF8 glass substrates with aluminum coatings. The UV grating is 240 nm blaze wavelength, while the VIO grating is 300 nm. The mirrors have spherical concave surfaces with 100 mm radii of curvature, with dielectric surfaces for maximum reflectance in the respective spectral ranges. Mirrors were manufactured by OPCO, and have reflectivity >94% between 240–340 nm for the UV spectrometer, and >98% between 380–500 nm for the VIO spectrometer. The diffracted images of the spectrometer slits are projected onto charge-coupled devices (CCDs). All three spectrometers have identical CCD assemblies with the exception of the CCD surface coatings, which were selected to optimize light collection at the respective ranges (Table 3). As with ChemCam, they are E2V 42-10 CCDs, 13.5 µm square pixels in an array of 2048 ×515 with 50 additional blind serial-register pixels on each side. The active area is 27.6×6.9 mm. Extensive testing was carried out on the CCDs and on replicates from the same factory lot. In addition to testing and characterization at the manufacturer (Table 3), read noise, dark current, image and serial pixel well capacities, and sensitivity to input voltages were all tested at LANL, and the best among several CCDs of each type were assigned for the flight instrument. A blemish was found in the flight VIO CCD after integration onto the CCD board, in which the sensitivities of pixels in rows between 240 and 250 at columns 1323–1326 are significantly reduced. This was not noted by the manufacturer. The LIBS spectral image from andesite standard JA-3 is shown in Fig. 12.Inthereflectance spectrometers, the light is spread across up to 180 central rows. (The VIO blemish is not seen.) ChemCam integrates the signal from 200 vertical rows to maximize the collection (Wiens et al. 2012). For SuperCam we plan to use several different vertical integrationrow settings on the reflection spectrometers (Table 3). Operation of the SuperCam CCDs and their timing with respect to the laser are discussed in Sects. 3.3.1 and 4.2. 4Page 18 of 87 R.C. Wiens et al. Fig. 11 Rendering of a SuperCam reflection spectrometer. Light enters from the fiber bundle and slit assembly at the upper right. It is collimated by the circular mirror at lower left. The grating at upper right provides spectral dispersion, the first order of which is focused by the rectangular mirror at the upper left onto the detector assembly at the lower right. Baffles can be seen along the upper and lower sides and next to the grating, used to absorb higher order reflections from the grating. The cylinder protruding at the left is a thermal switch 3.1.4 Transmission Spectrometer The transmission spectrometer is a significant innovation relative to ChemCam; it enables the remote Raman and TRL spectroscopy techniques, while providing enhanced capabilities for LIBS and also collecting passive VIS reflectance spectra. The primary functions are to provide high transmission and intensification at relatively high resolution for the weak Raman signal, and to provide time gating to 100 ns to minimize noise from background ambient light and fluorescence. The transmission spectrometer design is based very loosely on earlier commercial Raman transmission spectrometers such as the Kaiser Holospec, but with a long separate development period, first at U. Hawaii with modified commercial parts (e.g., Sharma 2007) and then at LANL with custom parts, starting around 2009. Figure 5shows the basic architecture of the spectrometer, while Fig. 13 shows an optical ray trace. Three diffracted bands are used to simultaneously cover the wavelength range required for LIBS (535–853 nm) while achieving the resolution needed for planetary greenlaser Raman spectroscopy (12 cm−1) over its range (∼105 to 4000 cm−1; 535–676 nm). Because the LIBS spectral resolution and light collection requirements are not as stringent in this spectral range, the transmission spectrometer was optimized over the narrower Raman range while still covering the broader LIBS spectral range. The feature that limits the resolution is the intensifier rather than the CCD. AsshowninFig.5and Fig. 13, light passes through the slit and is collimated before encountering a dichroic beam splitter assembly, which separates a red optical band. Two parallel beams travel to separate transmission gratings. A compound grating (one directly behind the other, with an angular separation) separately diffracts green and orange light, while an The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 19 of 87 4 Fig. 12 LIBS light from andesite standard JA-3 (shot in terrestrial atmosphere) as seen by SuperCam’s detectors. Shown are full images of the CCDs of UV, violet (VIO), and transmission (TS) spectrometers. These images are shaded as the natural logarithm of intensity. Rows and columns are indicated, as are the wavelengths. Two lines extending to the top of the VIO image are Hg emission lines from room lights. In the transmission spectrometer, the edge of the intensifier can be perceived by the edges of the continuum emission adjacent grating diffracts the red band. All three bands are focused on the intensifier. As a final step, intensified light is re-focused on the CCD in the relay section. A rendering of the spectrometer is shown in Fig. 14. The dichroic beam splitter (Figs. 14,15a) receives light from the collimating lens and reflects it down toward the gratings. Below the reflecting mirror, a custom dichroic mirror from OptoSigma, positioned at a 45◦angle, passes >93% of light from 720–850 nm directly down to the red-band grating while reflecting to the side >95% of light from 530–700 nm. This light forms the green and orange bands. It is reflected by another 45◦mirror straight down toward the compound green-orange grating. 4Page 20 of 87 R.C. Wiens et al. Fig. 13 Ray traces in a model of the transmission spectrometer, showing collimation, diffraction, and formation of slit images on the image intensifier entrance window and relay lens carrying intensified image to CCD sensor. Rays are colored by wavelength; 534 nm to 853 nm to the intensifier, and 545 nm from the intensifier phosphor to the CCD Fig. 14 Cutaway rendering of the transmission spectrometer showing the internal layout The two parallel beams are directed onto three transmissive diffraction gratings (Figs. 14, 15b); the long wavelengths >715 nm go to one grating, and the shorter wavelengths of 530– 715 nm go to a duplex compound grating. The diffractive element in the 715–855 nm beam The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 21 of 87 4 Fig. 15 Subsections of the transmission spectrometer. The spectrometer section (a)isshown with the fiber connector facing to the rear. The dichroic beam splitter is in focus. The grating assembly (b) is shown with the input side visible, and relay lens is seen in its housing (c), looking from the CCD side. Other components on the table are visible through the relay lens is a single 1800 lpmm grating. The compound grating contains two diffractive elements: a 2480 lpmm grating optimized for the 530–620 nm range (nominal center: 570 nm), and a 2110 lpmm grating optimized for the 610–700 nm range (nominal center at 670 nm). The rulings in the two gratings are rotated 1.8 degrees from each other, so that the diffracted spectra are spatially separated on the intensified CCD (ICCD) input window. The single 1800 lpmm grating is rotated by an additional 1.8 degrees so that its diffracted output is spatially separated from the other two. Volume-phase (VPH) transmission gratings from Wasatch Photonics Inc. were previously flight qualified as part of the ExoMars/RLS instrument (Rull et al. 2017). The output of the three VPH diffraction gratings is a set of three spectral bands stacked in the spatial direction on the face of the ICCD (see Fig. 16). An achromatic lens assembly, similar to a scan lens in design, focuses each spectral component 4Page 22 of 87 R.C. Wiens et al. Fig. 16 Map of spectrometer slit images at various wavelengths (in nm) with the field of view of the intensifier (circle) drawn for scale. The green band (middle, 530–618 nm) starts part-way in from the CCD’s edge in order to optimize the resolution. The orange band (598–720 nm) is projected to the left, and the red band (707–853 nm) to the right. The CCD is read-out from the right side in this orientation. Compare with the bottom panel of Fig. 12 onto the photocathode of the ICCD sub assembly. The images of the slit on the ICCD face are magnified 1.22 times by the scan lens assembly. The ICCD subassembly consists of three main elements: intensifier tube, CCD sensor, and relay lens (Fig. 15) to carry light from the intensifier to the CCD. The intensifier tube provides two essential functions for Raman and TRL spectroscopies: It amplifies the weak optical signal returned from the sample and allows very fast (requirement to ≤100 ns) gating. This gating is essential to isolate the brief Raman signal from the delayed and much longer-lived luminescence generated by the sample (Sect. 7.4) and to study time-resolved luminescence. Primary ICCD design considerations are optical gain, resolution, spectral responsivity, gating ability, and flight qualification. Working closely with ITT Exelis in Roanoke VA (now part of Elbit Systems), we identified an intensifier which meets all requirements. The intensifier features a continuously variable gain over 45 dB photon/photon, resolution exceeding 48 lpmm, and a photocathode with quantum efficiency (QE) over 30% from 530 nm to ∼900 nm. Gain is adjusted by selecting the voltage across the tube’s microchannel plate, and signals are time-gated by gating the voltage. A high-performance relay lens assembly couples light from the intensifier phosphor output to the CCD. A custom double-Gauss lens design is used to meet requirements for high coupling efficiency (f/2.7), >30◦FOV and compact size. Lens design takes advantage of the fact that light exiting the intensifier is nearly monochromatic at 545 nm. The intensifier’s P43 phosphor does generate minor outputs at other wavelengths, but these are suppressed by coatings on the relay lens surfaces. The resolving power of this relay lens assembly surpasses that of the intensifier tube. The paraxial magnification of the relay lenses is −1.26. The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 23 of 87 4 Fig. 17 Side-by-side rendering of ChemCam and SuperCam Body Units. SuperCam’s BU is 345 g lighter and occupies a slightly smaller volume A 20 µm feature on the intensifier maps to 25.2 µm on the CCD, which corresponds to 1.87 pixels. 3.2 Mechanical and Thermal Description 3.2.1 Mechanical The SuperCam BU mechanical design uses ChemCam as its starting point (Fig. 17), but incorporates many refinements and design changes driven by the specific SuperCam requirements. The BU consists of five major sub-assemblies (Fig. 4). These are the electronics box, demultiplexer, reflection spectrometers, transmission spectrometer, and connector interface bracket. The design is modular, allowing for each sub-assembly to be assembled and initially tested and aligned before final integration. This modularity was also intended to minimize the impact upon other modules in case disassembly was needed for trouble shooting. The electronics box (EBOX) (Fig. 18) is an aluminum structure that contains the spectrometer electronics (SE) printed circuit board (PCB), command and data handling (C&DH) PCB, and the low-voltage power supply (LVPS). It also functions as a stable platform for the other sub-assemblies mounted on its top panel. Additionally, the EBOX serves as the thermal interface to the RAMP; the interface is described in Sect. 3.2.2. The EBOX design uses a backplane/daughter-card architecture. Each daughter card can be easily and independently installed or removed without significant impact to the overall assembly, and with no impact to the opto-mechanical assemblies. This is an improvement over the ChemCam design. All electrical connections between boards are through the backplane using nano-D connectors. Each daughter board is mounted to an aluminum frame, which acts as a thermal sink and provides stiffness to the board in vibration. Each module slides into the housing and is clamped in place with wedgelocks. The wedgelock joints provide a reasonable thermal interface between the housing and the module frame. The LVPS frame also makes up the 4Page 24 of 87 R.C. Wiens et al. Fig. 18 An exploded view showing the major components of the electronics box (EBOX). Its backplane/daughter-card architecture allows easier removal and installation of circuit boards during assembly and testing. The EBOX acts as the support structure for the optical sub-assemblies base panel of the EBOX. On the top of the EBOX is a patch panel for distributing power to the decontamination heater systems. This panel uses nano-D connectors to minimize mass and volume, and allowed easy electrical connection of heaters during assembly of the major optical subsystems onto the EBOX. The demultiplexer uses the same overall architecture as ChemCam’s but was optimized for SuperCam. This included reducing the overall length by ∼25 mm to allow more room for the transmission spectrometer, and introduction of an improved fiber cable support bracket to better support the fiber cables in vibration and to protect the fiber optic cables during handling. The modifications also included accommodations for the refined optical elements. Figure 7shows the opto-mechanical design of the demultiplexer. The lenses used to focus the light onto the fiber tips are mounted in titanium barrels with anti-backlash springs. The focal distance was adjusted during assembly by turning the barrels which were easily accessible with the demultiplexer cover removed. Two-dimensional alignment of the fibers was accomplished by translation of the fiber tip and connector across the top surface of the demultiplexer body. Once aligned and focused, the barrels and fibers were locked in place to prevent movement in vibration. The demultiplexer is also equipped with decontamination heaters, thermostats, and a light-tight venting baffle. The reflection spectrometers were based on the same basic optical design as ChemCam but the mechanical designs were significantly refined (Fig. 11). The SuperCam reflection spectrometers are composed of two mechanically identical spectrometer assemblies bolted together. For ChemCam, the three reflection spectrometer housings were manufactured from beryllium. Due to the health-related complications of using beryllium and to improve ther- The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 25 of 87 4 mal stability, it was decided to use titanium for SuperCam’s spectrometers. Titanium has a significantly lower CTE than beryllium and 97% lower thermal conductivity which helps thermally isolate the optics. Titanium is 2.4 times denser than beryllium, and therefore the overall design required significant optimization to keep the mass as low as possible. In the end, each SuperCam spectrometer had almost the same mass as those on ChemCam. The optical mounts were also refined to reduce mass, improve bonding of optical components, and make optical adjustment easier than on ChemCam. Each optic is mounted on compact kinematic mounts that can be adjusted from the exterior of the housing. Once adjusted and staked, a cover was placed over the mounting region. This cover provides a light tight seal as well as a mounting location for decontamination heaters. Each spectrometer housing vents through a light-tight baffle equipped with an 18 micron filter mesh to protect from particulate contamination. The transmission spectrometer (Fig. 14) is an entirely new design and accounted for a large fraction of the developmental work to create a compact robust package that could fit within the allocated SuperCam volume and mass limitations. The lenses are mounted in Ti barrels with finger springs and retaining rings. Titanium was chosen to better match the CTE of the lenses to minimize thermal stresses on the lenses and allow for tighter tolerances. A number of the optical elements required the ability to adjust after assembly. These adjustments then required locking to prevent motion in vibration and thermal cycling. Some of the barrels could be adjusted for focus from the exterior of the spectrometer and then locked in place. The fold mirrors and diffraction gratings were mounted in Ti flexures that allowed tip/tilt adjustment. Below the transmission spectrometer is the high-voltage power supply (HVPS) (Fig. 14) that powers the intensifier. The HVPS is attached to the intensifier assembly only by the electrical leads; it is isolated thermally and structurally to eliminate any thermal leaks or induced stresses into the optical assembly. Packaging the HVPS in the small space below the transmission spectrometer required special care. All three spectrometers have mechanically identical CCD modules mounted to the rear of the spectrometer housings; they are thermally isolated from the housings (Fig. 19). The modules contain the CCDs and CCD PCBs as well as decontamination heaters and thermostats which interface with the cooling system on the EBOX. The CCD boards connect to the SE board via flex cables routed to maintain more than the minimum bend radius for the cables. Additionally, two of the CCD modules have thermostats that measure the external housing temperatures for the spectrometer units and report this information in the state of health data. Due to the very confined space SuperCam occupies in the rover, a connector interface bracket was added to provide easier access to the electrical connectors during integration with the rover (Fig. 3). Flex circuits route the connections from the EBOX PCBs to the top of the connector bracket. Despite significant increases in capability and performance, the final flight SuperCam Body Unit ended up 345 g lighter than ChemCam’s, occupying a slightly smaller volume. 3.2.2 Thermal The SuperCam BU thermal design consists of two active systems for CCD cooling and for decontamination, and passive design elements for cooling of electronics and thermal stability for optical systems. In all cases, the major thermal interface for the instrument is the RAMP, which is a fluid-loop-controlled mounting surface for avionics and other instrumentation. The RAMP provides an interface temperature within −40 to +50 ◦C. Based on experience with the MSL rover, which has the same overall thermal design, the RAMP interface to the BU is expected to be between 0 and +35 ◦C on Mars. In order to maximize 4Page 32 of 87 R.C. Wiens et al. Fig. 25 Schematic diagram for the BU spectrometer electronics board. ADC =analog-to-digital converter; DAC =digital-to-analog converter; Op Amp =operational amplifier; MOSFET =metal-oxide semiconductor, field-effect transistor. See main text for the meaning of other abbreviations second (Mbps) for commands (from RCE to SuperCam) and 7.5 Mbps for telemetry (from SuperCam to RCE). Reset and boot bank selection is provided via a set of RS-422 initialization discrete signals. The C&DH also acts as the digital interface to the MU over a 6-meter flex cable between the BU and the MU. An LVDS HSS interface is used to transmit RMI and Microphone data products from the MU to the BU at up to 10 Mbps. In the C&DH FPGA, this HSS interface is mapped to one of the Leon3FT’s four SpaceWire interfaces at 30 MHz. An LVDS universal asynchronous receiver-transmitter (UART) interface at up to 9600 baud is the primary command and telemetry interface between the two units, and a set of LVDS discrete signals are used as a laser trigger, an intensifier trigger, and a reset for the MU. A set of three SpaceWire links operating at 30 MHz interface between the SE module’s three CCD interfaces and the C&DH, which allows the use of direct memory access (DMA) to the C&DH SDRAM. A number of discrete signals to control the HVPS also originate on the C&DH and are passed through to the SE. The C&DH controls two power switches on the LVPS board, used to control power to the MU and to the BU TECs, respectively. SOH circuitry is also provided on the C&DH to monitor an array of temperatures and voltages. The SE module’s primary function is to clock and read the e2v CCDs. A schematic diagram is shown in Fig. 25. A set of three ChemCam-heritage 14-bit Maxwell/DDC analogto-digital converters (ADCs) are used. ChemCam-heritage CCD clock and voltage conditioning circuits are implemented on-board. A Microsemi FPGA is used to house the CCD clocking logic and SpaceWire firmware. The CCD pixels are digitized at 14 bits, but are over-sampled to 16 bits, taking the mean of four readings (Fig. 26). Additionally, a digital The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 33 of 87 4 Fig. 26 Oscilloscope traces illustrating the correlated quadruple sampling used to improve the signal– to-noise ratio. ORST is output reset pulse; R3 is readout register phase-3 clock pulse correlated double-sampling technique is employed to subtract the CCD reference voltage from the active pixel. These techniques together improved the detector read noise by a factor of over two versus ChemCam. One side effect is that a saturated pixel can read slightly below 216 digital numbers (DNs) due to the averaging and subtraction that occurs simultaneously with the digitization. Another side effect is that the maximum transfer rate using this method is 400 kHz, which is about 20% slower than used by ChemCam. The result is somewhat higher dark noise due to longer transfer times (Sect. 4.2). Nominally, the CCDs are read in a row-summed one-dimensional (1D) mode. A two-dimensional (2D) diagnostic mode is also available, and can be used to image all rows of the CCDs (e.g., Fig. 12). The SE board interfaces with the HVPS by providing the supply voltage from the LVPS, along with trigger signals from the C&DH and a gain voltage supplied by a 12-bit digitalto-analog converter (DAC). Board-level SOH circuitry is also provided on the SE for monitoring of SE, CCD, spectrometer, and HVPS temperatures, voltages, and currents. The CCDs are each individually mounted to identical circuit boards (Fig. 27), which contain front-end signal-conditioning electronics. These modules are physically mounted in the three BU spectrometers, and are electrically connected to the SE via flex connectors. A number of instrument temperature sensors are used to monitor the LVPS, C&DH, the CCD detectors, the HVPS, and the spectrometer housings. These sensors are only active when the instrument is powered on. They were calibrated during instrument thermal-vacuum tests. 3.3.2 High-Voltage Power Supply, Intensifier, and Timing with Respect to the Laser A custom high voltage power supply (HVPS) powers the transmission spectrometer’s optical intensifier. A simplified schematic is shown in Fig. 28. The HVPS applies three bias voltages to the intensifier: −600 V from the photocathode to the microchannel plate front, +1200 V across the microchannel plate front to rear, and +3800 V from the microchannel plate rear to the phosphor screen. Gain is adjustable via 0–5 V input from the SE board using a 12-bit DAC. This adjustment changes the bias voltage to the intensifier. Resolution is maintained across the range of gains used for this instrument. The HVPS circuitry consists of a low-voltage side and a high-voltage side. The highvoltage side is potted to prevent arcing in the thin Mars atmosphere. Thermal-cycling life 4Page 34 of 87 R.C. Wiens et al. Fig. 27 CCD board schematic diagram. Op Amp =operational amplifier; Temp =temperature. See main text for other abbreviations Fig. 28 Simplified schematic diagram of the high-voltage power supply that operates the intensifier depicted in Fig. 14.MCP=micro-channel plate tests were performed on several HVPS units to ensure that they could survive a large number of diurnal cycles as will be experienced within the rover body. The C&DH board provides gate-open and -close pulses that are synchronized with the laser and with the CCD and its readout process. Diagnostic testing of the intensifier’s rise time and stability was performed by firing the laser at a Raman-bright target and making separate collects with different delay times. The laser pulse (and hence the Raman signal) duration is 4 ns (Maurice et al. 2020, this journal). The precision of the laser-to-intensifier timing is approximately ±14 ns, as the two components are operated by different FPGAs in their separate parts of the rover (MU clock frequency is 20 MHz and the BU clock speed is The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 35 of 87 4 Fig. 29 Description of testing the duration of the intensifier gate by generating a Raman signal on a target multiple times with different delays for each collect. The technique is illustrated schematically in (a). The result is shown in (b) for a 100 ns gate. Each point represents data collection at the delay given on the x-axis. The y-axis represents the Raman signal peak intensity for a given delay setting. The rising portion of the curve on the left of the plot (b) comes from signals received at the end of the integration period (a, top); the falling portion on the right in (b), at longer delays, represents signals received at the beginning of the integration period, as shown at the bottom of (a) 50 MHz; timing uses the half-frequency steps for double the accuracy). Figure 29 shows the results for a 100 ns intensifier gate, collected from an ensemble of readings at different gate times. The rise and fall times of the signal are 30–40 ns, a little longer than the sum of the timing uncertainty and the laser pulse duration, indicating that the intensifier goes from full off to full on in a few (10–20) nanoseconds. The minimum recommended gate time for this system is 100 ns. Shorter gates can be used, however, the transistor-transistor logic (TTL) pulses start to overlap, which can stress the system, and the gain curve becomes compressed. Given the timing uncertainties and variable observation distances, the peak of the gain curve may not coincide with the arrival of the light at the intensifier if shorter gates were to be used. The top of the curve in Fig. 29b shows that stability is within ±10%. The slight dips at 640 and 670 ns are likely part of a fixed pattern. Tests with longer cables between the intensifier and the HVPS showed lower stability due to ringing caused by slight impedance mismatches in the cables. Given the expected reproducibility of the pattern, stability between observations made under the same conditions are generally better than the standard deviation across the plateau in the figure. The zero point of the timing units on the x-axis in Fig. 29 is arbitrary. The laser actually fires at around t=660 ns in every case, and the delay is adjusted to move the 100 ns window around the time the light arrives at the detector. For a target ∼2 m distant, it takes ∼33 ns for the light to travel from the laser to the target, back to the telescope, and down the 6 m optical fiber to the BU spectrometer. A delay setting of 690 ns results in just catching the Raman signal as the gate opens completely. At the other end, a delay of 590 ns results in the intensifier observing the Raman signal just as the 100 ns gate is starting to close. The BU controls the firing of the laser, and the zero point of the intensifier delay (t=0 extrapolation from Fig. 29) was designed to occur before the laser fires. A number of details are covered during the sequence containing the laser pulse and the ICCD gate. First, the intensifier HVPS must be powered on, the CCDs must be powered on, and the laser capacitors must be charged. If the CCD is not yet integrating (i.e., for the 4Page 36 of 87 R.C. Wiens et al. Fig. 30 Operating states of the SuperCam BU as commanded by the rover. POR =power-on reset, which is first achieved by providing power to the instrument first of several laser pulses to be collected on a single CCD exposure, or using single-shot exposures), then the pixels are continuously being dumped to the dump drain near the serial register. The CCD exposure is started and a “fire-the-laser” pulse is sent on a discrete line to the laser. The signal is received at the MU, which starts laser pumping and then triggers the Q-switch Pockels cell (Maurice et al. 2020). Simultaneously the MU sends a “laser sync” signal on a discrete line to the BU, which starts the delay timer for the intensifier, opening the gate by pulsing the cathode voltage at the proper time. The intensifier delay uses 24 bits; the gate uses 32, each bit representing a 10 ns time tick. These result in a maximum intensifier exposure of 42.95 s and a maximum delay of 167 ms. The intensifier voltage is refreshed every millisecond for exposures that exceed 1 ms. Tests with calibrated lamps showed excellent stability between successive spectra, with standard deviations of 0.5–1.0% for the reflection spectrometers and in the 1.5–3.0% range for the transmission spectrometer for 30 spectra taken at room temperature. Integration times for the reflection spectrometers were 5 ms, integrating across the center 16 vertical rows. The gate for the transmission spectrometer was 10 µs, with a gain setting of 2500. The light source was an Energetiq EQ-99 lamp positioned at a distance of 5 m, as described in Sect. 7.1.1. 3.3.3 Power Consumption The SuperCam BU consumes 12.0 W when idling. Additional power is used by the CCDs, the HVPS, and the intensifier when they are operating, bringing the maximum BU power to 19.1 W. The TECs consume an additional 24.0 W. 3.4 Software The BU flight software consists of a VxWorks 6.7 operating-system kernel and the custom SuperCam application. There are two copies of the kernel and application stored in radiation-tolerant magnetoresistive random-access memory (MRAM). A boot loader resides in a radiation-hard programmable read-only memory (PROM) that controls the startup of the instrument flight software. Figure 30 gives a state diagram of the instrument. The SuperCam BU software starts by executing a boot loader contained in a PROM. Based upon the initialization signal sent to the instrument, the boot loader selects one of the two kernels and applications stored in MRAM. The boot loader performs integrity checks on the software, loads it into synchronous dynamic random-access memory (SDRAM), and starts executing the flight software. Once the flight software begins execution, it initializes itself, performs The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 37 of 87 4 a built in self-test, then reaches the idle state awaiting commands from the rover compute element (RCE). When a command is received, the instrument goes to an operate state and executes the command. Once the command is complete it returns to the standby state. Only one command can be executed at a time and the RCE must wait until the instrument is in the standby state before sending another command. The only exception is the abort command. The instrument flight software is designed to handle off-nominal situations. The instrument handles errors in commanding, such as invalid parameters or invalid commands, using reply condition codes as defined by the project. Instrument-specific off-nominal situations are reported using condition codes and status flags as defined in the instrument command dictionary. The instrument maintains two logs in non-volatile memory: an instrument error log and a command history log. If a software processing error occurs, a timestamp and an error code are recorded that indicate the nature of the error that occurred. Both logs can be placed into the instrument’s data buffer and transmitted to the RCE using specific dump commands and the transmit data command as defined in the command dictionary. During the mission, it will be possible to upload new software, including new kernels, or applications, using the LOAD_MEMORY command as defined in the instrument command dictionary. It is possible to write to any area of memory using this command. 3.4.1 Command Handling and Instrument Status Two key responsibilities of the BU software are to receive and process commands coming from the RCE, and to provide telemetry and data to the RCE. Some commands like CONFIGURE_CCD_REGIONS are executed quickly and only a command reply is sent to the RCE. Other commands like DO_FOCUS, which take much longer to execute, result in the BU sending a command reply followed some time later by a science data frame. Both responses to the RCE contain status flags and a condition code to communicate the instrument’s state. There are fifteen persistent status flags that the software maintains and are communicated to the RCE as part of every instrument transfer frame. These include flags for MU power, MU direct current (DC) converters, IR spectrometer ready, continuous-wave laser (CWL) ready, laser frequency doubler ready, shutter status, laser stack ready, laser HVPS on, sun safety, BU HVPS state, and CCDs almost ready and ready. Some of the status flags are for MU functions that are explained in the companion paper (Maurice et al. 2020, this journal). If the instrument software detects a fault, the software will alert the RCE by setting the SEND_EVR, SYS_ERROR, or SHUTDOWN status flags depending on the severity of the fault. In addition to status flags, there are 30 condition codes that the instrument software can use to communicate status with the RCE. These range from indicating there was corruption of the received command, to an error communicating with the spectrometer electronics, or an error communicating the MU. Instrument commands represent one level in a large set of nested levels. Figure 31 illustrates the fact that the rover uses a higher level of (spacecraft) commands, such that a spacecraft command may parse several SuperCam instrument commands. Higher levels exist in the command structure such that a person using the Component-based Campaign, Planning, Implementation, and Tactical (COCPIT) tool for planning the rover operations may be able to use a template to carry out a significant portion of the remote-sensing activity for a given time frame. The instrument BU software only responds to instrument level commands. Those commands are combined to complete a higher level function. As an example, the following simplified sequence of instrument commands could be used to perform a LIBS collect: 4Page 38 of 87 R.C. Wiens et al. Fig. 31 Illustration of the nested nature of the commands that operate SuperCam, spanning from instrument-level commands (I-cmds) to spacecraft-level commands (S-cmds) up to components, which operate more than one instrument •COLLECT_SOH •MU_SEND_HOUSEKEEPINGS (housekeeping data generally consists of SOH and status flags) •CONFIGURE_CCD_VERT_TIMING (this command and the next two configure the vertical and horizontal timing of the CCD exposures, and the specific rows from which to collect data) •CONFIGURE_CCD_HORZ_TIMING •CONFIGURE_CCD_REGIONS •CONFIGURE_HVPS •CONFIGURE_INTENSIFIER •MU_DC_DC_CHAIN_ON (turns on the DC converter to enable various MU functions) •COLLECT_SOH •MU_SEND_HOUSEKEEPINGS •MU_CONFIGURE_LASER •DO_SPECTRA (fires the laser and collects the spectra using the specified laser, intensifier, and CCD configurations) •COLLECT_SOH •MU_SEND_HOUSEKEEPINGS •XMIT_DATA (sends the resulting data to the rover) The BU software has been designed such that all commands received and all data replies like state of health (SOH) or spectra are assembled into a single generic buffer in the instrument. Data markers and data lengths are used to identify each segment and make the generic buffer parsable. The entire generic buffer is transferred at once to the RCE using the XMIT_DATA instrument command. Section 4.4 provides detail on the formatting of the data product generated by the instrument. 3.4.2 MU Commanding, Focus Management, and Monitoring Other responsibilities of the BU flight software include powering and commanding of the MU, focus management, and monitoring of the MU. Instrument-level commands from the The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 39 of 87 4 RCE intended for the MU are in many cases just a pass-through for the BU software. However, there are times when the BU must modify command arguments before forwarding the command, and there are times such as the DO_FOCUS command where, as part of the command execution, the BU software will internally generate an MU-specific command. The commands and MU replies are sent over a UART data link between the BU and MU. Similar to the interface with the RCE, the MU indicates status in its command and data replies to the BU. The BU flight software is responsible for maintaining the state of the MU focus position. This includes keeping track of the focus position in motor steps (table position), converting the table position to distance to the target in millimeters, converting command arguments in millimeters to motor steps, calculating slight focus offsets needed to optimally perform either RMI or spectroscopy, and maintaining the state of the sun-safe flag, indicating that the focus position is not near infinity (Maurice et al. 2020), which would be dangerous for the optics if the sun passed through the FOV. The focus motor position is held in the BU’s nonvolatile memory. When the focus motor position changes either due to a manual move or an autofocus, the MU communicates the change to the BU. If the focus position maintained by the BU is deemed to be inaccurate, the proper position can be reestablished by driving the motor to a limit switch. There are seven components on the MU that the BU monitors to ensure they are within the allowable flight temperature (AFT): laser stack, laser doubling crystal, focus table, CWL laser, infrared spectrometer acousto-optic tunable filter (IRS AOTF), and IRS TEC hot side, and LPVS. The BU checks the temperatures every time an MU_SEND_HOUSEKEEPINGS command is sent, and, if there is no other command being executed, the BU will automatically send the command every ten seconds. If a temperature is found to be out of AFT, the BU will set an error status flag and communicate it to the RCE when the next command is received. 4 Details of Operation 4.1 Overview In general, SuperCam is designed to operate in a similar manner to ChemCam, which takes images and performs spectral analyses on a series of points arranged in a line or grid, referred to as a “raster.” It uses an autofocus routine and takes an RMI image at the beginning of each raster and occasionally thereafter, as required to maintain good focus and to cover the raster area with context imaging. Upon closer look, there are many more details involved in SuperCam’s operations due to the larger number of types of spectral observations and the increased complexity of some of them. Figure 32 shows a general flow diagram for operation of SuperCam. The figure and the description below includes all of the types of spectroscopy, but observations using only one type of spectroscopy (or only imaging), or a different combination or a different order of observations, are all equally feasible. Target acquisition, preparation, and RMI imaging: A target is selected by the science team, and its coordinates are uplinked to the rover as part of the sequence of commands. The target could also be selected by AEGIS (Autonomous Exploration for Gathering Increased Science), an onboard algorithm currently being used on MSL for autonomously selecting ChemCam targets (Francis et al. 2017). It is also expected to be able to perform closedloop pointing based on pattern recognition by the rover computer operating on Navcam or SuperCam images. Another possible option is to use blind targeting, in which the instrument 4Page 40 of 87 R.C. Wiens et al. Fig. 32 Flow diagram for typical SuperCam operations. This illustration includes all of the spectral techniques, though it is not necessary to use all techniques on a given observation, nor to use the order of taking spectra shown here. See text for description is pointed at a fixed location on the ground in rover coordinates without any knowledge of the terrain (Maurice et al. 2016). This was done on ChemCam prior to AEGIS. In any case, the rover is commanded to turn SuperCam on and to send it the proper commands for the sequence of analyses. Once SuperCam is on, it is commanded to do several things in preparation for analysis (Fig. 32): Assuming that LIBS, Raman, or VISIR spectroscopy will be done, the TECs will be turned on to start pre-cooling the CCDs. A pre-cooling time of 15–20 minutes is expected to be used prior to the start of taking spectra. If LIBS or Raman spectra are planned, the Nd:YAG laser will be warmed to its operating temperatures of −15 ◦C for the stack and −10 ◦C for the frequency-doubling crystal (Maurice et al. 2020). Finally, if the CWL will be used for autofocus, it needs to be warmed to −10 ◦C. The preferred autofocus method uses a series of RMI images, for which no warming is required (Maurice et al. 2020). During thermal equilibration, the RSM turns toward the proper coordinates for the target. Autofocus is the next step, followed by an RMI image centered at the first point of the raster. After the RMI data have been transferred, spectra are taken on the first point. LIBS: Typically, if LIBS plus other spectra are planned, we expect to perform LIBS first, as the LIBS shock wave removes dust from the target, clearing the surface for the other techniques. The laser pit is much smaller in diameter (∼250–400 µm; Maurice et al. 2020, this journal) than the FOVs of the Raman, VIS (both 0.74 µrad), and IR (1.15 µrad) observations. At a target distance of 2.5 m, the area of the LIBS laser pit represents ∼2.5% of the FOV for Raman and VIS observations, and ∼1% of the FOV of the IR observations. Because of this, the damage caused by the laser is expected to have negligible impact on the other observations (Fau et al. 2019). Instrument commands prepare the laser and spectrometers. Both active (with laser) and dark (background, without laser) LIBS spectra are taken at 3 Hz; the darks will be subtracted from the active spectra on the ground. We plan The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 41 of 87 4 to use ∼30 laser shots for each standard LIBS observation. A separate spectrum is returned for each laser shot, while we downlink only statistics (mean, median, and standard deviation for each channel) from the 30 dark collects. The separate active spectra potentially record changes as the laser profiles up to ∼100 µm into the sample (assuming 30 laser shots). Microphone recording of the LIBS plasmas is optional; acoustic data are sent to the rover after the recording is finished. After the LIBS data have been transferred to the rover, the next spectral activity can begin. Raman/TRL: For Raman and TRL spectroscopies, changes in the MU include deploying a beam stop to block the LIBS laser path and preparing to use the Pockels cell to cause the laser beam to be frequency doubled, to produce a green 532 nm beam (Maurice et al. 2020). Instrument commands prepare the transmission spectrometer. Both active and dark spectra are taken, usually using 100–200 laser shots at 10 Hz. For Raman spectroscopy, the active spectra can be averaged on board, as no changes to the sample take place as a result of successive laser pulses. In that case, the mean, median, and standard deviation are returned for active and dark collects, each. For TRL, short bursts of laser shots are fired, each burst with a different time delay between the laser and intensifier (see Sects. 4.2.2 and 7.4)to record a time spectrum of the intensity of the laser-induced luminescence. Once the Raman or TRL observation is complete, the LIBS laser beam stop must be removed; there is a time limit of two minutes for its deployment (Maurice et al. 2020). VISIR: For IR observations, the IR TEC in the MU must cool the IR photodiode ∼50 ◦C below the MU temperature before the observation. VIS spectra are taken using the VIO and transmission spectrometers in the BU, followed immediately by IR spectra (Maurice et al. 2020). The SuperCam IRS is wavelength-scanning, and typically takes <80 s to acquire a spectrum plus darks. The VIS spectral acquisition is done in several seconds at most. Typically, once these spectra are taken, the RSM moves to a new point on the same target, developing a raster pattern, usually with point-to-point spacing of 0.5 to 2 mrad (=1.25 to 5 mm at a typical distance of 2.5 m). Planned rasters include 1×5and1×10 line scans, and 2×2and3×3 grid patterns. As the raster proceeds, an additional autofocus may be needed to ensure proper focus is maintained, especially in the case of a vertical raster, where the distance to the target is more likely to change. Another RMI image is taken at the conclusion of a raster, and depending on the angular distance that the raster covers, an additional RMI image may need to be inserted in the middle to ensure imaging documentation of all raster points. Typically an RMI image should be taken every 8 mrad or less to keep the spectral observation points within its FOV (Table 1). This process can be repeated for successive targets. Several clean-up activities are performed to conclude the observations of one or several targets, shown in Fig. 32.Theseinclude moving the focus stage back to the sun-safe position, and performing a controlled warm-up of the IRS TEC. After that the instrument is powered off. The three critical resources for Mars operations are time, energy, and data volume. Tracking and optimizing each of these is important for efficient operations. A notional power profile is provided in Fig. 33. In this example, the BU TECs are powered on for 12 minutes before the MU laser warming is started. Activities that follow are CWL autofocus, RMI imaging, LIBS +Microphone, Raman, and VISIR, followed by a final RMI image, and ending with controlled shut-down of the instrument. Just one point is observed in this example, but using several spectral techniques. The turn-on and thermal equilibration takes over 15 minutes, while the analysis portion of the sequences takes a little under four minutes. Given the relatively long time it takes for turn-on and thermal equilibration, it is useful to spend enough time to make more observations once the effort has been made to prepare the instrument. The MU is required to be able to operate two hours at a time. After that, it needs to 4Page 48 of 87 R.C. Wiens et al. data, and auxiliary data (e.g. the position coordinates of the RSM) or the command history. EDRs containing science data will be processed and calibrated by a specific pipeline, and the resulting products are called clean data records (CDRs). 5 Model Development and Environmental Testing The overall plan of development of SuperCam emphasized early validation of new technologies, particularly the transmission spectrometer and its HVPS as applied to Raman spectroscopy, followed by an environmental qualification model (EQM) and flight model (FM). To facilitate the early validation, an engineering development unit (EDU) was developed and tested. Its MU consisted of a laser and associated electronics, telescope, RMI, and an infrared spectrometer that was fiber-coupled to the telescope. The EDU BU consisted of electronics, demultiplexer, reflection spectrometers, and a transmission spectrometer. TEC cooling of the EDU CCDs was done with commercial units via a non-flight copper bar from the CCDs. The EDU transmission spectrometer had only two spectral bands instead of the three in the EQM and FM, and so it had low resolution. Integrated testing of the EDU SuperCam instrument was done in the spring of 2016 at LANL, and was used to verify a number of details, including (a) overall coupling of the BU and MU, including the timing between the laser and spectrometers; (b) performance with the reflection spectrometers that was comparable to ChemCam; (c) basic Raman and TRL functionality and throughput; and (d) scientific studies of TRL spectra (Ollila et al. 2017). The latter was feasible because the luminescence peaks are often broad, and are thus relatively insensitive to the resolution of the spectrometer, which was not optimized at this early development stage. The EDU did not use flight-qualified electronic parts, and it was not built for environmental testing. Environmental testing was instead only carried out on new sub-assemblies, such as the transmission spectrometer, before integration into the BU. After the EQM was built, the MU EDU was stripped of its laser and infrared spectrometer. A special BU simulator was reconstructed from spare parts, with no optics. The two units were delivered to JPL in 2017 as the SuperCam Test Unit (TU), which supported verification and validation (V&V) activities and operations training. This unit will be mounted on the vehicle system test bed (VSTB) at JPL and will be maintained by the team for the duration of the mission on Mars. The BU EDU is planned to be used as a calibration model in the laboratory. A new MU will be rebuilt for that. The most significant change between the EDU and EQM BU was the transmission spectrometer (Sect. 3.1.4), which was redesigned with the dichroic beam splitter and compound grating to yield three wavelength bands instead of the two that the EDU had. This improved the resolution to meet the 12 cm−1requirement. The EQM BU underwent environmental testing in LANL, while the EQM MU underwent testing in Toulouse before integrating the units. Testing of the BU identified three issues: slight motion of the reflection spectrometer optics during vibration and shock, under-performance of the TECs, and a component failure in the HVPS after its cold thermal cycle. The reflection spectrometers were only slightly redesigned from ChemCam. For both instruments the optics are held in tension against a three-point mount. The problem was never seen on ChemCam, and for SuperCam, the only occurrences were at the qualification levels, with extremely minor movement at the flightacceptance instrument vibration, and no movement during rover shock or vibration. The HVPS component failure led to re-analysis of the mechanical stresses on the components, and some re-positioning of components. The TECs were not modified on the FM relative to the EQM, but a number of improvements were made for the flight TEC cooling system to The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 49 of 87 4 insulate it further from the environment and to spread the heat to the RAMP better. During testing of the EQM BU, the wavelength dependence on temperature for all three spectrometers was shown to be much less on SuperCam (Sect. 7.1.3) than on ChemCam (Wiens et al. 2012). The EQM is fully functional. The EQM BU and MU required some rework post-FMdelivery to match the FM performance. They will be used as a calibration model, particularly to complement the LIBS spectral database of known compositions, and to acquire a library of time-resolved Raman and luminescence signatures. Thus the EQM is critical to achieving SuperCam’s overall scientific mission. The flight BU has very few changes relative to the original EQM other than those mentioned above. Thermal, Mars-pressure testing of the FM identified one new issue: The intensifier created flashes of light at random times when at Mars pressure. This was due to an imperfection in the multi-channel plate and surrounding potting. Potting of the tubes was most successful on the EQM tube, and so that tube is being used for flight. It was also found that the EQM intensifier tube had a small tendency to create flashes at Mars pressure in N2 gas. Nitrogen is the test gas of choice, as it is easier and safer to use in a thermal chamber, and represents an over-test relative to the Martian atmosphere, for which breakdown does not occur as easily. The flashes were likely due to charge buildup on the output window of the intensifier due to electric field leakage through the unit. Extensive testing was carried out in Mars gas, showing that there were no flashes under Mars conditions using normal operating procedures. As part of the requirements for delivery the FM BU underwent a series of environmental qualification tests including random vibration and thermal vacuum, which were conducted using Los Alamos facilities. Random vibration testing was performed in each of the three axes at protoflight levels for one minute per axis. After vibrating each axis a visual inspection was performed and the accelerometer data was reviewed to assess any mechanical change to the unit. A comprehensive functional test was also conducted before the test, between axes, and after the test to determine any performance change after exposure to the vibration environment. The BU then underwent 204 hours of thermal vacuum testing. The chamber supported bulkhead feedthrough connectors that allowed power, command and telemetry signals, and an optical fiber to be coupled from support equipment outside the chamber to the unit in the chamber. When installed on the rover the SuperCam BU has its baseplate attached to the underside of the RAMP such that the thermal sink is above the CCDs. The orientation affects the performance of the heat pipes, so we needed to mimic the BU orientation in the rover for realistic thermal tests. For this, a secondary platen was mounted on legs inside the thermal vacuum chamber with its mounting surface facing down. The baseplate of the BU was then bolted to this surface. To complete the test setup a number of thermocouples were attached to the unit. During thermal vacuum testing the BU completed three thermal cycles across its protoflight temperature range of +60 C to −50 C. The first thermal cycle was conducted at high vacuum below 1 ×10−5Torr to replicate cruise conditions. The remaining two cycles were conducted at Mars pressure (∼7 Torr) using gaseous nitrogen. Each thermal cycle included long soaks at the maximum and minimum temperatures. During the soaks and at various intermediate temperatures, testing was conducted to verify the functionality of the electronics and the effectiveness of the TECs to cool the CCDs, and to assess the performance of the spectrometers. During the high-vacuum cycle the decontamination heaters were tested to verify their ability to maintain the optics at the desired temperature. Earlier in the project, pyro-shock testing was performed on the EQM BU using Los Alamos facilities. Results from this test were used to approve the BU FM without testing, 4Page 50 of 87 R.C. Wiens et al. limiting potential mechanical stress on the flight model. Additionally, an electromagnetic compatibility and electromagnetic interference (EMC/EMI) test was performed using the EQM BU coupled with the EQM MU at a JPL facility. Results from the EMC/EMI testing showed that the BU exceeded maximum radiated signal levels in a couple critical frequency bands. This was addressed on the FM by adding some additional EMI shielding to specific areas of the unit. Final testing of the instrument on the rover verified that SuperCam meets requirements, and can be used during UHF communications, which is an improvement over ChemCam. 6 Performance Testing The test and validation campaign of SuperCam was constrained by the availability and capabilities of the units and models, as described below. Table 8shows a summary of the testing of various models and configurations. The EDU was useful for testing of functionality, despite the fact that some of the EDU subsystems were not fully flight-like, and some were not present. For example, the microphone was not part of the EDU, and as mentioned, the transmission spectrometer only provided low resolution. The EQM was a better testbed, as it was flight-like in nearly every sense. It provided a stable platform for studies of Raman performance, although a few aspects of its performance (alignment, focus at cold temperatures) were not as good as that of the FM. Thus, Raman and TRL studies could be made at close distance (e.g., 2–2.5 m), but not at distances >5 m. Raman and TRL observations were also carried out with the FM units, often re-testing samples analyzed with earlier models. For the flight model, a problem arose in the fall of 2018 that required a complete rebuild of the MU (Maurice et al. 2020, this journal). Because of that, the flight MU was not ready until after the required delivery date of the BU to the rover. Validation of the BU in a realistic environment, development of a LIBS spectral library, and initial calibration of the instrument response function (sensitivity vs. wavelength, or IRF) of the system had to be done on a combination of FM BU +EQM MU. To validate critical MU functions such as the autofocus offsets (offsets as a function of temperature for different relative focus points for LIBS, RMI, and the CWL autofocus mode; Maurice et al. 2020), after delivery of the FM BU to the rover, the EQM BU was sent to Toulouse to support testing in the configuration of EQM BU +FM MU. Finally, as shown on the right side of Table 8, the last flight portion of the instrument was the FOC, the flight version of which never left JPL. So when the BU and later the MU were integrated onto the rover, all of the flight parts of the instrument were finally together. Because of the short time available in the final flight configuration, results (Sect. 7) will be presented from various configurations, as appropriate to represent the overall performance of the flight instrument. Because we could cross-check the FM units with the EQM units, and because the EQM units were generally very similar to the flight units, we believe the differences in characterization relative to the complete flight unit were relatively minor and limited to those aspects noted here. 6.1 Flight-Model Body Unit +Engineering Qualification Model Mast Unit Testing at LANL From 6 to 23 April 2019, the FM BU was tested in its final configuration with the EQM MU for performance and software validation, and to develop spectral libraries. Both the BU and MU were housed in a thermal chamber with dry nitrogen at ambient pressure, and at −10 ◦C to optimize the laser energy. For LIBS, a Mars sample chamber pressurized to 5.8±0.2Torr The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 51 of 87 4 Table 8 Summary of testing of the SuperCam models BU MU Location of testing #Data files Study Raman, TRL Validate functionality Cross calibration LIBS spectral library Autofocus offsets Realistic thermal environment Flight fiber EDU EDU LANL TRL X EQM EQM LANL 5534 Raman X FM EQM LANL 3100 X X X EQM FM IRAP 743 X X X FM FM JPL 1119 X X X X 4Page 52 of 87 R.C. Wiens et al. Table 9 Numbers of unique samples of rocks, minerals, and mineral groups observed by the FM BU and EQM MU during the validation period Rock/mineral type LIBS Mineral/type Raman VISIR Andesite 14 Carbonate 17 5 Trachyte 4 Silicate 12 6 Anorthosite 6 Plagioclase-feldspar 9 1 Basalt 33 Sulfate 8 4 Pyroxene 7 Olivine 6 1 Plagioclase-feldspar 3 Phosphate 5 Olivine 6 Phyllosilicate 4 9 Dolerite 3 Pyroxene 3 1 Gabbro 8 Serpentine 2 3 Norite 5 Mica 1 Syenite +granite 5 Amphibole 1 Carbonate 8 Diamond 1 Sulfate 7 Fluorite 1 Fe, Ti, oxides, banded iron 8 Metal oxide 1 9 Mn oxides 13 Sulfide 1 Metamorphic, Archean sediments 49 Nitrate 1 Clastic sediments including shale 91 Perchlorate 1 1 Phyllosilicate (kaolinite, smectite) 6 Miscellaneous other 31 Organic, mineral mix 51 15 SCCT 25 SCCT 14 13 TOTAL 332 TOTAL 138 68 with CO2was used to simulate the Mars atmosphere, which affects the plasma properties. During observations of samples in the Mars chamber, the window of the instrument thermal chamber was removed, and a pipe was installed between the instrument thermal chamber and the window of the sample chamber. In this way, there was a single window between the instrument and the samples, as will be the case on Mars. Raman, TRL, RMI, and VISIR observations were made with the instrument-chamber window in place and the samples in open air. The primary purpose of the LIBS observations was to establish a library of spectra of a diverse suite of geological materials, mostly at one distance, and then a smaller library at several distances. Experience with ChemCam has shown that a large number of standards are needed for quality calibration using multivariate methods for all of the major elements and a mixture of methods for minor and trace elements (Wiens et al. 2013; Clegg et al. 2017; Payré et al. 2017). Depth profiles were also produced using a large number of laser shots (Sect. 7.2.4). Table 9gives a summary of the LIBS observations. Three observations were made of each standard, with each observation consisting of 30 laser pulses. Each laser observation has a corresponding dark spectrum of 30 collects. Standards were loaded in the sample chamber on a turret that holds approximately fifteen 30-mm diameter pressed powder pellets, which is the physical form of the LIBS standards. With each turret load, a small titanium plate was installed and observed. Cross comparison of the Ti observations from one turret load to another can be used as a check against problems with focus, clipping of the laser beam in the pipe, laser energy, or ambient pressure. All of the LIBS targets were The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 53 of 87 4 observed at 2.85–3.00 m distance. The SCCTs were observed again at 1.55 m (the distance of the SCCTs on the rover) and at 4.25 m. A full list of the standards and the calibration parameters derived from the library will be published elsewhere. Raman and VISIR spectra were taken to validate the protocols and to obtain representative spectra from the instrument. Table 9lists the different unique targets that were observed by these techniques that were carried out with the FM BU +EQM MU. A few TRL spectra were also recorded. In addition to pure minerals, some mineral mixtures were tested, as well as mixtures of organic species with inactive binder materials. As mentioned above, many more such observations were carried out with the complete EQM instrument, especially for TRL spectroscopy (Ollila et al. 2017). The results are presented in Sect. 7. 6.2 Engineering Qualification Model Body Unit +Flight-Model Mast Unit Testing in Toulouse The rover schedule required the BU to be installed first, leaving several more weeks to complete and deliver the MU. During this time the EQM BU was sent to France to facilitate final testing of the FM MU. The primary objectives were to verify the communication between the two units and to perform safety checks by the BU software that included monitoring of MU temperatures and setting the correct status flag if out of range, and verifying that the BU set the appropriate motor speed, number of steps, distance to limit switches, pumping current, and laser rest time (30 s) between consecutive bursts. We also checked that the BU reported the correct condition codes, and the sun-safe status in all configurations, even if communication with MU was lost. The secondary objectives were to verify instrument functionality, and to characterize some key parameters that are unique to the FM MU. These include CWL-to-LIBS and CWLto-RMI focus-offset algorithms (Table 8) that were subsequently coded into the FM BU software. The test also checked the quality of the LIBS focus and the number of spectrometer CCD rows to use to avoid saturation at short distance. Some microphone noise tests were also performed. Finally, this configuration provided an opportunity to acquire some additional calibration data which could be used along with data from the EQM MU +FM BU tests at LANL (see Sect. 6.1). LIBS spectra were acquired at 1.56 m and 3 m, and compared to ChemCam spectra on Mars, confirming the comparative results to be presented in Sect. 7.1. A few tests at 7 m showed better SuperCam performance than ChemCam at that long distance, but the actual long-distance capabilities will have to be shown on Mars. Time-resolved Raman spectra were also obtained from gypsum. TRL spectra were obtained from apatite and showed REE and Mn2+ detections at ∼100 ppm levels for a 0.5 ms integration window. After the successful conclusion of these brief tests, the FM MU was delivered to the rover. 6.3 Testing on the Rover Extensive performance testing of the end-to-end FM SuperCam instrument on the rover took place in several campaigns. The general details are given in Table 10. Essentially all of the functions of SuperCam were tested at least once (except RMI z-stack). Alignment tests were done before and after dynamic and thermal environmental testing to verify that the FOV of the transmission spectrometer is well aligned (at the proper temperature) within the green laser beam, as needed for Raman spectroscopy (Maurice et al. 2020). These were done using a Spiricon camera and neutral density filters. Checks were also made of the CWL and LIBS 4Page 54 of 87 R.C. Wiens et al. Table 10 Testing of SuperCam on the rover in 2019 to early 2020; see text for explanation Campaign July August STT EMI/EMC December January Tests CWL autofocus RMI autofocus Autofocus Instrum. on Code update Dark collects RMI autofocus IRF LIBS RMI Alignment RMI HDR Alignment Thermal control LIBS +MIC IRF IRS side B LIBS (3 m) TRL sweep (5 m) MIC alone LIBS +MIC (pulsed mode) Raman (3 m) Passive VIS Raman RMI (camera model 2, 3, 5 m) 98-point IR scan VISIR RMI 2D spectra Additional testing (survival only) included rover-level vibration and RSM-release shock laser beam positions relative to the RMI FOV. Observations were made to support modeling of the pointing of SuperCam relative to Mastcams and Navcams (parallax and offset). This camera model is critical to being able to command the correct position to hit a target with SuperCam based on the location of the target in the Navcam or Mastcam images, all from a given rover position. RMI images taken in support of the camera model and other testing revealed a small number of unexpected particles near the focal plane of the RMI imager. Investigation indicated that these particles reduce the light only slightly in small areas of the image, and they can be removed from the images as part of the flat-field correction. The system thermal test (STT; Table 10) covered ∼2 weeks in October 2019 with the rover in a large thermal/vacuum chamber at a range of temperatures to simulate Mars. Most of the testing was done at ∼10 Torr of N2to simulate Mars thermal conditions, with some testing to simulate the cruise conditions (vacuum). Although the RTG power supply was not installed on the rover, the RAMP was heated as if it were present, to provide realistic temperatures for the instruments and other components in the rover. Two sets of plates with geological targets were mounted in the chamber. One was on the ground near the rover (2.6 m from SuperCam) and another plate was 4.6 m from the instrument, hanging on the wall. The SCCTs on the back of the rover constituted a third set of targets. In addition, SuperCam took one RMI image of the targets installed by the SHERLOC team. STT provided the first realistic test environment for the IR spectrometer, and also for the LIBS +microphone combination. All of the other main observation modes were carried out in STT: LIBS, Raman/TRL, VISIR, and RMI. Lastly, all end-to-end thermal control loops and associated hardware were confirmed in STT. The rover-level electromagnetic interference and compatibility (EMI/EMC) tests complemented other EMI/EMC testing done earlier at the instrument level. The test validated that SuperCam does not cause interference with the ultra-high frequency (UHF) antenna, and so the instrument can be operated at the same time as UHF uplink and downlink. December 2019 testing of SuperCam consisted of a minor software update and repeat observations of IRF and alignment. The latter two were important to verify that rover vibration, shock, and thermal testing had not adversely affected the optical system. The final IRF is given in Sect. 7.1.1. In January 2020, true “darks” were obtained by acquiring dark spectra with the lid on. Finally, we tested the backup IRS photo-diode, the RMI HDR mode and the microphone pulsed mode (Table 10). The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 55 of 87 4 7Results This section describes results of the integrated SuperCam instrument. Some additional results in terms of IR spectra, RMI images, and acoustic results can be found in the companion paper (Maurice et al. 2020), as these techniques are complete at the MU level and did not require an integrated instrument. 7.1 Overall Optical Properties We present some of the overall optical properties here. Radiometric information on the RMI is presented in the companion paper (Maurice et al. 2020).TheIRFoftheIRspectrometer is given in (Royer et al. 2020). 7.1.1 Instrument Response Function (IRF), BU Spectrometers Two different calibrated lamps were used to determine the IRF across the spectral range from 245–853 nm. For the violet, green, orange, and red spectral ranges of the transmission and VIO reflection spectrometers, a calibrated Labsphere integrating sphere and lamp assembly was used. This lamp provided uniform radiance generally at the 2% level or better across its aperture (∼100 mm) and for reasonable angles of a few degrees. The same lamp was used to calibrate ChemCam (Wiens et al. 2012). As a second source, a calibrated Energetiq EQ-99 laser plasma discharge lamp was used across the entire spectral range. The advantages of this lamp include a significantly brighter output at all wavelengths compared to the Labsphere, and a sufficient signal in the UV to allow the calculation of IRF. Overall, it is a good broadband source, but its output contains a few emission peaks, such as in the red range, and also ∼455 nm in the VIO range, that are not fully described in the manufacturer’s calibration data. Use of the Labsphere allowed an independent check of these regions. Other than the plasma emission peaks, cross checking of results obtained with the Labsphere and with the EQ-99 showed close agreement, within ∼2.5–3% across all but the UV spectral range, which is not covered by the Labsphere. Figure 36 shows the preliminary IRF for the BU spectrometers. This measurement was made with SuperCam mounted in the rover, with the system at room temperature (detectors at 29–32 ◦C). IRF observations were made shortly after integration onto the rover, and again near the end of the rover test period. IRF measurements were also made on the instrument test set-up at LANL, for the EQM MU and the FM BU, and also for the all-EQM instrument. At LANL, a calibrated fiber-fed Ocean Insight DH-2000 lamp was also used to calibrate the FM BU +EQMMUUVrange. As seen in Fig. 36, multiple observations were made using different numbers of integration rows for the reflection spectrometers (UV, VIO; see Table 3and Fig. 12). The reduction of integration rows from 200 to 40 and 16 allows the instrument to avoid saturation for nearby, bright LIBS targets, especially the rover calibration targets. The EQ-99 lamp was used for calibration of the UV range with the instrument on the rover. However, it was too bright for the 200-row integration, and so this curve was constructed from a scaling factor determined in the earlier IRF measurements that occurred shortly after integration. In this way, the 16and 40-row curves were scaled up to 200 rows as shown in Fig. 36a. For the VIO and transmission spectrometers, the Labsphere lamp was used for rover calibration, as it avoided alignment challenges incumbent with the EQ-99 lamp. For the data in Fig. 36c, the red window of the transmission spectrometer (∼715–853 nm) used a reduced set of 70 4Page 56 of 87 R.C. Wiens et al. Fig. 36 Preliminary instrument optical response functions for the SuperCam BU UV (a), VIO, (b), and transmission (c) spectrometers. The units are digital numbers (DN) per photon incident at the telescope aperture. Data were taken with the instrument mounted in the rover. For the reflection (UV, VIO) spectrometers, read-out of fewer CCD rows results in a lower optical response that can be used to avoid saturation for nearby targets. For the transmission spectrometer (c), a large range in response is needed to accommodate both bright LIBS signals and weak Raman signals. A log scale is used to present the gain settings planned for LIBS (2500) and Raman spectroscopy (3200). The response curves of the transmission spectrometer clearly show the three different (green, orange, red) optical windows, with dips in response at the transition regions. The exact position of the transition was selected to avoid any important LIBS emission peaks. ChemCam’s instrument response (Wiens et al. 2012)isshownfor comparison. ChemCam is limited to 14-bit numbers, in contrast to SuperCam’s 16 bits rows for integration. This trade accepts slightly lower IRF for better optical resolution in this region. The figure also shows that different intensifier gain DAC settings were used for Labsphere lamp observations with the transmission spectrometer (Fig. 36c). The two curves are for the gain settings planned to be used for LIBS (2500) and Raman and TRL spectra (3200). Observations were made at other gain settings, including 2100, 2300, 2600, 2900, 3200, and 3500, spanning a range of over three orders of magnitude in optical gain (see The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 57 of 87 4 Fig. 37 Gain produced across the SuperCam intensifier as a function of the digital-to-analog count value (0–4095). The typical DAC setting for LIBS is 2500; Raman spectroscopy uses a DAC setting of 3200, but these can be adjusted, depending on conditions, e.g., decreased to avoid saturation or increased to highlight a weak feature also Fig. 37). Other than scaling factors, these IRF curves at different gain settings appear identical, indicating that the IRF changes in magnitude only, and not in spectral distribution. Given that the telescope, FOC, and reflection spectrometers are similar to ChemCam’s, a comparison with that instrument is in order. ChemCam’s IRF was measured using 200 rows in all three spectrometers. That instrument’s highest sensitivity in the UV range is 8 × 10−4DN/photon between 280 and 300 nm (Fig. 36a; Wiens et al. 2012). SuperCam’s most sensitive region is at a little longer wavelength, between 295 and 310 nm, achieving 1.7× 10−3DN/photon using 200 rows of integration (Fig. 36a). The image pixels are effectively identical between the two units, but the digital representation of the pixel intensity is divided into 65536 steps on SuperCam instead of 16384 (see Sect. 3.3.1). At the low-wavelength end of the UV range, both ChemCam’s and SuperCam’s sensitivities drop similarly. In the VIO region, ChemCam achieves just over 1e-3 DN/photon in the 415–440 nm range, while SuperCam’s IRF reaches a maximum of ∼3.3e-3 DN/photon between 380 and 400 nm (Fig. 36b), the result of efforts to provide higher efficiency at shorter wavelengths. The maximum sensitivity in SuperCam’s UV and VIO spectrometers are short of the expected factor of four higher in the digitized signal. The SuperCam demultiplexer was redesigned to favor the Raman spectral range at the slight expense of all shorter wavelengths, so this is not surprising. A slight dip at ∼425.5 nm indicates the presence of the VIO CCD blemish, giving a maximum drop in instrument response of ∼4%. Because the signal is integrated over many vertical pixels for any given wavelength, after correction for instrument response, we do not expect to see any effect on the processed data. While the SuperCam transmission spectrometer is very different from the ChemCam reflection spectrometer that covers the same spectral range, we can still compare the IRF at the planned LIBS HVPS setting of 2500 with the IRF of ChemCam in this range. The latter achieves ∼1.1×10−3DN/photon, with a maximum at 670 nm. SuperCam’s IRF has two peaks around 1×10−2DN/photon, one around 580 nm and one near 660 nm (Fig. 36c). The 4Page 64 of 87 R.C. Wiens et al. Fig. 42 Total emission from target JA-1 as a function of distance for optimum performance. The flight instrument was only integrated at JPL, and the rover’s system thermal test—the only time the integrated instrument’s laser was cooled—did not facilitate a 7 m observation. A set of observations was made at increasing distances with the EQM instrument in a thermal chamber at −10 ◦C using a laser current of 140 A. The target, JA-1, was in air instead of in a Mars atmosphere. Focus was performed manually, as a z-stack. The results in terms of the change in total emission (sum of all channels) is shown up to 6 m (the longest distance performed, due to the size of the room) in Fig. 42.Some of the peaks saturated at distances closer than 2.8 m using all 200 rows of the reflection spectrometers and a nominal gain of 2500 on the intensifier, although the overall emission was not strongly affected, since much of the emission comes from the continuum with the target at ambient pressure. The overall trend with distance is close to r−3,whereris the distance. The trend is steeper than the r−2trend expected for passive observations due to the loss of the peak laser power density at longer distances (Maurice et al. 2020). Peaks were still relatively strong (some >3000 DN) at 6 m, and the FM MU +EQM BU demonstrated LIBS at 7 m (Sect. 6.2), so that achieving the 7 m requirement on Mars is well assured. Unpublished experience with ChemCam indicates that the limiting feature for detection at long distances is the ability of the laser to focus sufficiently well to optically couple with the target and produce a spark. If a spark is produced, the system is always sensitive enough to acquire its associated emission spectrum. (On Mars, ChemCam has succeeded in obtaining weak spectra from some iron meteorite targets as distant as 9.2 m; Johnson et al. 2020). Additionally, on SuperCam, for the spectral region of the transmission spectrometer, the gain can be boosted, which may increase the ability to detect peaks in the green, orange, and red portion of the spectrum at long distances. In terms of quantification, correction for a distance effect in the calibration is needed to provide reasonably accurate observations to 7m. 7.2.3 Time-Resolved LIBS SuperCam offers the ability to use time resolution for LIBS spectra in the transmission spectrometer, though not in the shorter wavelength ranges. The largest time difference within LIBS spectra is between atomic and molecular spectra. Our studies indicated that essentially all atomic emission is finished within 1 µs, while molecular emission continues over a longer The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 65 of 87 4 Fig. 43 Example of the ability of SuperCam to use time resolution to study LIBS processes. Shown are trends in the intensity of the CaF molecular peak measured at 603 nm as a function of the exposure duration up to 10 µs (a), and with up to 40 µs delay, on a log scale (b). In (a), all exposures start at a delay of 800 ns. In (b), all exposures have a duration of 10 µs period of time. To check the timing of the CaF molecular peak (Fig. 41c), LIBS was observed from a basalt target doped with nearly 50% CaF2at a distance of ∼2.3m;thistestusedthe EQM instrument. Two sets of data were taken. In the first set, the intensifier gate was set to 10 µs (long) and the delay was increased to see how much of the CaF peak remained after different times. The other data set used a constant delay of 800 ns (too long to capture all of the atomic emission) and different exposure durations. The results are shown in Fig. 43. The molecular emission peak persists to >40 µs (last data point in Fig. 43b) but most of the emission is gone much earlier; 6% of the emission remains at 5 µs and only 1% at 10 µs (fourth and fifth data points in Fig. 43b). Based on these data, it was decided for the mission to use a LIBS exposure duration of 10 µs with the transmission spectrometer. A relatively short LIBS exposure with the transmission spectrometer decreases the contribution from ambient light. On ChemCam, subtraction of ambient reflected light is a significant factor in this spectral range, and some solar features, especially the Balmer line subtraction for H at 656 nm, can cause problems if the surface albedo changes due to the excavation of a (dark) hole in soil (e.g., Schröder et al. 2015). Relative to ChemCam, which has an effective ∼8 ms exposure, the ambient light obtained with the transmission spectrometer is reduced by nearly three orders of magnitude, so the nolaser (“dark”) spectra will have much lower signal to be subtracted from the active spectrum. (The LIBS exposure durations for the reflection spectrometers are 1 time tick =34.133 µs, but the pixels are still active while being read, resulting in the same effective duration as ChemCam. However, ambient light intensity in the VIO and UV ranges are much lower than in the green, orange, and red ranges covered by the transmission spectrometer.) Several other time-resolution experiments were carried out on LIBS spectra with the transmission spectrometer. One question was whether it is feasible to remove the continuum by starting the exposure slightly later. Most of the continuum is gone within ∼100 ns. However, some of the most rapidly-quenching peaks start to disappear within that time scale, including sulfur peaks. Some spectra could be obtained showing S peaks and almost no continuum. However, SuperCam has some uncertainties on the order of 20 ns due to the two FPGAs that respectively operate the laser and the intensifier (Fig. 29 and Sect. 3.2). Because of this, for SuperCam LIBS data it is best not to try to remove the continuum with the intensifier gate. Other experiments may be of interest, such as temporally separating the H emission line at 656 nm from an overlapping C line (Schröder et al. 2018) or studying molecular emission lines of CaCl (Vogt et al. 2018) or CaO. Finally, it should be possible to collect atomic emissions largely to the exclusion of molecular emissions. This is potentially 4Page 66 of 87 R.C. Wiens et al. Fig. 44 Measured pit depths as a function of the number of laser shots for basalt (a) and dolomite (b)made with the EQM laser at a distance of 2.86 m in a simulated Mars atmosphere. Error bars show the standard deviations of three pits at each number of shots. Insets show the samples. Rows of pits can be seen in the dolomite sample (b) useful for observing one of the stronger P emission lines which is normally hidden within the CaF structure for minerals like apatite. 7.2.4 LIBS Depth Profiles LIBS has a special advantage of remotely profiling into rocks and soils, so that surface weathering or other surface coatings can be detected and characterized (e.g., Lanza et al. 2012,2015,2016). Several targets were profiled using variable numbers of shots per location with the EQM MU and FM BU; additionally, one depth profile was performed in ATLO with the flight configuration. The results may differ between the two because the EQM telescope’s primary mirror had a nickel layer between the structure and the reflective aluminum coating that was later discovered to cause deformation of the mirror at cold temperatures (Maurice et al. 2020). The EQM laser depth profile was performed with the instrument at −5to −10 ◦C, so the focus may have been somewhat poorer than on the FM, which does not have the Ni layer. Note that ChemCam has a Ni layer (Maurice et al. 2012) because the issue was not discovered until later. In any case, the EQM depth profile should represent a worst case in terms of the SuperCam flight depth-profile capabilities. We present the EQM results first. Two targets were used: a basalt and a sandy dolomite (Fig. 44, insets). The basalt sample is from the ∼1.5 Ma Black Point Lava Flow, which is part of the San Francisco volcanic field ∼40 miles north of Flagstaff, AZ (Ulrich and Bailey 1987); this basalt was also used for the LIBS depth profile experiments described in Lanza et al. (2012,2015). The dolomite sample is from the Moenkopi formation, Wupatki member, which composes the sedimentary bedrock that predates the lava flow in the same area (Stewart et al. 1972). Both samples were prepared as sawed billets that were not polished but were relatively flat. A series of LIBS shots were performed on each sample to produce pits with 5, 10, 20, 30, 50, 100, 150, 200, 300, 400, and 500 shots per location (3 pits each, e.g., Fig. 44b, inset). These were done at a standoff distance of 2.85 m under a simulated Mars atmosphere. Pit depths and volumes were then assessed using a Keyence VK-X100 3D Laser Scanning Confocal Microscope with a VK-X150 controller. Scanning heights were manually determined by focus at the surface of the sample through to the bottom of the crater. The Keyence MultiFile Analyzer was used to determine crater volume. The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 67 of 87 4 Fig. 45 Profile of a laser pit made with 500 shots in dolomite, and a side view of a slice near the surface. (a) Profile of the laser pit. (b) Backscattered electron image of the same dolomite sample in thin section; the natural exterior of the rock is seen near the top of the image, with the outline of the LIBS pit from (a) in black (the pit was not made at this location, but the scale of features can be compared). The weathered surface of the terrestrial dolomite is a different composition than the interior rock, and LIBS depth profiles sample and differentiate between exterior and interior compositions. The LIBS pit is approximately cone shaped but shows some variation due to the properties of the beam Both the pit depth and the volume of the pit increase rapidly at first, with the rate of ablation decreasing with increasing numbers of laser pulses (Fig. 44). This is expected because the surfaces in the pit are no longer normal to the laser beam as a generally cone-shaped pit develops (Fig. 45). The dolomite ablated more readily than the basalt, producing deeper pits. Pits made with 30 shots (typically used for ChemCam targets; Wiens and Maurice 2015; Maurice et al. 2016) ranged from 84–107 µm deep in basalt and 230–296 µm deep in dolomite. Pits made with 500 shots averaged 280 µm deep in basalt, while 500-shot pits in dolomite were nearly 1 mm deep (920–980 µm), with one outlier point at 400 shots that gave a greater depth of 1100 µm. These results confirm that the material properties of a target, in particular its hardness, play a large role in the total ablation depth that may be achieved with LIBS depth profile analyses. Conversely, the hardness of a target may be inferred by the relative pit size produced by LIBS analysis. The ChemCam FM achieved comparable results (Wiens et al. 2012). On the same samples as SuperCam reports here, the ChemCam laboratory unit achieved comparable results under similar analysis conditions, producing pits with depths ∼120 µm with 300 shots and ∼350 µm at 900 shots in the basalt, and ∼370 µm with 300 shots and ∼1000 µm at 900 shots in the dolomite (see images in Arvidson et al. 2014). Depth profile results on Mars may be affected by the stability of the rover and the RSM. For ChemCam depth profiles that used >150 spectra, a pause of over a minute was needed after 150 shots to transfer the spectral data to the rover. It was noted that in at least some cases in which depth profiles of >150 shots were made, the first few spectra after the pause showed stronger emission, thought to be due to a slight movement of the center of the beam on the target, possibly only by a few microns. Images of the resultant pits also suggested that there had been a very slight movement during the pause, causing a slight offset in the analysis location between 150 shot bursts. SuperCam will be able to complete 500 shots in a single burst, which will allow greater depth profiling capabilities than ChemCam regardless of mechanical stability. One depth profile was tested while the FM SuperCam instrument was mounted on the rover, during the system thermal test, with a pressure of 8.1 Torr of N2gas. The target (“ClinQzOrth”) consists of 49% clinoptilolite, 32% orthoclase, and 19% quartz, and was at 4Page 68 of 87 R.C. Wiens et al. a distance of 2.6 m. The laser was at −15 ◦C and used 140 A. A burst of 150 pulses was used, but it was at the same location as an observation using 30 pulses that occurred immediately before, so the pit was made from a total of 180 pulses. The maximum crater depth, measured with the Keyence microscope, is 194 µm, with a crater volume of 1.67 ×106µm3.The bottom of the pit slants slightly to one side; it is not clear if that is due to RSM movement during the depth profile or to the properties of the target or the laser beam. Note that the EQM-laser depth profile shown in Fig. 45 appears similar. 7.2.5 Quantitative Elemental Abundances from LIBS Calibration of SuperCam LIBS will be presented in another paper. Here a brief overview is provided. Pre-processing consists of subtracting a background (non-laser) spectrum, denoising, wavelength calibration, continuum removal, stitching of the overlapping spectral regions from the transmission spectrometer, and distance correction to provide radiance in photons/second/cm2/sr/µm as a function of wavelength. Further processing is done by masking and normalizing the spectra. Quantification will be done using either multivariate or univariate regression. Based on ChemCam experience, it is likely that a multivariate approach will be used for major elements and a univariate approach will be used for minor and trace elements. Quantitative elemental compositions will be reported for eight major elements and at least four trace elements. A spectral library consisting of 332 standards was developed with the FM BU +EQM MU, with targets at 2.86 m distance in a simulated Mars atmosphere. After pre-processing (in progress), the team will optimize and compare multivariate regression models for each element, similar to the methods described in Wiens et al. (2013)andCleggetal.(2017). For the trace elements, we expect to follow procedures outlined in Payré et al. (2017) and Cousin et al. (2020). Maintaining calibration over a large range of distances is challenging (Melikechi et al. 2014; Mezzacappa et al. 2016). Unpublished studies with ChemCam data indicate the need for distance corrections for data taken beyond 3.5 m with that instrument. Work is ongoing to apply distance corrections for the SuperCam data. The rover calibration targets play an important role. Once the rover is on Mars, cross comparison between spectra from Mars and equivalent spectra of replicate targets shot on Earth with a laboratory clone will allow channel-wise comparisons of the results. An “Earthto-Mars” correction will be applied to the spectra collected on Earth, as it has been for ChemCam (Clegg et al. 2017), to remove any differences between the instruments that remain after correcting for the instrument response. Additionally, the rover calibration targets allow checks of the accuracy of the quantification. 7.3 Raman Spectra 7.3.1 Overall Raman Results The SuperCam Raman technique operates remotely via a telescope using a pulsed laser and a gated detector set to 100 ns exposures. These factors distinguish SuperCam from typical commercial laboratory instruments which use a microscope and thus operate at short distances (i.e. ∼mm) with small spot sizes (i.e. ∼µm) and use a continuous-wave (CW) laser. Due to the large spot size of 0.74 mrad (the FOV of the spectrometer through the telescope), SuperCam’s Raman mode will likely probe several mineral phases simultaneously, depending on grain size. The Raman signal (lifetime ≈10−15 s) is effectively only produced during the excitation pulse, while other signals like luminescence (lifetimes from minerals The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 69 of 87 4 Fig. 46 Raman spectra of pure minerals taken using the FM BU and EQM MU. Spectra are averages from 100 laser pulses taken with a delay of 650 ns, a gate width of 100 ns, and a gain of 3200. Laser energy was measured at ∼8 mJ. The distance was 2.77 m. The talc is from an unknown locality while the apatite is from Durango, Mexico, the diopside from Tanzania, the oligoclase from Ontario, Canada, the gypsum from Durango, Mexico, the quartz from Minas Gerais, Brazil, the calcite from Mato Grosso do Sul, Brazil, the barite from Cumbria, UK, the hydromagnesite from Iran, and the olivine from San Carlos, AZ, USA. Spectra of the oligoclase, olivine, and talc have been multiplied by 4 to enlarge the peaks are generally >10−9s) are excited by the pulse but continue to decay long after the laser pulse is finished. The 100 ns gate synchronized with the laser pulse allows efficient rejection of interfering signals like mineral luminescence or daylight entering the telescope, thereby optimizing the collection of the Raman signal. Compared to CW instruments, SuperCam Raman benefits from signal intensification and filtering in the time-domain but uses considerably less excitation, and receives a much smaller fraction of the signal because of the distance. Representative Raman spectra obtained with the SuperCam EQM MU and FM BU are shown on Fig. 46. Various mineral targets were tested including silicates, phosphates, sulfates, carbonates and accessory phases (e.g., oxides). Most targets were used as raw single crystals but some were prepared as powder pellets. Several organic and natural rock targets were also tested during the campaigns. SuperCam’s spectra are generally high-quality for phosphates, sulfates, and carbonates with clear detection of internal molecular modes and lattice vibrations. Notably, (i) polymorphs can be unambiguously distinguished by studying lattice vibrations (e.g. calcite vs. aragonite), (ii) OH/H2O is clearly detected in relevant phases (e.g. gypsum, hydromagnesite, and talc) and (iii) spectral resolution is sufficient to detect differences in subtle Raman shifts of the main peaks due to compositional variations (e.g. Mg vs. Ca carbonates). Some silicates (e.g. olivine, quartz, diopside, and oligoclase) 4Page 70 of 87 R.C. Wiens et al. yield well-defined spectra as well, with similar information, while others are challenging to analyze due to low signal (e.g. some phyllosilicates). Accessory phases like opaque minerals are not detectable. Difficulty in detection is mostly due to the high value of a mineral’s optical absorption coefficient that prevents volumetric analysis and/or due to poor Raman efficiency (e.g., Fau et al. 2019); this is consistent with laboratory instruments. The spectra shown in Fig. 46 were taken under somewhat more favorable conditions than on Mars. First, the minerals were in many cases gem quality or collector’s versions. Minerals of this size and quality will not likely often be found on Mars. However, to challenge any concerns about these crystals being of non-representative sizes, grain-size studies are reported below. Secondly, the instrument was maintained at −10 ◦C. On Mars, the detectors will more likely be close to 0 ◦C. Additionally, the intensifier will likely be between 10 and 30 ◦C. It is not as sensitive to temperature as the CCDs, but we expect some increased electron background from the intensifier at these temperatures. Both of these factors will add to the noise floor of the instrument. 7.3.2 Grain Size Effects The intensity of Raman spectra has long been known to be affected by the grain size of the target (e.g., Schrader et al. 1991) and the mixtures of grains that are in contact with each other. Remote Raman spectroscopy is more sensitive to this detail than in-situ Raman spectroscopy, where plenty of signal is usually obtained (e.g., Pasteris and Beyssac 2020; Torre-Fdez et al. 2020). Several grain-size studies were carried out with SuperCam prior to delivery; we present the results of two such studies here. Selenite from the Glitter Mine, Utah USA, was cleaned with alcohol, partially crushed in a mortar and sieved through different mesh sizes. Reagent grade Mg sulfate was dissolved in water which was subsequently evaporated to produce large epsomite crystals. These were dried and partially crushed in a mortar and sieved through different mesh sizes. Grains were pressed into pellets and were observed at a distance of 2.25 m from SuperCam’s EQM with the instrument at <−10 ◦C for maximum laser energy. Figures 47 and 48 show the effects of grain size. In both cases, the fraction with sizes >500 µm yielded lower signal than the fraction between 250 and 500 µm. Grain sizes <45 µm showed clearly lower signal than the larger grain sizes, and in the case of selenite, a fraction with grain sizes <25 µm showed the lowest signal. If these relationships hold for other minerals, it suggests that the best Raman results will be obtained on sandstones with medium to fine grains, although coarse-grained sandstones may perform well too; siltstones and mudstones will give poorer Raman signals on Mars (e.g., Pasteris and Beyssac 2020). 7.3.3 Detection Limits and Distance Effects An important detail in SuperCam’s Raman detection capabilities is the fact that the laser beam covers a significantly larger area than the 0.74 mrad FOV of the telescope and spectrometer for targets at close distance. As distance increases, losses due to the decreasing solid angle of the telescope from the target become significant and temperature-dependent co-alignment of the laser beam and spectrometer FOV affects the signal (Maurice et al. 2020). Here we discuss the number of Raman photons detected from a test made at different distances using the EQM instrument. The rate at which Raman photons are generated by a given sample is governed by (i) the differential Raman scattering cross-section (i.e. cm2sr−1molecule−1) of a target molecule, The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 71 of 87 4 Fig. 47 SuperCam Raman spectra of pressed-powder pellets of different grain sizes of selenite gypsum (Glitter Mine, Utah), observed with the EQM at 2.25 m distance integrating 100 laser pulses. Wavenumbers are approximate. Insets show closer detail of the fingerprint (a) and water regions (b) of the spectrum Fig. 48 SuperCam Raman spectra of pressed-powder pellets of different grain sizes of epsomite, observed with the EQM at 2.25 m distance integrating 100 laser pulses. Wavenumbers are approximate. Insets show closer detail of the fingerprint (a) and water regions (b) of the spectrum 4Page 72 of 87 R.C. Wiens et al. (ii) the number of molecules of that species that are illuminated, and (iii) the transmitted energy of the excitation source. To gauge the number of Raman photons collected at SuperCam’s entrance aperture and at the photocathode of SuperCam’s intensifier, a crosscalibration study was performed at the University of Hawaii at M¯ anoa of the Raman scattering cross-section of cyclohexane and Raman efficiency of a pressed pellet sample of gypsum powder. A pressed pellet of powdered gypsum sample was measured by SuperCam and by UH’s Raman spectrometer. In the cross-calibration study, UH’s remote Raman spectrometer, described previously (Gasda et al. 2015), was radiometrically calibrated by a Labsphere calibrated lamp in a manner analogous to that utilized to calibrate SuperCam (Sect. 7.1.1). The radiometric calibration was verified by measuring the differential Raman-scattering cross section of cyclohexane’s 801.3cm −1Raman mode. Measurements were made with both 1 cm and 1 mm path-length Starna cells, which agreed well. The measured differential Raman scattering cross-section was found to be 4.58 ×10−30 cm2sr−1molecule−1,whichwas within 1% of the published values (Trulson and Mathies 1986; Acosta-Maeda et al. 2017) for that differential Raman cross-section. A field-stop, placed at the sample location 6.11 meters from the primary mirror of UH’s Raman spectrometer, circumscribed the projection of the spectrometer’s entrance slit at the sample location so that the laser energy within the FOV of the spectrometer at the sample location could be measured directly. After verifying the accuracy of the IRF, the Raman efficiency of gypsum’s ν1emission at 1008 cm−1 was determined to be 2.04 ×10−8, which is in general agreement with published values of gypsum samples (Stopar et al. 2004,2005). The gypsum sample was observed by SuperCam as a function of distance using the EQM instrument. The Raman spectrum of gypsum was corrected from counts (DN) at the CCD to photons at the aperture by application of the IRF. Following the IRF correction, the spectra were corrected by geometrical factors to convert photons at the entrance aperture to Raman photons emitted at the source, assuming Lambertian scattering into πsr. The results of this experiment are presented in Fig. 49. Three regions of the plot are readily distinguishable. First, the laser photons per pulse hitting the target (and also the Raman photons produced at the target) within the FOV of the transmission spectrometer increases with distance as the FOV encloses a greater portion of the laser beam (Maurice et al. 2020). Second, the number of laser photons hitting the target within the FOV plateaus as the portion of the laser beam with the highest energy density apparently starts to fall outside of the spectrometer FOV. Third, the laser energy declines as this effect outweighs the expanded spectrometer FOV. The system is sensitive to very slight (i.e. <0.35 mrad) laser-beam misalignments resulting from telescope temperature variations (Maurice et al. 2020), causing the laser spot to deviate relative to the spectrometer FOV, so more or less light is captured by the spectrometer FOV. As the distance increases, the coupling efficiency of the laser spot to the FOV decreases, depending on the amount of mis-alignment. For highly amplified systems, as is the case with the transmission spectrometer when the intensifier gain is at a high setting for Raman spectroscopy, it is useful to estimate the number of photons being collected to understand the limitations of the technique. This number is not given by the IRF, which also includes the amplification of the photons between the intensifier cathode and the CCD. To determine shot noise and SNR, the number of photons counted at the cathode is the relevant parameter. To determine this from the number of photons in the FOV at the target (Fig. 49), the number of photons entering the instrument aperture was calculated from the solid angle subtended by the instrument, and the transmission from the aperture to the intensifier photocathode was estimated from the transmission of each optic. Depending on the wavelength bin of interest, one to six percent of light entering the telescope aperture reaches the photocathode, with near 6% efficiency at the wavenumber The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 73 of 87 4 Fig. 49 Number of photons generated at the target within the spectrometer field of view (FOV) as a function of distance, for the ν1mode of gypsum powder using the EQM SuperCam instrument. The result is based on data taken at these distances and knowledge of the spectrometer FOV convolved with an independent determination of the Raman efficiency of the sample. For a given sample, the number of Raman photons produced within the FOV depends on the alignment of the laser beam and the spectrometer FOV, which can degrade beyond a certain distance. The number of photons collected and detected by the instrument is a convolution of this curve with the 1/r2losses with distance, as presented in Table 12 Table 12 Propagation of Raman photons from source to detector for gypsum ν1emission Units are photons between the FWHM of the Raman peak per laser pulse. For the source emission, only photons produced within the spectrometer FOV are counted (as in Fig. 49), resulting in an increase between 1.8 and 3.8 m. Numbers are based on measurements made with the EQM. The FM will have different results based on alignment at the time of measurement Distance (m) At source At aperture At cathode 1.8 1.8e7 4820 310 2.3 4.0e7 4080 260 2.8 7.5e7 3480 220 3.3 1.0e8 2400 150 3.8 1.3e8 1710 110 4.3 1.3e8 1080 68 4.8 1.3e8 690 43 6.0 1.1e8 250 14 used in this experiment. Table 12 shows the number of photons within the spectrometer FOV at the source, the aperture, and the intensifier photocathode. A minimum of one to ten photons per channel is necessary to achieve detection. SuperCam’s detection limit is wavelength-dependent, requiring the collection of 50 to 150 photons at the aperture. Each peak is at least five channels wide, and Raman observations are expected to use at least fifty laser pulses, so statistics are built up. This calculation was performed with the EQM; the alignment and beam profile is believed to be better with the FM, although the angular size of the laser beam is the same. This exercise, which used a relatively strong sulfate scattering spectral feature, illustrates the challenges of observing Raman signals at long distance. 4Page 80 of 87 R.C. Wiens et al. SuperCam’s imager (RMI) has a small field of view compared to the other cameras on board, but its contribution is paramount to the success of the investigation. As requested by the SDT, it is essential to document the context (morphology, texture) of each chemistry and mineralogy observation, and to place them into the broader perspective given by the Navcam (Maki et al. 2020, this journal) and Mastcam (Bell et al. 2020, this journal) imagers. One aspect of SuperCam is largely yet to be realized within this Mars 2020 mission, and that is the synergy that comes from fusion of the results from the different techniques. This data fusion can be applied within the SuperCam investigations, or at the level of all techniques on the Mars 2020 rover (Farley et al. 2020). Studies to date include one early field study by a small team representing various instruments (Martin et al. 2020), an exercise on a “mystery rock” within the SuperCam team (Ollila et al. 2019), and several remote operations and science team training (ROASTT) events organized within the Mars 2020 project (Lawson et al., in preparation). All of these were much more limited in scope than is expected to be experienced with the Perseverance rover covering multiple outcrops over a large field area (Sun and Stack 2020). We expect the synergy experienced at both the human level, and with tools such as machine learning, will provide surprising discoveries that would not be possible without combining the results of the individual techniques. Acknowledgements Many people contributed to this project in addition to the co-authors, and we are most grateful for their support. This project was supported in the US by the NASA Mars Exploration Program, and in France by CNES, CNRS, and local universities. Support in Spain was provided by the Spanish Science Ministry. SuperCam benefitted from LANL laboratory-directed research and development funding which provided early prototypes of the new technologies incorporated in the SuperCam BU. J. Bell, A. Yingst, and K. Bennett are thanked for reviewing this manuscript; editorial support by K. Williford is also gratefully acknowledged. SDG. Authors’ contributions All authors contributed to either the proposal or the development and testing of the SuperCam instrument as described in this paper. Funding Was provided in the US by NASA’s Mars Exploration Program. Funding in France was provided by CNES and CNRS. Funding in Spain was provided by the Spanish Science Ministry. Some funding of data analyses at LANL was provided by Laboratory-Directed Research and Development funds. Data availability Data presented in the Results section of this paper are being made available to the Planetary Data System Geosciences Node under Mars 2020/SuperCam. Code availability Not applicable. Conflicts of interest/Competing interests The authors declare that there are no conflicts of interest or competing interests. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 81 of 87 4 Appendix of Abbreviations, Acronyms, and Short Definitions ADC Analog-to-digital converter AEGIS Autonomous exploration for gathering increased science: a software package to analyze Navcam images onboard and select SuperCam targets AFT Allowable flight temperature AOTF Acousto-optic tuning filter, performs wavelength selection for the IR spectrometer APG Annealed pyrolytic graphite Be Beryllium BU Body unit C&DH Control and data handling (electrical board) CCD Charge coupled device CDR Calibrated data record CMOS Complementary metal oxide sensor CNES Centre National d’Etudes Spatiales CTE Coefficient of thermal expansion CWL Continuous-wave laser DAC Digital-to-analog converter DC Direct current Decon Decontamination DN Digital number EDR Experimental data record EDU Engineering development unit EMI/EMC Electromagnetic interference and electromagnetic compatibility EQM Engineering qualification model FITS Flexible image transport system (file format) FPGA Field programmable gate array FOC Fiber optic cable FOV Field of view FM Flight model FWHM Full width at half maximum GOR Green, orange, and red spectral regions covered by the transmission spectrometer HGA High-gain antenna HSS High-speed serial HVPS High-voltage power supply Hz Hertz ICER Lossy progressive wavelet image compression IR Infrared IRAP Institut de Recherche en Astrophysique et Planetologie, the lead SuperCam institution in France IRS Infrared spectrometer J Joules JPL Jet Propulsion Laboratory kHz Kilohertz LANL Los Alamos National Laboratory LIBS Laser-induced breakdown spectroscopy Lpmm Lines per millimeter 4Page 82 of 87 R.C. Wiens et al. LVDS Low-voltage differential signal LVPS Low-voltage power supply Mbps Million bits per second MHz Megahertz Mrad Milliradians MRAM Magnetoresistive random-access memory MSL Mars Science Laboratory MU Mast unit mV Millivolt NA Numerical aperture Navcam Navigation camera for the Mars 2020 rover Nd:YAG Neodymium-doped yttrium-aluminum garnet P43 Phosphor screen used in the image intensifier QE Quantum efficiency PCB Printed circuit board PROM Programmable read-only memory RAMP Rover accessory mounting plate, the main mounting surface for instruments and electronics in the interior of the rover body RCE Rover compute element (computer) RMI Remote micro-imager RSM Remote-sensing mast RTG Radiothermal isotope generator RTV Room-temperature vulcanizing SAGE Venus Surface and Atmosphere Geochemical Explorer SCCT SuperCam calibration targets SDRAM Synchronous dynamic random-access memory SE Spectrometer electronics (board) SHERLOC Scanning Habitable Environments with Raman and Luminescence for Organics and Chemicals, a UV Raman spectrometer mounted on the Perseverance rover’s arm SOH State of health Sr Steradian STT System thermal test, rover-level SUROM Start-up read-only memory TEC Thermal-electric cooler Ti Titanium TRL Time-resolved luminescence TTL Transistor-transistor logic UART Universal asynchronous receiver-transmitter UV Ultraviolet spectral range (∼245–340 nm) covered by one of SuperCam’s two reflection spectrometers V Volts VIO Violet spectral range (∼385–465 nm) covered by one of SuperCam’s two spectrometers VIS Visible spectral range encompassing the violet, green, orange, and red spectral ranges, used for passive reflectance spectroscopy VISIR Visible and infrared passive reflectance spectroscopy W Watts XDR Extracted data record The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 83 of 87 4 References T.E. Acosta-Maeda, A.K. Misra, J.N. Porter, D.E. Bates, S.K. Sharma, Appl. Spectrosc. 71, 1025–1038 (2017). https://doi.org/10.1177/0003702816668531 R.E. Arvidson, P. Bellutta, F. Calef, A.A. Fraeman, J. Garvin, O. Gasnault, J. Grant, J.P. Grotzinger, V. Hamilton, M. Heverly, K.A. Iagnemma, J. Johnson, N. Lanza, S. Le Mouelic, N. Mangold, D. Ming, M. Mehta, R.V. Morris, H. Newsom, N. Renno, D. Rubin, J. Schieber, R. Sletten, A.R. Vasavada, J. Vizcaino, R.C. Wiens, J. Geophys. Res., Planets 119, 1322 (2014). https://doi.org/10.1002/2013JE004605 J. Barbarand, M. Pagel, Am. Mineral. 86, 473 (2001) J. Bell III, S.W. Squyres, K.E. Herkenhoff, J.N. Maki, H.M. Arneson, D. Brown, S.A. Collins, A. Dingizian, S.T. Elliott, E.C. Hagerott, A.G. Hayes, M.J. Johnson, J.R. Johnson, J. Joseph, K. Kinch, M.T. Lemmon, R.V. Morris, L. Scherr, M. Schwochert, M.K. Shepard, G.H. Smith, J.N. Sohl-Dickstein, R.J. Sullivan, W.T. Sullivan, M. Wadsworth, J. Geophys. Res. 108, 8063 (2003). https://doi.org/10.1029/ 2003JE002070 J. Bell et al. (2020, this journal). https://doi.org/10.1007/s11214-020-00755-x M.Y. Berezin, S. Achilefu, Chem. Rev. 110, 2641 (2010) O. Beyssac, Elements 16, 117 (2020) R. Bhartia et al. (2020, this journal). SPAC-D-20-00105R1 J.-P. Bibring, A. Soufflot, M. Berthé, Y. Langevin, B. Gondot, P. Drossart, M. Bouyé, M. Combes, P. Puget, A. Semery, G. Bellucci, V. Formisano, V. Moroz, V. Kottsov, G. Bonello, S. Erard, O. Forni, A. Gendrin, N. Nanaud, F. Poulet, G. Poulleau, T. Envrenaz, T. Fouchet, R. Melchiori, F. Altieri, P. Pinet, B. Schmitt, C. Sotin, E. Hauber, H. Hoffmann, R. Jaumann, U. Keller, R. Arvidson, J. Mustard, F. Forget, ESA SP 1240, 37 (2004) S.A. Brokus, D.K. Silletti, J. Mark Lunderberg, P.A. DeYoung, G.F. Peaslee, D.E. Carpenter, J. Buscaglia, Am. Mineral. 100, 451 (2015) E. Cantelar, G. Lifante, T. Calderon, R. Melendrez, A. Millan, M.A. Alvarez, M. Barboza-Flores, J. Alloys Compd. 323–324, 851 (2001) D.M. Chew, M.G. Babechuk, N. Cogné, C. Mark, G.J. O’Sullivan, I.A. Henrichs, D. Doepke, C.A. McKenna, Chem. Geol. 435, 35 (2016) B. Chide, S. Maurice, N. Murdoch, J. Lasue, B. Bousquet, X. Jacob, A. Cousin, O. Forni, O. Gasnault, P.-Y. Meslin, J.-F. Fronton, M. Bassas-Portus, A. Cadu, A. Sournac, D. Mimoun, R.C. Wiens, Spectrochim. Acta B 153, 50 (2019). https://doi.org/10.1016/j.sab.2019.01.008 B. Chide, S. Maurice, D. Mimoun, N. Murdoch, R.C. Wiens, Lunar Planet. Sci. LI, 1366 (2020) P.R. Christensen et al., J. Geophys. Res. 108(E12), 8064 (2003). https://doi.org/10.1029/2003JE002117 S.M. Clegg, R.C. Wiens, A.K. Misra, S.K. Sharma, J. Lambert, S. Bender, R. Newell, K. Nowak-Lovato, S. Smrekar, M.D. Dyar, S. Maurice, Spectrochim. Acta 68, 925 (2014) S.M. Clegg, R.C. Wiens, R.B. Anderson, O. Forni, J. Frydenvang, J. Lasue, A. Pilleri, V. Payre, T. Boucher, M.D. Dyar, S.M. McLennan, R.V. Morris, T.G. Graff, S.A. Mertzman, B.L. Ehlmann, S.C. Bender, R.L. Tokar, I. Belgacem, H. Newsom, R.E. McInroy, R. Martinez, P. Gasda, O. Gasnault, S. Maurice, Spectrochim. Acta B 129, 64 (2017) S.M. Clegg, R.B. Anderson, W. Rapin, B.L. Ehlmann, D.E. Anderson, N.H. Thomas, P. Gasda, J. Frydenvang, O. Forni, H. Newsom, D. Blaney, W. Goetz, S. Maurice, R.C. Wiens, 50th Lunar and Planetary Science Conference (2019), p. 2768 E.A. Cloutis, P.M. Asher, S.A. Mertzman, J. Geophys. Res. 107 (2002). https://doi.org/10.1029/2000JE1467 E.A. Cloutis, F.C. Hawthorne, S.A. Mertzman, K. Renn, M.A. Craig, D. Marcino, M. Methot, J. Strong, J.F. Mustard, D.L. Blaney, J.F. Bell III, F. Vilas, Icarus 184, 121 (2006) E.A. Cloutis, M.A. Craig, R.V. Kruzelecky, W.R. Jamroz, A. Scott, F.C. Hawthorne, S.A. Mertzman, Icarus 195, 140 (2007). https://doi.org/10.1016/j.icarus.2007.10.028 G.B. Courreges-Lacoste, B. Ahlers, F. Rull Perez, Spectrochim. Acta A 68, 1023 (2007) A. Cousin, O. Forni, P.-Y. Meslin, S. Clegg, A. Ollila, O. Gasnault, S. Maurice, R.C. Wiens (the ChemCam Team), 51st Lunar and Planetary Science Conference (2020), p. 2160 M.C. De Sanctis, E. Ammannito, H.Y. McSween, A. Raponi, S. Marchi, F. Capaccioni, M.T. Capria, F.G. Carrozzo, M. Ciarniello, S. Fonte, M. Formisano, A. Frigeri, M. Giardino, A. Longobardo, G. Magni, L.A. McFadden, E. Palomba, C.M. Pieters, F. Tosi, F. Zambon, C.A. Raymond, C.T. Russell, Science 355, 719–722 (2017) B.L. Ehlmann, J.F. Mustard, S.L. Murchie, F. Poulet, J.L. Bishop, A.J. Brown, W.M. Calvin, R.N. Clark, D.J. Des Marais, R.E. Milliken, L.H. Roach, T.L. Roush, G.A. Swayze, J.J. Wray, Science 322, 1828 (2008) A. El Ali, V. Barbin, G. Calas, B. Cervelle, K. Ramseyer, J. Bouroulec, Chem. Geol. 104, 189 (1993) C. Fabre, S. Maurice, A. Cousin, R.C. Wiens, O. Forni, V. Sautter, D. Guillaume, Spectrochim. Acta B 66, 280 (2011). https://doi.org/10.1016/j.sab.2011.03.012 K. Farley et al. (2020, this journal). SPAC-D-20-00057 4Page 84 of 87 R.C. Wiens et al. A. Fau, O. Beyssac, M. Gauthier, P.-Y. Meslin, A. Cousin, K. Benzerara, S. Bernard, J.C. Boulliard, O. Gasnault, O. Forni, R.C. Wiens, C. Drouet, M. Morand, P. Rosier, Y. Garino, S. Pont, S. Maurice, Spectrochim. Acta 160, 105687 (2019) O. Forni, M. Gaft, M. Toplis, S.M. Clegg, V. Sautter, S. Maurice, R.C. Wiens, N. Mangold, O. Gasnault, V. Sautter, S. Le Mouelic, P.-Y. Meslin, M. Nachon, R.E. McInroy, A.M. Ollila, A. Cousin, J.C. Bridges, N.L. Lanza, M.D. Dyar, Geophys. Res. Lett. 42, 1020 (2015). https://doi.org/10.1002/2014GL062742 R. Francis, T. Estlin, G. Doran, S. Johnstone, D. Gaines, V. Verma, M. Burl, J. Frydenvang, S. Montano, R.C. Wiens, S. Schaffer, O. Gasnault, L. Deflores, D. Blaney, B. Bornstein, Sci. Robot. 2, eaan4582 (2017) N. Fray, A. Bardyn, H. Cottin, K. Altwegg, D. Baklouti, C. Briois, L. Colangeli, C. Engrand, H. Fischer, A. Glasmachers, E. Grün, G. Haerendel, H. Henkel, H. Höfner, K. Hornung, E.K. Jessberger, A. Koch, H. Krüger, Y. Langevin, H. Lehto, K. Lehto, L. Le Roy, S. Merouane, P. Modica, F.-R. Orthous-Daunay, J. Paquette, F. Raulin, J. Rynö, R. Schulz, J. Silén, S. Siljeström, W. Steiger, O. Stenzel, T. Stephan, L. Thirkell, R. Thomas, K. Torkar, K. Varmuza, K.-P. Wanczek, B. Zaprudin, J. Kissel, M. Hilchenbach, Nature 538, 72–74 (2016). https://doi.org/10.1038/nature19320 M. Gaft, R. Reisfeld, G. Panczer, P. Blank, G. Boulon, Spectrochim. Acta A 54, 2163 (1998) M. Gaft, G. Panczer, R. Reisfeld, I. Shinno, J. Alloys Compd. 300–301, 267 (2000) M. Gaft, R. Reisfeld, G. Panczer, Modern Luminescence Spectroscopy of Minerals and Materials (Springer, Berlin, 2005), p. 356 M. Gaft, R. Reisfeld, G. Panczer, Modern Luminescence Spectroscopy of Minerals and Materials, 2nd edn. (Springer, Berlin, 2015), ISBN-13: 978-3319247632, 606 pp. P.J. Gasda, T.E. Acosta-Mayeda, P.G. Lucey, A.K. Misra, S.K. Sharma, G.J. Taylor, Appl. Spectrosc. 69, 173 (2015). https://doi.org/10.1366/14-07483 M.P. Golombek, R.C. Anderson, J.R. Barnes, J.F. Bell III, N.T. Bridges, D.T. Britt et al., J. Geophys. Res. 104, 8523 (1999) T.A. Goudge, J.F. Mustard, J.W. Head, C.I. Fassett, S.M. Wiseman, J. Geophys. Res. 120, 775 (2015). https://doi.org/10.1002/2014JE004782 T.A. Goudge, R.A. Milliken, J.W. Head, J.F. Mustard, C.I. Fassett, Earth Planet. Sci. Lett. 458, 357 (2017) K.P. Hand, A.E. Murray, J.B. Garvin, W.B. Brinkerhoff, B.C. Christner, K.S. Edgett, B.L. Ehlmann, C.R. German, A.G. Hayes, T.M. Hoehler, S.M. Horst, J.I. Lunine, K.H. Nealson, C. Paranicas, B.E. Schmidt, D.E. Smith, A.R. Rhoden, M.J. Russell, A.S. Templeton, P.A. Willis, R.A. Yingst, C.B. Phillips, M.L. Cable, K.L. Craft, A.E. Hoffman, T.A. Nordheim, R.P. Pappalardo (the Project Engineering Team). Report of the Europa Lander Definition Team, posted 2017 (2017) P. Hoess, K. Fleder, Proc. SPIE 4128, Image Intensifiers and Applications II (2000). https://doi.org/10.1117/ 12.405878 B.H.N. Horgan, R.B. Anderson, G. Dromart, E.S. Amador, M.S. Rice, Icarus 339, 113526 (2020). https://doi.org/10.1016/j.icarus.2019.113526 J.R. Johnson, J.F. Bell III, S. Bender, D. Blaney, E. Cloutis, L. DeFlores, B. Helmann, O. Gasnault, B. Gondet, K. Kinch, M. Lemmon, S. Le Mouelic, S. Maurice, M. Rice, R. Wiens (the MSL Science Team), Icarus 249, 74 (2015). https://doi.org/10.1016/j.icarus.2014.02.028 J.R. Johnson, C. Achilles, J.F. Bell III, S. Bender, J. Berger, E. Cloutis, B. Ehlmann, A. Fraeman, O. Gasnault, V.E. Hamilton, S. Le Mouelic, S. Maurice, P. Pinet, L. Thompson, D. Wellington, R.C. Wiens, J. Geophys. Res., Planets 122, 2655 (2017). https://doi.org/10.1002/2016JE005187 J.R. Johnson, P.-Y. Meslin, J.F. Bell III, R.C. Wiens, S. Maurice, O. Gasnault, Lunar Planet. Sci. LI, 1136 (2020) M. Kayama, H. Nishido, T. Sekine, T. Nakazato, A. Gucsik, K. Ninagawa, J. Geophys. Res. 117, E09004 (2012) M. Kayama, H. Nishido, S. Toyoda, K. Komuro, A.A. Finch, M.R. Lee, K. Ninagawa, Am. Mineral. 99,65 (2014) A. Kiely, M. Klimesh IPN Progress Report 42-155 (2003). https://ipnpr.jpl.nasa.gov/progress_report/42-155/ 155J.pdf J.R. Lakowicz, Principles of Fluorescence Spectroscopy, 3rd edn. (Springer, Berlin, 2006), p. 954 N.L. Lanza, S.M. Clegg, R.C. Wiens, R.E. McIlroy, H.E. Newsom, M.D. Deans, Appl. Opt. 51, B74 (2012) N.L. Lanza, A. Ollila, A. Cousin, R.C. Wiens, S. Clegg, N. Mangold, N. Bridges, D. Cooper, M. Schmidt, J. Berger, R. Arvidson, N. Melikechi, H. Newsom, R. Tokar, C. Hardgrove, A. Mezzacappa, R. Jackson, B. Clark, O. Forni, S. Maurice, M. Nachon, R.B. Anderson, J. Blank, M. Deans, D. Delapp, R. Leveillé, R. McInroy, R. Martinez, P.-Y. Meslin (the MSL team), Icarus 249, 43 (2015) N.L. Lanza, R.C. Wiens, R.E. Arvidson, B.C. Clark, W.W. Fischer, R. Gellert, J.P. Grotzinger, J.A. Hurowitz, S.M. McLennan, R.V. Morris, M.S. Rice, J.F. Bell III, J.A. Berger, D.L. Blaney, N.T. Bridges, F. Calef III, J.L. Campbell, S.M. Clegg, A. Cousin, K.S. Edgett, C. Fabre, M.R. Fisk, O. Forni, J. Frydenvang, K.R. Hardy, C. Hardgrove, J.R. Johnson, J. Lasue, S. Le Mouelic, M.C. Malin, N. Mangold, J. Martin-Torres, S. Maurice, M.J. McBride, D.W. Ming, H.E. Newsom, S. Schröder, L.M. Thompson, The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 85 of 87 4 A.H. Treiman, S. VanBommel, D.T. Vaniman, M.-P. Zorzano, Geophys. Res. Lett. 43, 7398 (2016). https://doi.org/10.1002/2016GL069109 S. Le Mouelic, O. Gasnault, K.E. Herkenhoff, N.T. Bridges, Y. Langevin, N. Mangold, S. Maurice, R.C. Wiens, P. Pinet, H.E. Newsom, R.G. Deen, J.F. Bell III, J.R. Johnson, W. Rapin, B.L. Barraclough, D.L. Blaney, L. Deflores, J. Maki, M.C. Malin, R. Perez, M. Saccoccio (the MSL Science Team), Icarus 249, 93 (2015). https://doi.org/10.1016/j.icarus.2014.05.030 C. Lenz, D. Talla, K. Ruschel, R. Skoda, J. Götze, L. Nasdala, Mineral. Petrol. 107, 415 (2013) C. Lenz, L. Nasdala, D. Talla, C. Hauzenberger, R. Seitz, U. Kolitsch, Chem. Geol. 415, 1 (2015) J.N. Maki, D. Gruel, C. McKinney, M.A. Ravine, M. Morales, D. Lee, R. Willson, D. Copley-Woods, M. Valvo, T. Goodsall, J. McGuire, R.G. Sellar, J.A. Schaffner, M.A. Caplinger, J.M. Shamah, A.E. Johnson, H. Ansari, K. Singh, T. Litwin, R. Deen, A. Culver, N. Ruoff, D. Petrizzo, D. Kessler, C. Basset, T. Estlin, F. Alibay, A. Nelessen, S. Algermissen, Space Sci. Rev. 216, 137 (2020). https://doi.org/10.1007/s11214-020-00765-9 J. Manfredi et al. (2020, this journal). SPAC-D-20-00067R1 N. Mangold, G. Dromart, V. Ansan, F. Salese, M.G. Kleinhans, M. Massé, C. Quantin, K.M. Stack, Astrobiology 20, 994–1013 (2020). https://doi.org/10.1089/ast.2019.2132 J.A. Manrique, G. Lopez-Reyes, A. Cousin, F. Rull, S. Maurice, R.C. Wiens, M.B. Madsen, J.M. Madariaga, O. Gasnault, J. Aramendia, G. Arana, P. Beck, S. Bernard, P. Bernardi, M.H. Bernt, A. Berrocal, O. Beyssac, P. Caïs, C. Castro, K. Castro, S.M. Clegg, E. Cloutis, G. Dromart, C. Drouet, B. Dubois, D. Escribano, C. Fabre, A. Fernandez, O. Forni, V. Garcia-Baonza, I. Gontijo, J. Johnson, J. Laserna, J. Lasue, S. Madsen, E. Mateo-Marti, J. Medina, P.-Y. Meslin, G. Montagnac, A. Moral, J. Moros, A.M. Ollila, C. Ortega, O. Prieto-Ballesteros, J.M. Reess, S. Robinson, J. Rodriguez, J. Saiz, J.A. Sanz-Arranz, I. Sard, V. Sautter, P. Sobron, M. Toplis, M. Veneranda, Space Sci. Rev. 216, 138 (2020). https://doi.org/10.1007/s11214-020-00764-w P. Martin, B.L. Ehlmann, N.H. Thomas, R.C. Wiens, J.J.R. Hollis, L.W. Beegle, R. Bhartia, S.M. Clegg, D.L. Blaney, Earth Space Sci. 7, e2019EA000720 (2020). https://doi.org/10.1029/2019EA000720 S. Maurice et al. (2020, this journal). SPAC-D-20-00069R1 S. Maurice, R.C. Wiens, M. Saccoccio, B. Barraclough, O. Gasnault, O. Forni, N. Mangold, D. Baratoux, S. Bender, G. Berger, J. Bernardin, M. Berthé, N. Bridges, D. Blaney, M. Bouyé, P. Cais, B. Clark, S. Clegg, A. Cousin, D. Cremers, A. Cros, L. DeFlores, C. Derycke, B. Dingler, G. Dromart, B. Dubois, M. Dupieux, E. Durand, L. d’Uston, C. Fabre, B. Faure, A. Gaboriaud, T. Gharsa, K. Herkenhoff, E. Kan, L. Kirkland, D. Kouach, J.-L. Lacour, Y. Langevin, J. Lasue, S. Le Mouélic, M. Lescure, E. Lewin, D. Limonadi, G. Manhes, P. Mauchien, C. McKay, P.-Y. Meslin, Y. Michel, E. Miller, H.E. Newsom, G. Orttner, A. Paillet, L. Pares, Y. Parot, R. Perez, P. Pinet, F. Poitrasson, B. Quertier, B. Sallé, C. Sotin, V. Sautter, H. Seran, J.J. Simmonds, J.-B. Sirven, R. Stiglich, N. Streibig, J.-J. Thocaven, M. Toplis, D. Vaniman, Space Sci. Rev. 170, 95 (2012). https://doi.org/10.1007/s11214-012-9912-2 S. Maurice, S. Clegg, R.C. Wiens, O. Gasnault, W. Rapin, O. Forni, A. Cousin, V. Sautter, N. Mangold, L. Le Deit, M. Nachon, R.B. Anderson, N. Lanza, C. Fabre, V. Payré, J. Lasue, P.-Y. Meslin, J.-B. Sirven, N. Melikechi, S. Le Mouelic, J. Frydenvang, A. Vasavada, J.C. Bridges, S.C. Bender, S. Schröder, R. Francis, P. Pinet, H. Newsom, A. Ollila, K. Herkenhoff, M.B. Madsen, G. Dromart, P. Beck, E. Lewin, J.-L. Lacour, Y. Langevin, B. Gondet, L. d’Uston, G. Berger, M. Toplis, J.R. Johnson, M.D. Dyar, N. Bridges, D. Vaniman, B. Barraclough, J. Anal. At. Spectrom. (2016). https://doi.org/10.1039/c5ja00417a N. Melikechi, A. Mezzacappa, A. Cousin, N.L. Lanza, S.M. Clegg, R.C. Wiens, G. Berger, S. Maurice, S. Bender, O. Forni, D. Delapp, J. Lasue, O. Gasnault, H. Newsom, A.M. Ollila, E. Lewin, E.A. Breves, M.D. Dyar, J. Frydenvang, D. Blaney, B.C. Clark (the MSL Science Team), Spectrochim. Acta B 96, 51 (2014). https://doi.org/10.1016/j.sab.2014.04.004 A. Mezzacappa, N. Melikechi, A. Cousin, R.C. Wiens, J. Lasue, S.M. Clegg, R. Tokar, S. Bender, N.L. Lanza, S. Maurice, G. Berger, O. Forni, O. Gasnault, M.D. Dyar, T. Boucher, E. Lewin, C. Fabre (the MSL Science Team), Spectrochim. Acta B 120, 19 (2016). https://doi.org/10.1016/j.sab.2016.03.009 G. Montagnac, G. Dromart, P. Beck, F. Mercier, B. Reynard, A. Cousin, S. Maurice, R.C. Wiens, J. Raman Spectrosc. (2018). https://doi.org/10.1002/jrs.5406 S. Murchie, R. Arvidson, P. Bedini, K. Beisser, J.-P. Bibring, J. Bishop, J. Boldt, P. Cavender, T. Choo, R.T. Clancy, E.H. Darlington, D. Des Marais, R. Espiritu, D. Fort, R. Green, E. Guinness, J. Hayes, C. Hash, K. Heffernan, J. Hemmler, G. Heyler, D. Humm, J. Hutcheson, N. Izenberg, R. Lee, J. Lees, D. Lohr, E. Malaret, T. Martin, J.A. McGovern, P. McGuire, R. Morris, J. Mustard, S. Pelkey, E. Rhodes, M. Robinson, T. Roush, E. Schaefer, G. Seagrave, F. Seelog, P. Silverglate, S. Slavney, M. Smith, W.-J. Shyong, K. Strohbehn, H. Taylor, P. Thompson, B. Tossman, M. Wirzburger, M. Wolff, J. Geophys. Res. 112, E05S03 (2007). https://doi.org/10.1029/2006JE002682 N. Murdoch, B. Chide, J. Lasue, A. Cadu, A. Sournac, M. Bassas-Ports, J. Merrison, J.J. Iversen, C. Moretto, C. Valsco, L. Pares, A. Hynes, V. Godiver, R.D. Lorenz, P. Cais, P. Bernardi, S. Maurice, R.C. Wiens, D. Mimoun, Planet. Space Sci. 165, 260 (2019). https://doi.org/10.1016/j.pss.2018.09.009 4Page 86 of 87 R.C. Wiens et al. J.F. Mustard, M. Adler, A. Allwood, D.S. Bass, D.W. Beaty, J.F. Bell III, W.B. Brinckerhoff, M. Carr, D.J. Des Marais, B. Drake, K.S. Edgett, J. Eigenbrode, L.T. Elkins-Tanton, J.A. Grant, S.M. Milkovich, D. Ming, C. Moore, S. Murchie, T.C. Onstott, S.W. Ruff, M.A. Sephton, A. Steele, A. Treiman, Report of the Mars 2020 Science Definition Team, 154 pp., posted July, 2013, by the Mars Exploration Program Analysis Group (MEPAG) at http://mepag.jpl.nasa.gov/reports/MEP/Mars_2020_SDT_Report_Final.pdf (2013) A. Ollila, R.C. Wiens, R. Perez, A. Nelson, M. Bodine, S. Maurice, S.K. Sharma, P.J. Gasda, J.-M. Reess, T. Fouchet, E. Cloutis, J. Johnson, S. Bender, F. Montmessin, P. Bernardi, J. Frydenvang, K. McCabe, P. Cais, Lunar Planet. Sci. 48, 2339 (2017) A.M. Ollila, N.L. Lanza, O. Beyssac, M. Gauthier, S. Clegg, A. Misra, S. Sharma, R.C. Wiens, S. Maurice, O. Gasnault, Lunar Planet. Sci. 49, 2786 (2018) A.M. Ollila, R.C. Wiens, S. Maurice, A. Cousin, R. Anderson, O. Beyssac, L. Bonal, P. Beck, S. Clegg, B. Chide, L. Deflores, G. Dromart, W. Fischer, O. Forni, T. Fouchet, O. Gasnault, J. Grotzinger, J. Johnson, J. Lasue, J. Laserna, J.M. Madariaga, M. Madsen, N. Mangold, T. Nelson, R. Newell, J. Martinez-Frias, S. McLennan, A. Misra, F. Montmessin, S. Robinson, S. Sharma, F. Rull, D. Venhaus, P. Bernardi, J.-M. Reese, A. Reyes-Newell, F. Poulet, N. Lanza, I. Torre-Fdez, J. Aramendia, R. Perez, in Ninth International Mars Conference #6352, Pasadena, California (2019) J.D. Pasteris, O. Beyssac, Elements 16, 87 (2020) V. Payré, C. Fabre, A. Cousin, V. Sautter, R.C. Wiens, O. Forni, O. Gasnault, J. Lasue, A. Ollila, W. Rapin, P.-Y. Meslin, S. Maurice, M. Nachon, L. Le Deit, N. Lanza, S.M. Clegg, J. Geophys. Res. 122, 650 (2017). https://doi.org/10.1002/2016JE005201 F. Postberg, N. Khawaja, B. Abel, G. Choblet, C.R. Glein, M.S. Gudipati, B.L. Henderson, H.-W. Hsu, S. Kempf, F. Klenner, G. Moragas-Klostermeyer, B. Magee, L. Nölle, M. Perry, R. Reviol, J. Schmidt, R. Srama, F. Stolz, G. Tobie, M. Trieloff, J.H. Waite, Nature 558, 564 (2018). https://doi.org/ 10.1038/s41586-018-0246-4 S. Potin, O. Brissaud, P. Beck, B. Schmitt, Y. Magnard, J.-J. Correia, P. Rabou, L. Jocou, Appl. Opt. 57, 8279 (2018). https://doi.org/10.1364/AO.57.008279 W. Rapin, P.-Y. Meslin, S. Maurice, D. Vaniman, M. Nachon, N. Mangold, S. Schröder, O. Gasnault, O. Forni, R.C. Wiens, G.M. Martinez, A. Cousin, V. Sautter, J. Lasue, D. Archer (the MSL science team), Earth Planet. Sci. Lett. 452, 197 (2016). https://doi.org/10.1016/j.epsl.2016.07.045 W. Rapin, B. Chauviré, T. Gabriel, A. McAdams, P.-Y. Meslin, B.L. Ehlmann, B. Rondeau, E. Dehouck, N. Mangold, C. Hardgrove, S. Maurice, R.C. Wiens, O. Gasnault, A. Cousin, O. Forni, J. Frydenvang, S. Schröder, J. Geophys. Res., Planets (2018). https://doi.org/10.1029/2017JE005483 W. Rapin, B.L. Ehlmann, G. Dromart, J. Schieber, N.H. Thomas, W.W. Fischer, V.K. Fox, N.T. Stein, M. Nachon, B.C. Clark, L.C. Kah, L. Thompson, H.A. Meyer, T.S.J. Gabriel, C. Hardgrove, N. Mangold, F. Rivera-Hernandez, R.C. Wiens, A.R. Vasavada, Nat. Geosci. (2019). https://doi.org/10.1038/ s41561-019-0458-8 C. Royer, F. Poulet, J.-M. Reess, C. Pilorget, V. Hamm, T. Fouchet, S. Maurice, O. Forni, R.C. Wiens, Rev. Sci. Instrum. 91, 063105 (2020). https://doi.org/10.1063/1.5145390 F. Rull, S. Maurice, I. Hutchinson, A. Moral, C. Perez, C. Diaz, M. Colombo, T. Belenguer, G. Lopez-Reyes, A. Sansano, O. Forni, Y. Parot, N. Streibig, S. Woodward, C. Howe, N. Tarcea, P. Rodriguez, L. Seoane, A. Santiago, J.A. Rodriguez-Prieto, J. Medina, P. Gallego, R. Canchal, P. Santamaria, G. Ramos, J.L. Vago, Astrobiology 17, 627 (2017). https://doi.org/10.1089/ast.2016.1567 M.R. Salvatore, T.A. Goudge, M.S. Bramble, C.S. Edwards, J.L. Bandfield, E.S. Amador, J.F. Mustard, P.R. Christensen, Icarus 301, 76 (2018) V. Sautter, M. Toplis, A. Cousin, O. Forni, R.C. Wiens, C. Fabre, O. Gasnault, M. Fisk, S. Maurice, A. Ollila, N. Mangold, L. Le Deit, P. Beck, W. Rapin, P. Pinet, J. Blank, S. Clegg, P.-Y. Meslin, M.D. Dyar, H. Newsom, N. Bridges, N. Lanza, S. Le Mouelic, D. Vaniman, Nat. Geosci. 8, 605 (2015). https://doi.org/ 10.1038/NGEO2474 V. Sautter, M.J. Toplis, P. Beck, N. Mangold, R.C. Wiens, P. Pinet, A. Cousin, S. Maurice, L. LeDeit, R. Hewins, O. Gasnault, C. Quantin, O. Forni, H. Newsom, P.-Y. Meslin, J. Wray, N. Bridges, V. Payre, Lithos 254–255, 36 (2016) B. Schrader, A. Hoffmann, S. Keller, Spectrochim. Acta 47A, 1135 (1991) S. Schröder, P.-Y. Meslin, O. Gasnault, S. Maurice, A. Cousin, O. Forni, N. Mangold, W. Rapin, S. Le Mouelic, A. Ollila, M. Nachon, J. Lasue, M.D. Dyar, S. Clegg, R. Jackson, R.C. Wiens, Icarus 249,43 (2015). https://doi.org/10.1016/j.icarus.2014.08.029 S. Schröder, K. Rammelkamp, D.S. Vogt, S. Frohmann, A. Cousin, O. Gasnault, H.-W. Huebers, Lunar Planet. Sci. Conf. 49, 1962 (2018) S.K. Sharma, Spectrochim. Acta A 68, 1008 (2007) J.-L. Starck, F. Murtagh, Astron. Astrophys. 288, 343 (1994) J.-L. Starck, J. Fadili, F. Murtagh, IEEE Trans. Image Process. 16(2), 297 (2007) The SuperCam Instrument Suite on the NASA Mars 2020 Rover... Page 87 of 87 4 J.H. Stewart, F.G. Poole, R.F. Wilson, R.A. Cadigan, U. S. Geol. Surv. Prof. Pap. 691 (1972). https://doi.org/ 10.3133/pp691 J.D. Stopar, P.G. Lucey, S.K. Sharma, A.K. Misra, H.W. Hubble, Proc. SPIE 5163, 99 (2004). https://doi.org/ 10.1117/12.505090 J.D. Stopar, P.G. Lucey, S.K. Sharma, A.K. Misra, G.J. Taylor, H.W. Hubble, Spectrochim. Acta A 61, 2315 (2005). https://doi.org/10.1016/j.saa.2005.02.030 V.Z. Sun, K.M. Stack, Geologic Map of Jezero Crater and the Nili Planum Region, Mars. U. S. Geological Survey Scientific Investigations Map 3464 (2020). https://doi.org/10.3133/sim3464 N.H. Thomas, B.L. Ehlmann, W. Rapin, F. Rivera-Hernandez, N.T. Stein, J. Frydenvang, T. Gabriel, P.-Y. Meslin, S. Maurice, R.C. Wiens, J. Geophys. Res., Planets 125, e2019JE006289 (2020). https://doi.org/ 10.1029/2019JE006289 I. Torre-Fdez, K. Castro, O. Beyssac, A.M. Ollila, G. Arana, S. Bernard, P. Bernardi, P. Cais, S. Clegg, M. Egan, O. Forni, O. Gasnault, I. Gontijo, J.M. Madariaga, J.A. Manrique, S. Maurice, A. Misra, G. Montagnac, T. Nelson, R. Newell, P. Pilleri, S. Robinson, F. Rull, S.K. Sharma, R.C. Wiens, P. Willis (the SuperCam Science Team), Lunar Planet. Sci. 51, 1868 (2020) M.O. Trulson, R.A. Mathies, J. Chem. Phys. 84, 2068 (1986) G.E. Ulrich, N.G. Bailey, Geologic map of the SP Mountain part of the San Francisco volcanic field, northcentral Arizona. U. S. Geological Survey, Miscellaneous Geologic Investigations Map MF-1956, scale 1:50000 (1987) D. Vaniman, M.D. Dyar, R.C. Wiens, A. Ollila, N. Lanza, J. Lasue, M. Rhodes, S.M. Clegg, Space Sci. Rev. 170, 229 (2012). https://doi.org/10.1007/s11214-012-9886-0 D.S. Vogt, K. Rammelkamp, S. Schröder, W.H. Hübers, Icarus 302, 470 (2018) G.A. Waychunas, Rev. Mineral. Geochem. 78, 175 (2014). https://doi.org/10.2138/rmg.2014.78.5 D.C. Wells, E.W. Greisen, R.H. Harten, Astron. Astrophys. Suppl. 44, 363 (1981) R.C. Wiens, S. Maurice (the ChemCam and MSL Science Teams), Elements 11, 33 (2015) R.C. Wiens, S.K. Sharma, J. Thompson, A. Misra, P.G. Lucey, Spectrochim. Acta A 61, 2324 (2005) R.C. Wiens, S. Maurice, B. Barraclough, M. Saccoccio, W.C. Barkley, J.F. Bell III, S. Bender, J. Bernardin, D. Blaney, J. Blank, M. Bouye, N. Bridges, P. Cais, R.C. Clanton, B. Clark, S. Clegg, A. Cousin, D. Cremers, A. Cros, L. DeFlores, D. Delapp, R. Dingler, C. D’Uston, M.D. Dyar, T. Elliott, D. Enemark, C. Fabre, M. Flores, O. Forni, O. Gasnault, T. Hale, C. Hays, K. Herkenhoff, E. Kan, L. Kirkland, D. Kouach, D. Landis, Y. Langevin, N. Lanza, F. LaRocca, J. Lasue, J. Latino, D. Limonadi, C. Lindensmith, C. Little, N. Mangold, G. Manhes, P. Mauchien, C. McKay, E. Miller, J. Mooney, R.V. Morris, L. Morrison, T. Nelson, H. Newsom, A. Ollila, M. Ott, L. Pares, R. Perez, F. Poitrasson, C. Provost, J.W. Reiter, T. Roberts, F. Romero, V. Sautter, S. Salazar, J.J. Simmonds, R. Stiglich, S. Storms, N. Streibig, J.-J. Thocaven, T. Trujillo, M. Ulibarri, D. Vaniman, N. Warner, R. Waterbury, R. Whitaker, J. Witt, B. Wong-Swanson, Space Sci. Rev. 170, 167 (2012). https://doi.org/10.1007/S11214-012-9902-4 R.C. Wiens, S. Maurice, J. Lasue, O. Forni, R.B. Anderson, S. Clegg, S. Bender, B.L. Barraclough, L. Deflores, D. Blaney, R. Perez, N. Lanza, A. Ollila, A. Cousin, O. Gasnault, D. Vaniman, M.D. Dyar, C. Fabre, V. Sautter, D. Delapp, H. Newsom, N. Melikechi (the ChemCam team), Spectrochim. Acta B 82, 1 (2013). https://doi.org/10.1016/j.sab.2013.02.003 R.C. Wiens, P.J. Gasda, A.K. Misra, T.E. Acosta-Maeda, S.K. Sharma, H. Quinn, K. Ganguly, R. Newell, S. Clegg, B. Sandoval, L. Ott, S. Maurice, C. Virmontois, Lunar Planet. Sci. 51, 1780 (2020)