MIRS: an imaging spectrometer for the MMX mission

The MMX infrared spectrometer (MIRS) is an imaging spectrometer onboard MMX JAXA mission. MMX (Martian Moon eXploration) is scheduled to be launched in 2024 with sample return to Earth in 2029. MIRS is built at LESIA-Paris Observatory in collaboration with four other French laboratories, collaboration and financial support of CNES and close collaboration with JAXA and MELCO. The instrument is designed to fully accomplish MMX’s scientific and measurement objectives. MIRS will remotely provide near-infrared spectral maps of Phobos and Deimos containing compositional diagnostic spectral features that will be used to analyze the surface composition and to support the sampling site selection. MIRS will also study Mars atmosphere, in particular spatial and temporal changes such as clouds, dust and water vapor.


Introduction
MIRS (MMX InfraRed Spectrometer) is an imaging spectrometer working from 0.9 to 3.6 µm with spectral resolution better than 20 nm. MIRS will work during the 3 years of the MMX stay in circum-Martian space (Kuramoto et al. 2021). The imaging spectrometer will provide global spectral characterization of Phobos and Deimos to map the surface composition, to support the Phobos sampling site selection and give detailed surface composition of the two sampling sites. MIRS will also study Mars atmosphere with particular attention to local time variation on the spatial distribution of water vapor, dust, clouds, and other atmospheric species, which could not be traced by previous satellites on the sun-synchronous polar orbit.
MIRS will contribute to the two major goals (Kuramoto et al. 2021) of the MMX mission: 1. Clarify the origins of the Martian moons and constrain processes for planetary formation and material transport in the region connecting the inner and outer solar system 2. Clarify the driving mechanism of the transition of the Mars-moons system and add new knowledge to the evolution history of Mars.

Controversy on the origin of the Martian moons
The origin of Phobos and Deimos is a debated question and attacking this problem represents the major MMX mission goal. Remote sensing observations by Viking 1 and 2, Mariner 9 and Phobos-2 revealed that the heavily cratered surface, the mass and the low density of Phobos and Deimos shared similarities with asteroids.
In addition, as detailed in the following sections, their spectra are very different from those of the martian surface, being red sloped (i.e., reflectance increases with wavelength), without strong absorption features and resembling the spectra of primitive C-and D-type asteroids.
Following the aforesaid observations, it was suggested that Phobos and Deimos could be captured asteroids. Asteroidal capture by a planet is possible if a dissipative process is at play to extract orbital energy from the passing body. This requires either that the object is braked in a circumplanetary atmosphere (Sasaki 1990), or in 3-body encounters where a binary asteroid is split by tides, and one body remains trapped in planetocentric orbit, and the others escape the planet (Hansen et al. 2018), or if tidal dissipation is intense enough (Kaula 1964). However, none of these processes can easily explain the close-to-zero inclination of Phobos and Deimos (with respect to Mars' equatorial plane). Indeed objects coming from the Asteroid belt may encounter Mars with high inclinations (Mars obliquity varies between 10° and 40° (Laskar and Robutel 1993;Touma and Wisdom 1993). Captured objects may remain on high eccentricity and high inclinations orbits with respect to the planet. Whereas tides may damp orbital inclination and eccentricity of a captured Phobos in less than 4.6 Gyrs (Burns 1992), owing to its larger size and closer distance to Mars, they are much less efficient for the case of Deimos that orbits beyond 6 Martian radii (Szeto 1983;Burns 1992).
An alternative explanation for the formation of a Martian disk was fostered by Craddock (2011) who suggested that an impact onto Mars may have formed a circumplanetary disk, and that Phobos and Deimos accreted locally from disk material. Reaccumulation of disk material into rubble piles naturally leads to the formation of a collection of a dozen of moons (Rosenblatt and Charnoz 2012; Rosenblatt et al. 2016;Canup and Salmon 2018) with masses between those of Deimos and up to 1000 times Phobos. Most moons would be tidally driven inward over time and impacts on the planet. In this view (Rosenblatt et al. 2016), Phobos would be the last surviving moon of this old population, whereas Deimos would be the only moon that accreted beyond Mars' synchronous orbit, and have experienced only little dynamical evolution.
Recently, Hesselbrock and Minton (2017) suggested that Phobos suffers multiple cycle of tidal destruction and re-accretion close to Mars' Roche limit (about 3 Mars radii), implying potentially that the today Phobos may have accreted only recently (< 100 Myrs).
In the giant impact hypothesis, Phobos and Deimos would be made of a mix (50-50%) of Martian mantle and impactor material (the origin of which is unconstrained). The temperatures that were reached during the impact are just high enough to melt silicate minerals, in the range of 2000 K only (Hyodo et al. 2017). However, this would be enough to devolatilize material and the resulting composition could resemble devolatilized Martian mantle with maybe substantial signs of oxidation if the impactor was water rich Pignatale et al. 2018).
Regardless of the origin of Martian moons, exogenic materials from the natural meteoroid impact and Martian materials ejected by small impacts on Mars would be mixed in the regolith of Phobos and Deimos (Ramsley and Head 2013;Hyodo et al. 2019), whereas, these late accreted masses are of the order of ~ 1000-10,000 ppm fraction in their regolith (Hyodo et al. 2019) and thus the change in the bulk surface composition would be negligible. If Phobos is formed by a giant impact, Phobos should contain high-temperature phase materials representing a mixture of Martian silicates (from crust and mantle).
MIRS will be able to characterize the global composition of Phobos and Deimos surface material (and subsurface through the mapping of crater ejectas) and will be able to support the selection of the two Phobos sampling sites. In combination with the sample return from Phobos, MIRS data will help deciphering whether Phobos composition is closer to that of C-D type asteroids (with similarity to carbonaceous chondrites and presence of organics or ices), which would imply a capture origin, or is more similar to a devolatilized and/or hydrated Martian mantle. This latter case would indicate, at the same time, a giant impact origin for Phobos and Deimos and would also give clues to the characteristics of the putative impactor onto Mars (see paper by Nakamura et al. 2021).

MIRS scientific objectives
MIRS is expected to characterize Phobos and Deimos surfaces and Mars atmospheric composition by remotely identifying diagnostic features in the near-infrared spectral range.
MIRS is used to achieve several of the mission requirements (MR), in particular:

Phobos
Due to the proximity of Phobos and Deimos to Mars, accurate analyses were possible only with the arrival of space missions (Mariner and Viking orbiters). The first images of Phobos from Mariner 7 allowed the estimation of the size and of the albedo of the satellite, found to be the darkest object known in the Solar System at that time (Smith 1970). Later measurements confirmed the low albedo of Phobos (7.1 ± 1.2%, Simonelli et al. 1998). Higher-resolution imaging from Viking 1 and 2 revealed the complex surface geomorphology of Phobos, characterized by km-long grooves, craters in a variety of degradation states with variable diameter up to the larger 9.4-km-sized Stickney crater (Thomas et al. 1979).
Several spectra of Phobos were acquired by Mariner 9, Viking 1 and 2 confirming that it is a dark body with featureless spectra in the visible-UV range comparable to low-albedo asteroids Pollack et al. 1978). Thanks to the Phobos-2 mission, spatially resolved spectra were acquired with the imaging spectrometer for Mars (ISM), revealing heterogeneity in reflectance and spectral properties across Phobos surface at spatial resolution of 0.7-1 km (Bibring et al. 1989. Two spectral units were identified, the 'red unit' occupying most of the surface, and the 'blue unit' associated with Stickney crater. The two units differ mainly on spectral slope and brightness, the red one having a higher spectral slope, and the bluer one being brighter and with a lower slope (Murchie et al. 1991;Murchie and Erard 1996). Blue unit is associated with fresher areas close to the Stickney crater and initially considered as fresh ejecta deposit. However, high resolution images from HiRISE instrument on Mars Reconnaissance Orbiter (MRO) revealed that blue unit is seen both inside and outside the crater rim, and that within it several underlying spots of the red unit are present (Thomas et al. 2011), indicating that Phobos has a heterogeneous composition made of blocks of red and blue units (Basilevsky et al. 2014).
No clear absorption bands were detected from Phobos-2 data, and spectra acquired from ground-based telescopes observations indicate no 3 µm absorption feature due to hydrated minerals with an upper limit of 5-10% ( Fig. 1 from Rivkin et al. 2002).
Further high-resolution spectra were acquired by the CRISM and OMEGA spectrometers, respectively, onboard MRO and Mars Express missions (Fig. 1). These instruments acquired high signal-to-noise ratio (SNR) spectra revealing spectral features in both Phobos units ( Fig. 1), at a spatial resolution of about 350 m/pixel (Fraeman et al. 2014). The red unit of Phobos and also Deimos show a weak 0.65 µm absorption feature consistent with desiccated Fe-phyllosilicate such as nontronite (Murchie et al. 2008;Fraeman et al. 2014). An alternative interpretation of this feature includes anhydrous silicates and space weathering effects producing micro-and nanophase iron which acts as Rayleigh scatters generating the 0.65 µm feature (Clark et al. 2012). This feature on Phobos was confirmed from ground-based observations covering the red unit (Murchie et al. 2015), by Pathfinder data (Simonelli et al. 1998), and by ROSETTA mission observations with its OSIRIS cameras (Pajola et al. 2013).
CRISM also detected a sharp absorption at 2.8 μm in both units of Phobos, and on Deimos, with depth varying from 1 to 10% (Fraeman et al. 2014). Note that CRISM data in the 2.70-2.76 μm wavelength range were compromised by a boundary between filters mounted to the IR detector that blocks higher orders from the diffraction grating; these wavelengths were not routinely downlinked by the instrument. This feature is stronger on Deimos and Phobos red unit, and weaker on Phobos' bluer unit. An absorption feature in the 3-μm region was observed on several asteroids, and it is typically caused by a combination of the very strong OH-radical absorption feature and the first overtone of the 6-μm H 2 O fundamental (Rivkin et al. 2002). Adsorbed water has a symmetric stretch, giving rise to an absorption feature at 3.1 μm, and an antisymmetric stretch feature at 2.9 μm. Fig. 1 Left: single scattering albedo spectra (SSA) for the Phobos blue and red units, and for Deimos, derived from CRISM-MRO spectra, extracted from Fraeman et al. (2014). Right: combination of several spectra covering the 0.3-3.6 μm range of Phobos and Deimos, extracted from Rivkin et al. (2002) Structural hydroxyl (OH) that has been incorporated into mineral lattices produces a stretch absorption at 2.7 μm. The depth of the fundamental H 2 O/OH feature depends on many parameters besides abundance of hydrated material, including particle size, albedo, and temperature. In the case of Mars satellites, the strength, position, and shape of the 2.8 μm band is attributed to OH, and there is no clear evidence of absorptions in the 3-3.1 micron region due to adsorbed water or water ice (Rivkin et al. 2002;Fraeman et al. 2014).
No other diagnostic absorptions appear in the VIS-NIR range in either Phobos or Deimos spectra, in particular any related to common ferrous minerals (olivine, pyroxene). The red slope and the absorption bands detected in the visible and NIR region are consistent with those of hydrated asteroids interpreted to have primitive composition (C-, G-, P-, and D-class asteroids, Fraeman et al. 2014). The emissivity spectra of Phobos derived by TES data suggest an ultramafic composition with the presence of phyllosilicates and feldspathoids in some regions (Giuranna et al. 2011). Recently, Glotch et al. (2015) from MIR data, suggest the presence of bound molecular water and carbonate on the surface of Phobos. This result needs confirmation, but these aforementioned observations seem to point to a spectral analogy between the Mars moons and primordial asteroids, thus favoring the asteroid capture hypothesis.

Deimos
The first data of Deimos were also acquired by Mariner 9 and Viking orbiters, suggesting that Deimos was spectrally similar to C-type objects (summarized by Burns 1992), an idea that was challenged by Grundy and Fink (1992) who found much steeper spectral slopes for this body that were inconsistent with carbonaceous chondrites and more consistent with reddish D-type asteroid. As indicated in the previous section and shown in Fig. 1, Deimos shares similar spectral properties with Phobos red unit, in terms of high spectral slope, and it shows the same faint features at 0.65 μm and 2.8 μm, possibly indicating a composition rich in desiccated phyllosilicates. Rivkin et al. (2002) observed Deimos using the IRTF and combined several spectra obtained from the literature to generate a complete spectrum from 0.3 to 3.6 μm ( Fig. 1). While there is strong caveat in the whole nearinfrared range and that the spectral resolution of the data is usually low, they confirm, from their analyses, the anhydrous nature of the surface of Deimos, placing limits on the 3-μm band depth at roughly 10-20%. They also state that spectral analogs for the Martian satellites are difficult to find in the asteroid belt, and it seems that the outer-belt D asteroids have spectra which bracket spectral pattern of Deimos.
The Viking data suggested the trailing side is about 10% brighter than the leading side. This leads us to assume a modest difference in materials composition or properties on the two sides of Deimos. This is consistent with the successive Hubble Space Telescope data that also showed slight leading/trailing differences on Deimos (Zellner and Wells 1994). Differences exist in the visible albedo and are correlated with topography rather than found in more coherent units as on Phobos (Thomas et al. 1996). High spatial resolution color observations of Deimos were not obtained by Phobos-2, and the Viking data which exist are in a very restricted wavelength range. The Imager for Mars Pathfinder (IMP) obtained disk-integrated, visiblerange spectrophotometric data for the sub-Mars hemispheres of Deimos (Murchie et al. 1999), showing Deimos is quite homogeneously red, consistent with the results of Grundy and Fink (1992).
Although there may be small differences in the material properties of Phobos and Deimos, the drastic visual difference most probably derives from the smoothing effects of the large impact that formed the 10-km wide southern polar depression on Deimos (Thomas et al. 1996). This impact is relatively much larger than the Stickney impact crater on Phobos. Thus some previous suggestions that the lack of grooves on Deimos stemmed from a lack of large craters (Thomas et al. 1979) need revision, either in the fracture origin of grooves or on the timing of ejecta re-accretion covering the effects of fracturing. Extend these analyses in the near-infrared spectral range is fundamental in order to explore the possible spatial variation of the chemical composition of Deimos and answer the still open questions. Fraeman et al. (2014) proposed two hypotheses to explain the spectral features on Phobos and Deimos. The first invokes the presence of highly desiccated Fe-phyllosilicate minerals inherent to Phobos and Deimos, consistent with the presence of primordial material in the Moons' formation region, and the second appeals exogenic processes associated with the solar wind which may have created Fe and OH on the moons' surface. However, there are no telescopic-or space-based data sets with sufficient spectral range and SNR levels to independently validate this feature's presence. OMEGA observations of Phobos do cover the necessary wavelength range, but a detector boundary with an unknown offset near 2.7 μm makes it impractical to detect a weak absorption. Observations of Phobos and Deimos with MIRS will allow to check the presence of this absorption feature and its variation on the whole Phobos and Deimos surfaces, and possibly discard one of these hypotheses described by Fraeman et al. (2014) by characterizing the 2.7-μm absorption band features (i.e., band peak position and shape; Donaldson et al. 2016, andPieters et al. 2009, for instance). At further extent, detection of phyllosilicates would imply that the origin of Phobos from a Giant impact is less likely. Yet, a global assessment of composition is needed to identify and evaluate the properties and geologic context of different materials. As presented in Pieters et al. (2014), diverse materials are known to exist across the surface of Mars moons and it is essential to document their characteristics, distribution, and the spatial relationships between them. MIRS will be on front seat to fulfill this goal.

MIRS in the exploration of the Phobos and Deimos surface
MIRS will observe Phobos and Deimos in the 0.9-3.6 μm range with a spectral resolution better than 20 nm. For Phobos, MIRS will acquire spectra at SNR > 100 and at a spatial resolution of about 20 m (for ± 30° latitude) during the QSO-M (quasi-satellite orbit-medium altitude, see Table 3) global survey. A higher spatial resolution of 1 m will be reached over an area within 50 m from the selected sampling sites. The spectral radiometric absolute accuracy is expected to be of 10%, and the relative accuracy of 1%. The high SNR and unprecedented spatial resolution achieved by MIRS will permit to fully characterize the composition and mineralogy of both the red and blue units on Phobos and further investigate the local compositional heterogeneity associated with the different geomorphological features across Phobos surface. The capability of MIRS fulfills completely mission requirements for Phobos. It will also possibly provide new insights into space weathering processes thanks to resolved spectral observations of small fresh craters.
MIRS is expected to spectroscopically detect and characterize, if any, water (ice) (absorption bands at 1.5, 2.0 and 3.0-3.2 μm), hydrous silicate minerals (features at 2.7-2.8 μm, and minor overtones at 1.4 and 1.8 μm), or anhydrous silicates (bands in the 0.9-1 and 2.0 μm regions). If possible, also to measure organic matter (3.3-3.5 μm). Thanks to high SNR, it will also allow a detailed characterization of the absorption bands detected, permitting by the precise investigation of the band center, depth and area to well constrain the mineralogy, species abundances and composition of Phobos. These unprecedented data will allow scientists to cast light on the origin of Mars moons.
For Deimos, MIRS is expected to spectroscopically map major regions at a spatial resolution better than 100 m and at SNR better than 100 below 3.2 µm in less than 2 s, and detect the same major absorption bands as observed in Phobos. Such observational conditions will allow us to spatially probe the surface of Deimos and constrain the surface composition as well as search more efficiently possible weak spectral features (a full characterization of the bands is possible for band depth as weak as 3%, estimated from a Monte Carlo approach, and based on synthetic spectra reaching SNR of about 100), as in the case of Phobos. The data will also favor the search for physical and chemical heterogeneity that could be linked to topography or linked to the Martian phase aspect. Do the particle size or the chemical content of the leading and trailing sides are different, as suspected from the observation performed with HST in the visible range? The lack of current data of Deimos, especially in terms of spatially resolved spectral data, considerably reduce our ability to constrain the surface properties of it, and even if Deimos is not the target for the sampling, we will clearly be able to take advantage of this mission to considerably improve the physical and chemical knowledge of the surface of this satellite, and bring new insights about the history of the Martian environment.

Temperature on Phobos and Deimos
Temperature is a fundamental physical parameter that affects a number of physical processes important for remote sensing of planetary surfaces, including the reflected and emitted radiance. On solar system bodies the temperature is primarily determined by the heliocentric distance, its time derivative and by variations in the illumination caused by the cycles between day and night and seasons. In addition, on atmosphere-less bodies, such as Phobos and Deimos, temperatures are also controlled by the physical properties of their surfaces, the most important one being the albedo, the roughness, the heat capacity, the density, and the thermal conductivity of the surface. The latter is function of the porosity of the material and grain size in the case of regolith. When illumination conditions vary periodically and thermal conductivity is independent of depth, it is possible to define the thermal inertia as the square root of the product of thermal conductivity, density, and heat capacity, and show that surface temperatures are governed by the value of the thermal inertia (e.g., Spencer et al. 1989). When its value is lower than some hundreds J m −2 s −1/2 K −1 at heliocentric distances of few AU, temperatures can vary dramatically between winter and summer and between day and night (e.g., Delbo et al. 2015).
The latter is the case of the surfaces of Phobos and Deimos that express temperatures ranging between ~ 50 K and ~ 300 K (Kührt and Giese 1989). These large temperature variations are in part due to the low value of the thermal conductivity (Gatley et al. 1974) or thermal inertia (Lunine et al. 1982;Kührt et al. 1992). The large temperature variations are also caused by the complex orbital geometry of these satellites, including their synchronous rotations with their orbital periods, the eclipses they are subject to, and the seasons. In the case of Phobos, which orbits Mars at an altitude of only ~ 6000 km from the surface of red planet, the reflected scattered light and thermal emission from Mars are also important in determining the surface temperature of the satellite (Kührt and Giese 1989). The low albedo of these satellites and its variations across the surface, with values ranging between 0.05 and 0.10 in the visible band (Fraeman et al. 2014), also contribute to temperature differences.
It is known that temperature can affect the near-infrared spectral features (e.g., Singer and Roush 1985;Roush and Singer 1987;Lucey et al. 1998;Moroz et al. 2000). For instance, at low temperature, the visible and nearinfrared bands narrow and the band contrast increases compared to their measurements at higher temperatures. The change in the spectral band properties is related to the composition of the surface and the grain size of dust potentially covering it. It is therefore conceivable that spectral observations of the martian moons' surfaces at different temperatures (e.g., different time of the day) will help constraining their composition.
By means of computer codes that implement the socalled thermal (or thermophysical) models (e.g., Delbo et al. 2015 for a review) we calculate model temperatures, or the corresponding emitted infrared radiance as a function of physical parameters such as the albedo, roughness, and thermal inertia, which are adjusted until best match between the model predictions and the observations are obtained, thereby constraining the physical properties of the surface (Delbo et al. 2015). Modeling of surface temperatures will be essential to estimate the heat and radiation environment for near surface operations of MMX, which is important to support the sampling sites selection and operations.
On the other hand, the physical parameters estimated from the analysis of the infrared observations will be interpreted in terms of grain sizes of the surface regolith (e.g., Gundlach and Blum 2013; Sakatani et al. 2017;Ryan et al. 2020) or porosity of surface rocks (as in Grott et al. 2019), using radiative transfer models. This will contribute to provide geologic context of Phobos returned samples. Thermal inertia (or conductivity) values have been derived for Phobos by several authors using measurements from different space missions and also from ground-based data: Gatley et al. (1974) found thermal conductivity values that would correspond to a thermal inertia < 15 J m −2 s −1/2 K −1 using Mariner 9 radiometric data, while Lunine et al. (1982) obtained a range of thermal inertia values between 40 and 70 J m −2 s −1/2 K −1 from Viking data, and Kührt et al. (1992) determined thermal inertia values between 20 and 40 J m −2 s −1/2 K −1 . All these values are quite low and would imply that the surface is made of a fine-grained particulate material (regolith). Kührt and Giese (1989) proposed that thermal properties dependent on temperatures must be used on Phobos and Deimos. Bandfield et al. (2018) found that the best effective thermal inertia value is 150 J m −2 s −1/2 K −1 , a value that they claim to be still compatible with particulate regolith. The same authors also reported a more successful model fit where thermal inertia varies with depth, from 50 at the surface to 1000 J m −2 s −1/2 K −1 in the shallow subsurface. Smith et al. (2019) derived the first disk resolved map of the thermal inertia of some regions of Phobos. They found a mean thermal inertia value of 42 J m −2 s −1/2 K −1 with a range of values between 20 and 100 J m −2 s −1/2 K −1 .
MIRS will measure the spectral radiance of the surface within the instrument footprint (Fig. 2). This is essentially the reflected solar spectrum with superimposed minerals absorption bands. At wavelengths longer than about 2.5 μm the so-called thermal tail kicks due to Phobos thermal radiation become more important than the reflected sun-light. The thermal tail is a strong function of the temperature and thus MIRS data can be used to derive the surface temperature as long as the latter is high enough to produce a radiance signal above the instrument sensitivity. The thermal tail is also function of the infrared emissivity. Modeling the thermal tail is an important aspect of the MIRS data analysis, in particular in the region around 2.7-2.8 μm where the reflectance spectra will be investigated for the presence of the phyllosilicate and water bands.
Despite MIRS is not a proper thermal instrument due to its limited wavelength extension in the infrared, the thermal tail can be used to assess the surface temperature and its spatial and temporal variations. From temperature measurements, information about the surface thermal inertia of Phobos will be derived. This has been done at similar wavelengths with the OSIRIS-REx Visible and InfraRed Spectrometer (OVIRS; Reuter et al. 2018) onboard NASA's OSIRIS-REx asteroid sample return missions (Lauretta et al. 2015) to successfully determine the thermal inertia of the surface of the asteroid Bennu with spatial resolutions of some meters (DellaGiustina et al. 2019).

Mars atmosphere
Mars atmosphere is strongly coupled to the surface. This is especially the case regarding volatiles species in strong interaction with the polar caps. This effect, together with seasonal variation of the radiative balance, induces a high seasonality in the volume mixing ratio of species such as H 2 O, CO or O 2 . The suspended dust particles from the surface contribute to localized heating of the atmosphere and have a significant impact on the global climatology on Mars, especially during the dust storm seasons. All of these atmospheric components (Table 1) present significant signatures in the MIRS spectral range, from 0.9 to 3.6 µm (see Fig. 3).
While being the main Martian atmospheric component, CO 2 presents a strong seasonal cycle due to its condensation at the polar caps, which induces strong fluctuation of Martian surface pressure. On top of that, small oscillations of the surface pressure, which are key diagnostic to meteorological events, were observed by OMEGA onboard Mars Express Spiga et al. 2007). MIRS will have the capability to monitor CO 2 and thus pressure through inversion of its 2.0 µm band as it was performed with the OMEGA data. While the large-scale annual variation of water vapor has been stable for more than 20 Martian years (Montmessin et al. 2017), questions remain on the interaction between water vapor, ice clouds and dust on regional scales. The spatial resolution of MIRS for Mars observations (< 10 km) will extend the existing records on these observations and better constrain the interdependencies between these major atmospheric components. It is noteworthy that the MIRS measurements provide the first opportunity for investigating the diurnal variation and the transport processes of water vapor with time-resolved pictures of the atmospheric phenomena from the equatorial orbit.
Most of the information about water vapor on Mars has been obtained for more than two decades from MAWD onboard Viking 1 and 2, TES onboard Mars Global Surveyor, SPICAM, OMEGA, and PFS onboard Mars Express, CRISM onboard MRO missions (e.g., see review for SPICAM in Montmessin et al. 2017). These observations revealed a global cycle of water vapor, which is mainly driven by dynamics and sublimationcondensation process between the atmosphere and the polar caps, with a column integrated abundance peaking up to 40 pr-µm during northern summer in the polar region. On the other hand, local events such Left: preliminary temperature calculation on Phobos using the Gaskell shape model (Gaskell 2011). The model was reduced to 98,000 facets to speed-up calculations. Temperatures decrease from yellow to red to blue. Right: spectral radiance calculated using the thermal model and assuming that the reflected light component is a reddened solar spectrum. The spectral radiance is normalized to 1 at the peak of the curve. Notice the strong thermal tail that arises in MIRS relevant wavelengths and its strong dependence on temperature as surface-atmosphere interaction and water vapor transport due to so-called rocket dust storm, that can lift dust rapidly and efficiently to high altitude (Spiga et al. 2013), are not well understood. Characterization of such local events requires high-frequency monitoring of water vapor with different local times. MIRS will have the first capability to monitor water vapor at lowto mid-latitude on a daily basis using its 2.6 µm band (Maltagliati et al. 2008). The adsorbed water in the regolith also can be monitored by the 3.0 µm band (Audouard et al. 2014). Water ice clouds also present large seasonal variations and spatial distribution, associated with both the water cycle and general atmospheric circulation ). The most predominant ice clouds formation is found in a belt around the equator during Mars' aphelion season. These clouds are due to the release of large amount of water from the northern polar cap during spring and summer: this induces an increase in water vapor content at mid-to low-latitude where it can be uplifted by the Hadley circulation and this uplifting brings air charged in water to higher altitudes where low temperatures and pressures favor cloud condensation, allowing the seasonal persistence of this so-called Aphelion cloud belt. Water ice clouds can also be found at higher latitude during winters over polar caps (Fedorova et al. 2006;Montmessin et al. 2017;Vincendon et al. 2011). Water ice clouds will be identified primarily through their spectral features between 0.9 and 3.6 µm (Olsen et al. 2017).
In addition to the water ice clouds, MIRS may be able to investigate the mesospheric CO 2 ice clouds. Recent observations have revealed the presence of the CO 2 ice clouds at remarkably high altitudes (above 40 km; mesosphere) (e.g., Montmessin et al. 2006;Aoki et al. 2018). However, the formation process of these peculiar clouds has not been measured even though it has been suggested that the clouds form in supersaturated pockets of air created by the interference of thermal tides and gravity waves (e.g., Clancy and Sandor 1998;Spiga et al. 2012). Its monitoring in the morning or evening will provide a first insight on its formation process. Vincendon et al. (2011) proposed a procedure to process CRISM observations to discriminate H 2 O and CO 2 ice clouds by combining its visible images and near-infrared spectra and demonstrated its usefulness by comparing it with previous observations. This  (Ignatiev et al. 2005). A typical condition of Mars atmosphere is assumed. The surface albedo of Mars is assumed to be 0.15 (uniform). The thick curves in the figure show the spectra for the clear sky condition, and the dashed curves include the effects of dust or water ice clouds. For gases, CO 2 , CO, and water vapor are calculated separately (red, orange, and blue curves, respectively). The calculated synthetic spectra are convolved with the instrumental line shape of MIRS that is a Gaussian function whose FWHM is 20 nm. The sampling step is 10 nm procedure will be tentatively applied to coordinate MIRS and OROCHI observations to investigate mesospheric CO 2 ice clouds. The relatively long lifetime of CO has a capability to trace the atmospheric dynamics on Mars. Its detection and quantification thus would allow for further refinement of current atmospheric dynamical models (e.g., Daerden et al. 2019). CO presents only a weak band at 2.3-2.4 µm within MIRS wavelength range. However, previous observation with OMEGA and CRISM, both instruments having similar SNR as MIRS, demonstrated the detectability of this compound. In addition, while the formation of CO is intrinsically linked to the CO 2 cycle, its non-condensation during polar winter, contrary to CO 2 , induces a relative enrichment at mid-to highlatitudes. While we expect global detection of CO, the equatorial orbit of the MMX mission will make Hellas Planitia, where clear seasonal variations due to CO 2 condensation during southern winter have been evidenced by OMEGA and CRISM, our prime target for CO monitoring (Encrenaz et al. 2006;Smith et al. 2009).
O 2 is an oxidant tracer for atmospheric photochemistry on Mars. During the diurnal variation, it is formed through ozone UV photodissociation in the lower atmosphere (Fedorova et al. 2006). This induces a dayglow at 1.27 µm whose mapping thus provides a tracer for O 3 and its variations. O 2 dayglow shows its maxima during early spring at high latitudes, while low latitudes show a maximum at aphelion due to a peak in ozone concentration related to lower abundances of hydrogen radicals (Lefèvre et al. 2004). At high latitudes, O 2 can also be produced by CO 2 photodissociation, as it can be observed on the nightside at high latitudes via recombination of O brought in a downwelling general atmospheric circulation on the dayside in the upper atmosphere (Bertaux et al. 2012). This nightglow emission is a unique tracer for downward advection transport mechanism to constrain the mesospheric dynamics. The dayglow of O 2 formed via O 3 dissociation will be detectable with the MIRS instrument in nadir observations but also measurable at limb, which will provide its vertical profiles as well as an index of water vapor vertical profiles (Clancy et al. 2017). The nightglow would require limb observations at high latitudes that could technically be doable by MIRS, but may be extremely challenging to perform given MMX equatorial orbit. Another possible tentative measurement within the MIRS wavelength range would be the one of OH Meinel bands at 1.45 and 2.9 µm as detected by the CRISM instrument ). However, the detection of these extremely faint bands would require the accumulation of observations of the polar night limb and thus be subjected to the same limitation as the O 2 nightglow observations. As such, and while these two observations will be tentative, they are not considered a prime target for the MIRS instrument.

Mars surface and dust
Although the specifications of the MIRS imaging spectrometer are quite similar to previous instruments, such as OMEGA aboard MEX, the orbit of MMX can offer some advantages for global mapping of the red planet. The expected geometric resolution is ~ 2 km at the footprint on Mars surface, so a spatial binning can be possible to increase the SNR and save data volume. Being placed on the same orbit as Phobos around Mars, at about 6000 km of altitude, the spacecraft will have a ~ 7 h orbit around the planet, allowing to complete a global mapping of the surface of Mars within a few days. In addition, the well-controlled scanner system enables the specific pointing of the instrument, thus allowing MIRS measurements to obtain wide spatial coverage in hourly timescale. Both these observing modes will be very useful to follow the temporal evolution of the atmospheric aerosols with a meso-scale spatial resolution. In that respect, MIRS will complement adequately the instruments that are already observing the Martian surface and atmosphere from orbit. Its coverage will be more similar to that obtained with the MEX/OMEGA instrument, with the interesting capability to analyze the same area at high temporal resolution (5 times every 30 min), or to obtain almost global mapping of the sunlit hemisphere with 5 h of observations. This capability will allow to survey rapidly changing atmospheric events. Based on the similarities with previous observations, existing techniques that have already been applied to analyze the spectral cubes from the atmosphere of Mars with previous missions can be also applied to the MIRS data and improved.
The analysis of the spectra taken on the slopes of large volcanoes, such as the Olympus Mons can help to probe different atmosphere thicknesses and detect the column density of H 2 O, CO 2 and dust constituents in the Martian atmosphere (Combes et al. 1991;Rosenqvist et al. 1992). To go further into the analysis of the different atmospheric constituents, Mars limb observations should allow to directly recover at first order the spectral signature of aerosols present in the atmosphere though it will depend on the properties and composition of the aerosols at different altitudes (Erard et al. 1994). Furthermore, limb observations by MIRS have a capability to investigate the vertical structure of aerosols in a global view. It is interesting to note that high-altitude (above 60 km) H 2 O was first identified by SPICAM occultation (Maltagliati et al. 2011(Maltagliati et al. , 2013 especially in the southern summer, which happens to be a dusty season. They reported the links between such high-altitude H 2 O and aerosols in their vertical profiles within a short time-scale. This implies the importance of aerosols for key processes in the Martian water cycle and climate as a whole. MIRS will provide further information of the new pathway of water loss proposed by recent studies Heavens et al. 2018).
A better estimate of the reflectance spectrum from the light backscattered by atmospheric aerosols can be obtained by comparing the successive observations of identical zones on the surface of Mars within a short time and with varying phase angles. Such a set of measurements permits to decorrelate the surface reflectance signal from the light scattered by the aerosols suspended in the atmosphere of Mars and gives information on the aerosols size and composition properties. Such an analysis for the ISM data (on-board Phobos-2 spacecraft) taken over Tharsis and Ophir Planum showed that the spectra thus obtained are consistent with a particle size distribution having an effective radius of 1.2 ± 0.2 μm (Erard et al. 1994). Features of the spectrum indicate the presence of water ice, and possible absorption due to clays and/ or sulfates. The overall parameters that can be obtained from such an analysis include the scattering albedo, the phase function of the aerosols, the optical thickness, τ, of the atmosphere and the aerosols size distribution with possible indications on their shape. Vincendon et al. (2007Vincendon et al. ( , 2009) have shown that the introduction of a radiative transfer model can improve on the first estimates for these values, which was demonstrated as an application to the OMEGA data. It was also possible during Martian dust storms to obtain typical properties of the Martian suspended dust in terms of scattering albedo and optical depth by applying independent components multivariate methods to separate the spectrum signal of the aerosols from the one from the surface and the atmospheric gas (Määttänen et al. 2009). An improvement of the method used to correct for aerosols light scattering by multi-angle analysis over 11 different viewing geometries to retrieve the Mars surface photometric properties was developed for CRISM by including non-isotropic surface materials reflectance Fernando et al. 2013). Further CRISM observations of the limb of Mars allowed to retrieve locations and properties of the aerosols on the atmosphere of Mars . To go further, combined observations of CRISM limb observations, MCS and MARCI were used to determine the composition and sizes of Mars mesospheric aerosols and ice clouds dynamics .
Finally, the combination with lander infrared measurements can allow to confirm and better constrain the local properties of atmospheric dust from the ground (Ockert- Bell et al. 1997). Assuming that Mars rovers are still active on the surface of the planet during the mission, it can be envisioned to combine the orbital measurements with those that will be made locally by using the infrared channel of the SuperCam instrument on Mars, for example Wiens et al. 2021) and the MastCam-Z multispectral camera observations (Bell et al. 2021). All these studies will improve our understanding of the optical properties of the Martian aerosols, their evolution with time and the exchanges between the surface and the atmosphere of the planet. In the end, understanding fully the contribution of aerosols to the measured reflectance from orbit will enable an improved understanding of the scattering properties of the Martian surface.

Push-broom spectrometer
MIRS is a spectrometer that uses the well-known pushbroom acquisition principle (Fig. 4). A single detector acquisition (2D matrix) provides the image of a strip in one direction (spatial), and the spectrum of each point of the strip in the second direction (spectral). The second spatial dimension results from the motion of MIRS Line of Sight in the along-track direction either thanks to the spacecraft speed or by actuation of MIRS scanner.
The start of each image acquisition has to be adjusted so that: -All images are as contiguous as possible: no overlaps, no holes. -Integration time is sufficient to guarantee the required SNR, and short enough not to degrade the spatial resolution. We define the subsatellite point as the projection of MMX-Z axis on the celestial body of interest. Subsatellite ground speed varies and might not be adapted to MIRS acquisition principle. For Phobos observation on low-altitude trajectories, ground speed can be too high to ensure sufficient integration time between 2 images. However, for Phobos observation during descent phases, ground speed is nearly null. For Mars observation, ground speed is low enough to cover large zones.
In order to adjust ground speed of MIRS Line of Sight projection on celestial body of interest to these situations, MIRS is equipped with an along-track scanner with ± 20° optical amplitude. This scanner can also be used to perform observations of zones with specific phase angle and local solar time by commanding a constant along-track bias. Meanwhile, the cross-track latitudes can be reached using the spacecraft maneuvers.

Instrument concept
MIRS is an infrared imaging spectrometer (Fig. 5). Based on scientific objectives, the MIRS technical requirements have been derived as presented in Table 2. The spectral dispersion is obtained with a low-density groove grating working in order 1. The telescope is composed of two free form mirrors (highly aspherical mirrors) focusing the scene target on the entrance slit of the spectrometer.
The collimator is also composed of two free form mirrors projecting the entrance slit at infinity. The grating is located at the pupil of the instrument. Then two dioptric objectives are used. The first one projects the spectral image on a filter that sorts the grating orders. The second one projects this spectral image on the detector, but also images the pupil on a cold stop in order to limit the background flux due to the thermal emission of the spectrometer. The detector and the cold stop are encapsulated in a cryostat and cooled down to 110 K (120 K for the cold stop).
A shutter is placed in the slit plane in order to close the spectrometer cavity after the telescope and acquire background images that can be subtracted to science data.
A cover is positioned at the entrance of the instrument to limit dust pollution when landing on Phobos.
MIRS is mounted on the exploration module (Fig. 6) and is connected to the MDP for data management (TC/ TM and image compression) and the PDCU for power supply (Fig. 7).
MIRS consists of three subunits: -An Optical Box (OBOX) (Fig. 8A, B) containing all the optical components (Fig. 8C): mechanisms (scanner, shutter, dust cover), a telescope, a spectrometer, a calibration source, and a detector package with its dedicated proximity electronic. MIRS OBOX is mounted on a dedicated bracket outside MMX eXploration module, with a nadir pointing (Fig. 6) -An Electronic Box (EBOX) (Fig. 9A) containing the low-voltage power supply board (LVPS) (Fig. 9B) and the interface control unit board (ICU) (Fig. 9B). MIRS EBOX is located inside MMX exploration module (Fig. 6) -An Interconnection harness linking the EBOX to the OBOX.
The OBOX is thermally uncoupled from the interface bracket. Two radiators ensure a passive cooling of the OBOX: -1 Dedicated to the spectrometer. An open-close mechanism called dust cover is implemented on the MIRS OBOX entrance to protect the spectrometer from dust contamination (Fig. 8A).
The optical alignment is mechanically ensured with a stiff foot called Reference foot and a flexible one, i.e., a bipod called Axis foot. The third foot is mounted with regard to the Detector package center of mass. Figure 8B shows the external mechanical interfaces on the MIRS OBOX, while Fig. 8C shows the inside of the OBOX and the optical layout.
The EBOX (Fig. 9A) is the electronic box of the MIRS-F spectrometer. Its mass is about 1.2 kg. This unit is thermally coupled to the spacecraft descent module and is black anodized on its external surfaces. Figure 9 shows the two-subunits of the EBOX, i.e., the LVPS and the ICU.

MIRS-PE-007
Modulation transfer function shall be higher than 8% at Nyquist frequency in the spatial direction MIRS-PE-008 MIRS boresight orientation with respect to MMX spacecraft shall be known with an accuracy better than ± 1.4 mrad MIRS-PE-009 MIRS boresight relative orientation between successive acquisitions of an observation sequence shall be known with an accuracy better than ± 0.17 mrad Note: an observation sequence is: A complete observation run when the scanner is not used (e.g., Phobos global mapping) The observation phase corresponding to a scan when the scanner is continuously rotating during images acquisition (e.g., Mars, Phobos landing)
The instrument will produce 2D data cubes (λ, y) with typically images of 256 × 200 px 2 .
MIRS has a data interface with the Mission Data Processor (MDP) which also has a data interface with the MMX spacecraft Management Unit (SMU) and the Mission Data Recorder (MDR). All telemetry data will be first transferred into the MDR via MDP without any compression. Telemetry data will be reloaded from MDR to MDP for the compression process (CCSDS122.0-B-1 standard).

MIRS mechanisms
Three mechanisms are implemented in MIRS.

Scanner mechanism
The along-track scanner mechanism (Fig. 10) is used to adjust ground speed with respect to the line of sight and also to reach specific phase angles. It is a singleaxis scanner. A cam and lever mechanism transforms the rotation of the stepper geared motor to obtain the mirror movement. The full mechanical angular science stroke is [− 10°, + 10°] (which allows a [− 20°, + 20°] optical range for MIRS line of sight). The equipment is also able to target a calibration position at + 17°. Mirror hinge is made with 2 cantilevered flexural pivots. No angular sensor in the equipment. The main characteristics are:

Shutter mechanism
The MIRS shutter mechanism (Fig. 11) masks the spectrometer entrance slit (1 × 10 mm 2 ) in order to allow background acquisitions that can be subtracted to the science data. The design has a high space heritage and has already been qualified and flown on different space instruments: DIRS/Cassini-Huygens, OMEGA/Mars Express, VIRTIS/Venus Express and Rosetta, SuperCam/ NASA Mars 2020 mission. The mechanism concept is based on a mono-stable actuator using steppers and permanent magnets: it keeps one stable position when the actuator is un-energized. The main characteristics are: • Dimensions: 20 × 20 × 24 (mm).

Cover mechanism
The MIRS cover mechanism (Fig. 12) closes the optical entrance (~ 50 × 90 mm 2 ) to protect optics from dust during landing phases on Phobos surface. The design is based on a space qualified Phytron moto-reduction system and a 3 bars mechanism. It allows opening and closing the cover (5 actuations). The cruise stable position is obtained with actuator in un-energized mode when the connecting rods are aligned. The cover mechanism is also equipped with an additional HDRM (hold down release mechanism) to prevent failure mode due to dust intrusion inside cover or moto-reduction system. The main characteristics are: • Dimensions: 200 × 130 × 45 (mm). • Mass: < 500 g.

MIRS detector
The MIRS detector (Fig. 13) is based on COTS SWIR hybrid CMOS detector (Neptune) developed by Lynred and delivered with a cryocooler by Ricor (K508 reference). This assembly has already been used for several space missions and is fully qualified. The estimated lifetime of this cryocooler (of about 10,000 h) is well adapted to the MIRS operation lifetime requirements (estimated The detector format is 500 columns by 256 lines with square pixels of 30 µm pitch, sensitive to 0.45-3.8 µm made of photovoltaic diode in HgCdTe coupling via indium bump to a ROIC allowing different types of readout, gain and integration time control. Its nominal working temperature is 110 K, with active cooling. To reduce background a cold pupil around 120 K is physically present in the cryostat. Thanks to backside illumination and anti-reflection coating optimized in the optical bandwidth, the quantum efficiency is up to 0.8, insuring high sensitivity, and low PRNU (< 20%). The dark current @110 K is given under 62.5 ke/s (n-on-p technology).
The ROIC allows the readout of one image frame while the next is under integration, two independent readout output may be programmed to decrease the readout time. The gain of each line may be programmable with low and high gain to optimized the readout noise with the signal, typically the noise is < 150 e with integrated charge of 0.5 Me and < 340 e with 2.5 Me, respectively. The maximum readout frequency of each single output is 8 MHz, allowing a minimal readout time of 3.2 ms for the whole area. The windowing readout mode is possible by programming the serial register of the ROIC, 1 indicates a line to be read (then the whole pixel of the line as to be red) and 0 indicates a line to be swapped out (during the programming of the ROIC, the gain of each line is chosen).
For the purpose of MIRS, only one half of the total sensitive surface will be used, leading to a readout of 250 × 256 pixels. The register is programmed at each power-on of the ROIC, the integration time is then managed by a single input signal. Proximity electronics inherited from previous instrument will only use one output @1 MHz with 12 bits ADC.
The detector has 2 integrated temperature sensors, one of them is used for the control of the cryocooler, and both will be monitored. A specific proximity electronics close to the cryostat is connected to the detector via a flex-ribbon to minimize the thermal link. It will insure the management of the bias polarization of the detector, the digitalization of the video signal and temperature sensor and the management of the cryocooler. A specific thermal strap links the cryocooler to a dedicated radiator to dissipate the 4-8 W needed to cool down the detector head.

Thermal architecture
OBOX is an external onboard subunit and thus sees spacecraft external thermal environment (Fig. 14). OBOX is located on an MMX bracket on which it is thermally insulated in a semi enclosure (on the + Y side of the spacecraft (S/C)). OBOX lateral sides are radiatively insulated, thanks to MLIs. OBOX is split in two thermal zones: a "warm" zone containing the cryocooler and a "cold" zone containing the spectrometer and the optical components.
The cryocooler is a Ricor K508 Stirling rotary cryogenic cooling machine which maintains the FPA at 110 K during science operations. OBOX + Y side has a "warm" radiator which evacuates the cryocooler heat dissipation. The cryocooler and "warm" radiator assembly are thermally insulated from the spectrometer.
A "cold" radiator is providing to the spectrometer an operating temperature below 250 K. OBOX mechanisms are thermally insulated from the optical bench and directly linked to this radiator. A survival system composed by heaters and mechanical thermal switches prevents OBOX from being colder than 230 K when MIRS is not operating.
EBOX is an internal onboard subunit. It is located inside MMX cavity and is thermally coupled to a S/C internal panel thanks to several screws and to thermal filler. EBOX is black painted in order to be radiatively coupled with S/C internal cavity walls.
Both MIRS subunits have one thermometer read by MMX system at any time, and several internal sensors read by MIRS when operating.

Electrical architecture
MIRS is composed of three electronics boards: one located inside the optical box (OBOX), and the two others inside the electronics box (EBOX). These electronics boards (Fig. 15) are: • The low-voltage power supply board (LVPS) which generates, from an unregulated 32-50 V primary bus (provided by the power distribution control unit), all secondary voltages needed to supply the different subsystems of MIRS instrument. This board is located inside the EBOX. • The interface control unit board (ICU) which accommodates a reprogrammable FPGA to manage raw data handling and storage, and TM/TC communication with the mission data processor (MDP) through a redundant SpaceWire link (SpW). The ICU board controls all mechanisms (scanner, dust cover, shutter), switch on/off the calibration lamp, controls the Neptune detector and its readout electronics. The ICU includes also two full bridges motor drivers to control step motors dedicated to dust cover and scanner, and a slow ADC for housekeeping monitoring (component temperatures, sec-ondary voltages and primary current). This board is also located inside the EBOX. • The readout electronics board of the MWIR detector. This board, called proximity electronics (PE), which accommodates a low-power 12-bit ADC at 1 MSps to sample one video output of the detector. A video amplifier is implemented on this board to match the dynamics of the detector signal to the ADC full-scale. The ADC integrates an 8-channel multiplexer, and the other channels are used to monitor the temperature of each mechanism and the spectrometer through a PT100 temperature sensor. A particular attention is given to PE board in order to provide very stable voltages for detector polarization by the use of linear regulator and common mode and differential filters. This board is located inside the OBOX, as close as possible to the detector.
Three shielded wire harnesses are used to connect the EBOX to the OBOX. The first one is used to provide all power signals to each mechanism, the two others connect directly the ICU board to the PE board for data and power signals. Low-voltage differential signaling (LVDS) transceivers are implemented on ICU and PE boards for digital data link transmission.
An 8-layer flex PCB is used to connect the PE board to both detector connectors: the first one (A) for analogue signals and the other one (B) for digital signals.

Digital electronics
The MIRS digital electronics is based on a FPGA device (Fig. 16), a NG-Medium from NaNoXPlore, and an embedded processor (soft IP, 8051 family). The main functions to be handle are communication with the MDP through Space Wire RMAP protocol, command dictionary support with TC/TLM (science data, housekeeping telemetry, and status telemetry), control of MIRS subsystems (detector and cryocooler, shutter, calibration lamp, scan mirror, dust cover), algorithm for image auto-exposure, data acquisition (2D images and housekeeping), and data processing.
The specificity of the design is its ability to perform the digital signal processing on 2D images (including image stacking, dark subtraction, region of interest extraction, spatial binning, spectral binning), on the fly, during detector read out and image transfer on Space Wire link. Additionally, detector readout and image transfer are performed during detector integration, in order to avoid (or minimize, depending on the integration time selected) dead time in the observation. The embedded processor is responsible for determining the best detector integration time and stacking factor to optimize SNR and avoid blurred images (Fig. 17). Also the processor has the capability to compute and update the scanner parameters (e.g., speed) during observations.

MIRS performances have been simulated considering:
• Spectral resolution and spectral sampling. • Spatial resolution.
In order to meet the spectral resolution requirement, the slit width is chosen so that its image on the focal plane convolved by the instrument PSF is spread on two pixels. The current optical design performances yield to a spectral sampled resolution less than 22 nm. The spatial resolution is linked to the optical quality specified in terms of MTF (modulation transfer function) amplitude at Nyquist frequency. The requirement is > 8% over the field of view and spectral range. The instrument MTF is assessed by product of the optical FTM, the detector FTM and the smearing FTM.
Values @ Nyquist frequency: The SNR is limited by the maximum integration time allowable for an observation.
To calculate the maximum integration time, we need to consider two limits: • Detector saturation that limits the elementary acquisition integration time, • Pixel scroll due to MIRS movement with respect to the target in order to preserve the spatial resolution that limits the number of accumulations of elementary acquisitions. For Phobos observations, the typical SNR profile is reported in Fig. 18. Note that SNR can be efficiently improved by integrating photons during a longer duration. For example, on QSO-M, the pixel scroll is higher than 3 s on 85% of the orbit.
For Mars observations, the different SNR profiles for a single accumulation are shown in Fig. 19. Depending on the various cases, the integration time is between 75 and 130 ms. The geometric nadir resolution is 2 km on ground (requirement is 10 km) so a spatial binning is possible to increase the SNR and save data volume.

MIRS observations
For several reasons, among which the fact that Phobos sphere of influence is very close to its surface, pure Keplerian motion of a satellite around it is not possible. Quasi-satellite orbits (QSOs) are therefore used and correspond to an MMX orbit around Mars with carefully set differences with Phobos orbit in order to induce a relative movement of MMX spacecraft around Phobos. Altitude varies during each revolution of MMX around Phobos and QSOs with various altitude ranges are defined in order to allow both global mapping and detailed analysis of landing site candidates (see paper by Nakamura et al. 2021).

Phobos
Phobos spectral observation by MIRS will consist of: • A global mapping, performed from High and Medium altitude (see Nakamura et al. 2021 for orbit definition), to determine Phobos composition. • A detailed observation of the landing site candidates, from medium and low altitude, to identify their thermal and compositional characteristics and provide the necessary information to select the 2 most interesting sites. • A close observation of the 2 landing sites, during the vertical descent phase, to have a detailed spectral map to contribute to the selection of the sampling spot.

Global mapping
The first 2 months of Phobos observation will be performed from high-and medium-altitude equatorial quasi-satellite orbits (QSO), at a time when Sun elevation on Phobos (β angle) is low, which is favorable for the observation of North and South hemispheres. Priority is given to the zone within ± 30° latitudes, but higher latitudes (poles observation) are also important to settle the question of Phobos origin. On the maps in Fig. 20, it is shown that high-resolution coverage of Phobos can be achieved by MIRS from medium equatorial quasi-satellite orbit. Coverage can be global from a 3D quasi-satellite orbit that allows visibility of high latitudes.

Landing sites candidates observation
Landing site candidates will be observed from medium (QSO-M) and low altitude equatorial quasi-satellite orbits (QSO-LA, -LB' , -LC' , Table 3). They will be in a flat zone for safe MMX landing and within latitudes ± 30°. Analysis of data collected from a given QSO will allow to select the remaining best candidates to observe on the next QSO. It is therefore very important to ensure that data volume and data downlink rate are compatible with the timely analysis and selection process. The lower the altitude is, the faster MMX relative speed to Phobos is. Therefore, for low-altitude QSOs MIRS along-track scanner will be used during observations to reduce the ground speed of MIRS Line of Sight projection on Phobos so that it is compatible with geometric pixel resolution and integration time.
In order to get as much information as possible for spectrographic and thermal inertia characterization, each site will be observed with several couples of phase angle and solar local time, in addition to the couple 10°/12 h.

Landing sites close observation during descent
During the 2 descent phases, MIRS will perform two 50-m large spectrographic maps, from altitudes 2 km down to 1 km, with a geometric resolution between 0.7 and 0.35 m. Additionally, to contribute to the selection of the sampling spot that will be chosen within a 2-m square zone centered on landing site, MIRS will perform a finer map from lowest possible altitude to get the best resolution.

Mars
In order to better understand the transport processes of dust and water between atmosphere and surface, large zones of Mars atmosphere will be observed several times with a time-resolution below 1 h. During most of its 3-year mission, MMX trajectory will be in quasi-satellite orbit around Phobos, which is very similar to Phobos orbit around Mars: nearly equatorial and circular, with a period of 7 h40' . Consequently, Mars observation slots will occur every 7 h40' at best and will last a maximum of 3 h50' (corresponding to the time MMX is above Sunlighted part of Mars).
Each observation of a zone will consist of a series of west-east stripes: MIRS along-track scanner will move the line of sight along west-east while MMX spacecraft maneuver will perform the cross-track offsets before the acquisition of each stripe. To optimize coverage, it is proposed that MMX compensates during observations the yaw offset of the stripes induced by the small inclination of MMX orbit.
In Fig. 21, we present 3 examples of observations: on the left, a zone on the south hemisphere is observed 3 times with a time resolution of 45 min. On the center, a central zone is observed 5 times with a time resolution of 30 min: the longer the total time span is, the shorter the coverable west-east span of the zone is. On the right, a single observation of most of Mars disk lighted zone is performed. To gather further information on Mars atmosphere composition, MIRS will also perform various kinds of limb observations.

Deimos
Deimos being around 2.5 times farther from Mars center than Phobos, it will be observed during multiple fly-bys, from the elliptical orbits of MMX around Mars (Fig. 22). Current baseline is that MMX will remain on each of the three foreseen orbits the necessary time to perform ten Deimos fly-bys. It is a real challenge for observations as MMX relative velocity to Deimos being very high, distance will evolve very quickly, reducing the duration of observation slots with a geometric resolution below 100 m. A key issue is the setting of the minimum distance to Deimos: the farther the safer for MMX trajectory, the shorter the better for geometric resolution.

Conclusions
MIRS is an imaging spectrometer that uses the pushbroom acquisition principle. The near-infrared wavelength range in 0.9-3.6 µm has been chosen to optimize the study the Phobos and Deimos surface composition as well Mars atmosphere with a spatial resolution of 20 m or better on Phobos surface. It will be also a fundamental instrument to evaluate sampling site candidates.
For Phobos, MIRS will allow to obtain a global mapping, performed from high and medium altitude. It will allow to investigate the surface composition as well the thermal characteristics and to provide, with observations during the closer orbits, the necessary information to select the most interesting sampling sites. MIRS will allow also to have a detailed spectral map to characterize the sampling spot with the close observation of the two landing sites, during the vertical descent phase down to a distance from the surface of at least 400 m. MIRS will allow to characterize the Deimos surface composition to investigate the compositional analogy with Phobos. MIRS will provide the first opportunity for investigating the Martian diurnal variation and the transport processes of dust and water vapor with timeresolved pictures of the atmospheric phenomena from the equatorial orbit.
The MIRS instrument together with the MMX payload and sample analysis will be able to clarify the origin of the Martian moons and may also be able to elucidate the process of the evolution of the Mars environment.