Infrared Spectrometer for ExoMars: A Mast-Mounted Instrument for the Rover

2.1. Contribution to overall rover mission science

The scientific objectives of the ExoMars program are defined as follows (Vago et al., 2017):

1. To search for signs of past and present life on Mars;

The first two objectives are the most relevant to the ExoMars rover. The ISEM experiment will contribute to achieving each of these objectives in the following ways:

1. Many of the minerals and rocks detectable by ISEM are good indicators of past habitable conditions and of biological processes [e.g., carbonates, oxalates, borates, nitrates, NH₄-bearing minerals (Applin et al., 2015; Berg et al., 2016; Cloutis et al., 2016)] and may even contain biomolecules in detectable concentrations. These include a class of biogenic minerals that are further known to provide a substrate for, and catalyze, prebiotic reactions. Organic compounds, including polycyclic aromatic hydrocarbons (PAHs) and those containing aliphatic C-H molecules, can also be distinguished by ISEM (e.g., Clark et al., 2009; Izawa et al., 2014).

In terms of where the rover will land, ESA has issued the following scientific criteria for landing site selection (Vago et al., 2015):

For the ExoMars rover to achieve results regarding the possible existence of biosignatures, the mission has to land in a scientifically appropriate setting:

1. The site must be ancient (older than 3.6 Ga)—from Mars' early, more life-friendly period: the Noachian to the Noachian/Hesperian boundary;

ISEM addresses these criteria in the following ways:

1. The suite of samples that we have used to characterize ISEM in laboratory tests includes materials similar to those found in Noachian terrains (e.g., phyllosilicates and carbonates).

2.2. Synergies with other instruments

The main goal of ISEM is to establish the mineral composition of Mars' surface materials remotely. ISEM, together with PanCam (Fig. 1), offers high potential for the remote identification and characterization of any scientifically high-value target in the vicinity of the rover, including proximal and distant rocks, outcrops, and other geological formations. ISEM will help to establish the geological context of each site along the rover traverse, discriminating between various classes of minerals and rocks. ISEM will also be important in the selection of promising sites for subsurface sampling.

OPEN IN VIEWER

FIG. 1.

Schematic view of the ExoMars rover mast instruments: PanCam, navigation cameras, and ISEM. ISEM, Infrared Spectrometer for ExoMars; PanCam, panoramic camera.

Infrared reflectance spectroscopy allows the study of the composition in the uppermost few millimeters of a rock's surface. It allows discriminating between various classes of silicates (hydr-) oxides, hydrated/hydroxylated salts, and carbonates. As shown in Figure 2, the 1.3° field of view (FOV) of ISEM lies within the 5° FOV of the color PanCam HRC, and they both are within a much wider FOV (38.6°) of the WACs with multispectral capabilities (Coates et al., 2017). The multispectral data are produced using a filter wheel with 11 filter positions for each of the two WACs. Out of the 22 filters, six are devoted to red, green, and blue broadband color and are duplicated in both cameras. Twelve are optimized for mineralogy in the 400–1000 nm range, and four “solar” filters are dedicated to atmospheric studies. By extending the wavelength range beyond PanCam, ISEM will enable many more spectral features diagnostic of specific mineralogy to be detected. Together, PanCam and ISEM provide spectrally resolved information from 0.4 to 3.3 μm.

OPEN IN VIEWER

FIG. 2.

Schematic representation of possible ISEM and PanCam joint observation scenario showing a sequence of ISEM measurements acquired together with WAC and HRC PanCam frames. HRC, high-resolution camera; WAC, wide-angle camera.

The identification and mapping of the distribution of aqueous alteration products in the upper surface layer, combined with subsurface data from the neutron detector ADRON and the ground-penetrating radar WISDOM, will help to understand the subsurface structure and the exobiology potential at each prospective drilling site.

The collected drill samples will be analyzed in the rover's Analytical Laboratory Drawer by several instruments. The first is an infrared hyperspectral microscope MicrOmega (Bibring et al., 2017). The principle of MicrOmega is very similar to that of ISEM. The reflectance spectroscopy is performed in the near-IR, but the analysis is done at the microscopic scale. The sample is illuminated by a monochromatic light source. In contrast to MicrOmega, ISEM has a much wider FOV and range of detection. The distance to a target is not significantly limited, and practically may reach hundreds of meters. Thus, ISEM is better suited for accommodation on the rover mast, where it can be used for target identification of far away objects and also for investigating outcrop, rock, and soil mineralogy at close range.

There is no specific instrument dedicated to environmental characterization on the rover. Although hampered by the limited number of observation cycles, ISEM, jointly with PanCam, will deliver information regarding atmospheric aerosol opacity and the atmospheric gaseous composition. The data on water vapor content and aerosol will be retrieved as a by-product of reflectance spectra, from in-flight calibration, from the direct Sun imaging by PanCam, and from sky observations by PanCam and ISEM.

2.3. The method

Near-IR spectroscopic observations of the Mars surface have not been performed from any surface platform to date. Spectroscopy was used on the Mars Exploration Rovers (MER) in the thermal IR range (the radiation emitted by the surface, 5–29 μm) with the Mini-TES instrument (Christensen et al., 2003, 2004a, b). These remote observations proved very useful for selecting targets for in situ analyses by the alpha proton X-ray spectrometer (APXS) and Mossbauer instruments on Spirit and Opportunity (e.g., Squyres et al., 2004). They also allowed the identification of carbonates in Gusev Crater (Morris et al., 2010) and undertook a number of atmospheric investigations (Smith et al., 2006).

Conversely, the near-IR spectral range (i.e., the infrared solar reflected radiation) is being widely used in orbital observations, as it allows for significantly better characterization of aqueous minerals than does the thermal infrared. To illustrate this point, the TES instrument on Mars Global Surveyor in the wavelength range from 6 to 50 μm has mapped the distribution of both mafic and anhydrous high-silica minerals (Bandfield et al., 2000; Christensen et al., 2001), but could not unambiguously detect clay/clay-like minerals and salts, including carbonates. In the near-IR, the OMEGA hyperspectral instrument on Mars Express (MEx) measured the martian surface reflectance from 0.5 to 5.2 μm and was able to recognize a number of different phyllosilicates and sulfates (Bibring et al., 2006). Building on OMEGA's success, the higher spatial resolution of the CRISM instrument on Mars Reconnaissance Orbiter (MRO), which measures reflectance from 0.362 to 3.92 μm, permitted the first detections of carbonates and serpentine (Ehlmann et al., 2008, 2010). One shortcoming of the orbital observations is the limited surface resolution (300–500 m per pixel for OMEGA/MEx and ∼20 m for CRISM/MRO), which hinders the detection of small-scale exposures. In addition, such instruments suffer from limited detection sensitivity owing to the combination of the subtle nature of aqueous mineral spectroscopic features and to the short dwell time of orbiting spacecraft. The wide variety of minerals on Mars detected by the higher spatial resolution CRISM instrument compared to the lower resolution OMEGA instrument demonstrates the greater mineralogical diversity that can be identified as spatial resolution improves. Therefore, the close-up near-IR capability of ISEM on the ExoMars rover offers a very high diagnostic potential.

The small size and low mass of the ISEM instrument allow it to be accommodated more easily on a rover platform. A similar AOTF-based near-IR spectrometer has been selected to enhance the capabilities of the successor to the ChemCam instrument on the Curiosity rover (Fouchet et al., 2015). The new SuperCam instrument planned for the NASA 2020 rover utilizes a combination of the laser-induced breakdown spectrometer, Raman, and near-IR reflectance spectroscopy, offering complementary elemental and mineralogical analysis techniques.

2.4. Potentially detectable mineral groups

The greatest share of what we know about martian mineralogy has been gleaned from orbital near- and thermal-infrared spectroscopy measurements. The infrared instruments have been successful in identifying igneous minerals, while the near-IR instruments were more important for detecting ancient hydrated minerals, such as phyllosilicates, as well as carbonates and sulfates, which could be used to develop a timeline of changing environmental conditions on Mars and evidence of previous more clement epochs on Mars. The capabilities of ISEM can be gauged against these detections to assess its expected science return, as well as expectations for detecting other mineral species of high scientific value, all in the context of the proposed ExoMars landing sites.

2.4.1. Phyllosilicates

The identification of hydrated phyllosilicates on the martian surface is one of the discoveries driving in situ exploration of Mars. Phyllosilicates are commonly detected in ancient, Noachian-aged martian terrains (Poulet et al., 2005). Various types of phyllosilicates have been observed, including Fe/Mg and Al-rich smectites, micas, vermiculites, kaolinite, chlorite, and serpentine (Ehlmann et al., 2009; Murchie et al., 2009; Carter et al., 2013). The presence of these minerals likely indicates that conditions in the past were favorable for the presence of liquid water at or near the surface. They are considered good reaction templates for organic molecules and excellent for biosignature preservation (Bishop et al., 2013, and references therein). Phyllosilicates have been detected from orbit within the ellipse of two of the three candidate landing sites for the ExoMars rover, Oxia Planum, and Mawrth Vallis (Vago et al., 2017). They have not been detected in Aram Dorsum yet; even if they are there, it may be difficult to see them from orbit. ISEM will have the capabilities to detect sharp features between 1.35 and 2.6 μm in phyllosilicate reflectance spectra, which are related to combinations and overtones of OH-M stretching and bending bound to various cations (M) as well as the H-O-H stretch and bend from water bound in interlayer regions or adsorbed on mineral surfaces (Clark et al., 1990) (Fig. 3a–c).

OPEN IN VIEWER

FIG. 3.

Example spectra of minerals and rocks in the ISEM spectral range, which may be expected at the ExoMars landing sites (from CRISM spectral library, http://pds-geosciences.wustl.edu/missions/mro/spectral_library.htm; sample labels from this library). The upper half of the figure (a–f) shows the spectra of minerals the most significant for habitability, while the spectra shown in the lower half (g–l) are more relevant to background minerals and dust. The spectral resolution of CRISM is close to that of ISEM in the center of the range; ISEM resolution is better at shorter wavelengths and coarser at 3 μm (Table 1).

2.4.4. Silica

Silica, in a variety of forms, is present in a number of terrains on Mars (Bandfield et al., 2004; Smith and Bandfield, 2012; Bandfield et al., 2013; Carter et al., 2013; Smith et al., 2013). Differences in reflectance spectra can be used to distinguish different forms of silica, some of which are associated with habitability (Rice et al., 2013). The region most useful for distinguishing different forms of silica is located near 2.2 μm and is due to Si-OH overtones. The nature and abundance of these bonds vary among different types of silica (Rice et al., 2013). This diagnostic feature can be clearly seen in the reflectance spectra of two different grain sizes of quartz (Fig. 3g).

2.5. Atmospheric studies (aerosol, gaseous content)

There is no dedicated instrument, such as a meteorological station on the ExoMars rover, to characterize the environmental conditions. Therefore, any information available by other means is of particular value. Spectrally resolved data in the visible-NIR and in the thermal IR ranges have been used to assess the atmospheric state from Mars Pathfinder and MERs. In the spectral range of ISEM, there are absorption bands of atmospheric CO₂ (at 1.43, 1.6, 2.0, and 2.7 μm) and H₂O (at 1.38 and 2.56 μm), and it is well suited, in particular its short-wave subrange, to characterize the main mode of the martian dust (Fedorova et al., 2014). The condensation clouds in the equatorial region consist predominantly of water ice (Vincendon et al., 2011), with multiple absorption bands within the ISEM spectral range (Section 2.4). Carbon dioxide clouds, if present at high altitudes, may be detected by distinct features of solid CO₂ near 1.2, 1.4, 1.5, 2.0, 2.7, and 3.0 μm.

The gaseous absorptions will be measured by ISEM as a by-product of every surface measurement. Depth of the CO₂ absorptions can be used to determine the surface pressure, and the H₂O absorptions quantify the total column water contents in the atmosphere above the site. To disentangle the possible overlapping atmospheric and mineral features in the reflected surface spectra, preferentially the observations of the in situ calibration target will be interpreted. The estimate of accuracy to retrieve the atmospheric pressure by ISEM is given by the OMEGA study of the 2 μm CO₂ absorption from the orbit: 7–10 Pa (∼1%) 1-sigma (Forget et al., 2007). With the optical path being half of the OMEGA case, our retrieval accuracy will be comparable.

More precise measurements of surface pressure, minor gases, and better characterization of the atmospheric dust and cloud situation by ISEM will be possible using dedicated atmospheric observations. They will be coordinated with those by PanCam. The PanCam atmospheric study (Coates et al., 2017) will include direct solar observations just before sunset (to take advantage of the maximal atmospheric path lengths) and cross-sky brightness measurements. Such observations, for example, performed with Mars Pathfinder camera (Markiewicz et al., 1999; Titov et al., 1999) or MERs (Lemmon et al., 2004; Smith et al., 2006) can be significantly strengthened by extending the spectral range to the near-IR.

Direct solar observations by ISEM are not so far in the baseline. Their implementation would allow for high-accuracy pressure and water retrievals, better optical characterization of dust, direct and localized detection of H₂O ice and, if present, the CO₂ ice clouds. Observing the Sun at different zenith angles helps putting constraints on the vertical distribution of water vapor. However, ISEM is optimized for weak reflected light, and measuring the direct Sun signal may not be feasible. We successfully tested this possibility with the lunar prototype, but the ISEM aperture is larger, and a solar-blind filter within the foreoptics might become necessary. Such a filter would cause a several percent signal loss for the baseline ISEM observations. A final assessment of the direct Sun mode will be done during the characterization of the flight model.

The cross-sky brightness measurements by ISEM are the best suited for characterizing the aerosol component of the Mars atmosphere. Because of quasipermanent aerosol loading, the brightness of the martian sky at low zenith angles is not significantly lower than that from light scattered from the surface. Measurements at a range of phase angles will allow the science team to extract the size distribution and optical properties, and even assess the shape of aerosols suspended in the atmosphere. Water vapor absorption at 2.56 μm is much stronger than that at 0.94 μm to be observed by WAC in solar filters, and the ISEM data might put additional constraints on the vertical distribution of water in the boundary layer of atmosphere.

The variety of atmospheric measurements by ISEM calibrated by using an in situ calibration target will allow the refinement of atmospheric scattering models and therefore refinement of the calibration of orbital spectral measurements of the martian surface.

3.1. Instrument concept

The measurement principle of ISEM is based on the use of an AOTF. The core element of an AOTF is a birefringent crystal (typically of paratellurite, TeO₂, due to the combination of acoustic and optical properties) with a welded piezotransducer. The radio frequency (RF) applied to the transducer generates an acoustic field in the crystal, implementing acousto-optic interactions in Bragg's regime. The spectral selectivity of the acousto-optic diffraction allows the filtering of light. The diffraction occurs for a single wavelength, and there are no diffraction orders. The applied RF controls the tuning of the AOTF. AOTFs are technologically mature and widely used for spectral analysis. The robust design, small dimensions, and mass, coupled to the absence of moving parts in an AOTF-based spectrometer, make them popular for space applications. So far, AOTF-based spectrometers have been used in space science in the following ways: (1) to study the atmospheric composition of Mars and Venus (Korablev et al., 2006, 2012); (2) on the Moon within the Chang'e-3 VNIS spectrometer (0.45–2.4 μm) mounted on the Yutu rover (He et al., 2014); (3) for isolation of Echelle-spectrometer diffraction orders in high-resolution instruments (Nevejans et al., 2006; Korablev et al., 2011, 2014; Neefs et al., 2015); and (4) to illuminate the sample of an IR microscope with monochromatic light (Pilorget and Bibring, 2013; Bibring et al., 2017).

Wider application of the AOTFs in remote sensing is hampered by the inherent requirement to sequentially scan the spectrum. On an orbital mission with a short dwell time, this scanning interferes with the spacecraft or line of sight motion, complicating the analysis. Even for a pencil-beam device, different parts of the acquired spectrum would correspond to different observed areas. Conversely, observations from a static point, such as a planetary lander, or a rover, which remains immobile during the measurement, are well suited for AOTF-based instruments.

As described in Section 1, ISEM is a close equivalent of LIS, which is being developed for two Russian lunar landers. The LIS development is more advanced with respect to that of ISEM—by about 2 years. LIS benefited from the experience gained from the pencil-beam spectrometer design of the Mars Express and Venus Express instruments, and from the early developments of MicrOmega. The SPICAM-IR AOTF spectrometer with the spectral range of 0.9–1.7 μm (Korablev et al., 2006) has been operating in Mars orbit since 2004. A similar SPICAV-IR instrument using a double-range AOTF operated in Venus orbit from 2006 to 2014 (Korablev et al., 2012). A double-range AOTF for the spectral range of 0.7–4.1 μm serving as a prototype to the LIS and ISEM AOTFs was developed for the Phobos-Grunt mission (Leroi et al., 2009).

To reduce the influence of the extreme temperature conditions at the rover's mast on the electronics, the instrument is implemented as two separate boxes, a mast-mounted optical box (OB) and the electronics box (EB) (Fig. 4). The thermally stabilized EB is mounted within the rover's body. The parameters of ISEM are summarized in Table 2.

OPEN IN VIEWER

FIG. 4.

The ISEM instrument, the photographs of the optical (a) and electronics box (b) structural and thermal models.

3.2. The optical box

Although most of the electronics are in the EB, some electronics such as the detector's preamplifier and the RF conditioning electronics remain in the OB. Neither the weak photodiode current nor the power RF can be transmitted via the 5-m harness. The block diagram of ISEM is presented in Figure 5.

OPEN IN VIEWER

FIG. 5.

The block diagram of the ISEM control electronics.

The OB contains all the optical elements, the AOTF with associated electronics (ultrasound frequency synthesizer and amplifier boards), the photo detector, and the photo detector board (Fig. 6).

OPEN IN VIEWER

FIG. 6.

The three-dimensional model open view of the ISEM optical box. The main elements are visible: 1-baffle, 2-entry optics; 3, 7-AOTF collimating optics; 4, 6-polarizers; 5-AOTF crystal; 8-detector; 9-detector's preamplifier; 10-AOTF RF proximity electronics. AOTF, acousto-optic tunable filter; RF, radio frequency.

The spectrometer is built following a standard layout for an AOTF spectrometer, with the AOTF in the path of a quasiparallel beam. The optical scheme is based on a Galileo system with remote pupil built using CaF₂ and ZnSe lenses, and it is presented in Figure 7. The image is transferred through the optical system, and the FOV of 1.3° is formed on the detector-sensitive area. The achromatic lens entry telescope (1) has the aperture of 25 mm. The AOTF crystal (5) is placed in a quasiparallel beam between collimating lenses (3, 7) and a pair of polarizers (4, 6); the output collimating lens (8) serves also as a focusing optic for the detector (8).

OPEN IN VIEWER

FIG. 7.

ISEM optical scheme. The numbering is the same as in Figure 6: 2-foreoptics; 3, 7-AOTF collimating optics; 4, 6-polarizers; 5-AOTF crystal; 8-sensitive area of the detector.

A wide-angle AOTF from NII Micropribor in Zelenograd, Russia, is manufactured on the base of a tellurium dioxide crystal. The ultrasound frequency range of 23–82 MHz provides a spectral range from 1.15 to 3.3 μm with two piezotransducers. The transducers operate in the subbands of 23–42 and 42–82 MHz. The crystal cutoff angle is 12.5° in the (110) crystallographic plane. The anisotropic Bragg diffraction regime is used. The incident optical radiation has ordinary polarization, and the diffracted optical beam has the extraordinary polarization. The angle between the passed and diffracted optical beams is 6° at the output of the AO crystal. A pair of polarizers with crossed polarizing planes is used to filter out the nondesired zero diffraction order.

The AOTF and its electronics are assembled in a single functional unit, which includes the acousto-optic cell, the polarizers, a proximity RF matching board, an RF synthesizer, the driver of the AO crystal (the RF power amplifier), and a dedicated internal microcontroller, which communicates with the main controller (MC) in the EB and controls all the functions of the acousto-optic module (AOM). Commands define magnitude and frequency of the RF signal applied, as well as the RF driver ON/OFF states. Within the RF range of 23–82 MHz, the minimum step of frequency sweeping is 10 kHz (5900 frequency points). The RF amplitude may be set at 1 of 16 even levels. The AOM transmits back to the MC a few housekeeping parameters, such as measured RF voltage and AO crystal temperature. The MC and the AOM are connected via an RS-485 interface running at a speed of 115.2 Kbit/s.

During the measurement, the RF level is being alternated between ON and OFF states, with the cadence being defined by integration time. The measured signal is then processed as the AC allowing to remove offsets caused by the detector's dark current and stray light, and improving the dynamic range of the instrument.

The detector is a single-pixel InAs thermoelectrically cooled photodiode. A detector module J12TE3-66D-R01M from Teledyne Judson Technologies is used. The built-in three-stage thermoelectric Peltier cooler maintains a detector temperature of about 90°C below that of the hot side. For the ISEM OB operating in the range from −10° to +30°C, it results in the detector's temperature ranging between −100° and −60°C, the corresponding detector's shunt resistance is therefore 60–400 kOhms. The temperature of the sensitive area is monitored by a built-in thermistor.

The detector's photocurrent is amplified with a two-stage circuit. The first stage is a transimpedance amplifier, AC coupled (1-s time constant) to the second stage. The second stage has a gain of 83 and a time constant of 80 μs and is based on an ADA4610 (Analog Devices) operational amplifier characterized by low current and voltage noise (50 fA Hz^−1/2, 7.5 nV Hz^−1/2). Its output signal is transmitted via the harness to the electronics box.

3.3. The electronics box

The EB is mounted inside the rover at the rear balcony, and it is thermally stabilized. It includes the ADC, the main and auxiliary controllers, power conditioning (power supply unit, PSU), and the interface and bridge boards, which support RS-422 communication between the ISEM and the rover data and command system.

The two ISEM blocks are connected via a 5-m harness running from the inside of the rover to the mast. It is fabricated by Thales Alenia Space-Italy (TAS-I). The cable includes the analog signals from the detector and the thermal sensor, the power supply lines, and a digital RS-485 connection to control the AOFT RF synthesizer and the power amplifier.

The MC located in the EB commands the operation of all modules of the instrument. It uses an MSP430FR5739 Texas Instruments circuit running at 19.68 MHz. The controller is equipped with 1 kB of RAM and 16 kB of ferroelectric memory (FRAM) used for program and data storage. Compared to commonly used FLASH memory, the FRAM has better radiation immunity. A 10-bit internal ADC of the MC is used to digitize the signal from the detector's thermistor.

The main ADC serves to digitize the signal from the detector, amplified in the OB. A 16-bit ADC (ADS8320; Texas Instruments) is used. The full-scale range of the ADC is 2.9 V and the peak-to-peak noise is 3 LSB (22 μV RMS). The MC receives the ADC output data via a serial interface operating at 2 Mbit/s. The preamplifier and the ADC contribute little to the total noise in the signal path, and thus, the Johnson noise of the detector controls the limit of signal detection.

The auxiliary controller, based on C8051F121 Silicon Laboratories circuit operating at 12.25 MHz, incorporates a 12-bit ADC and serves to monitor power supply voltages and dispatch the operation of different memory devices. In all, ISEM uses two built-in FLASH memories within the microcontroller and two external chips of FRAM and magneto-resistive memory (MRAM).

The two microcontrollers hold four identical copies of their firmware in program memories. At the start of operation, the controllers check the copies, repair damaged ones, and run a validation program. To increase its reliability, ISEM contains two redundant MC units. After turning on, first the MC1 is powered. Its sequence, if performed correctly, commands the PSU to keep power at MC1. If not commanded in 5 s, the PSU automatically powers on the MC2, and so on.

The rover onboard computer (OBC) controls ISEM operations via RS-422 interface. There are two (nominal and redundant) RS-422 links. Communication via only one of the links is available at a given time. The baud rate is 112.179% ± 1% Kbit/s.

3.4. The calibration target

To determine the incident solar illumination spectrum for deriving I/F data and to verify the in-flight performance and stability of the instrument, ISEM will observe the radiometric calibration target before each measurement. The target is used for the in-flight radiometric calibration of both PanCam and ISEM and will be located on the front deck of the rover as shown in Figure 8. At this location, it will be interrogated by ISEM from a distance of 1.1 m at an emittance angle of 24°. The calibration target occupies an area of 67 × 76 mm² and has a mass of 40 g.

OPEN IN VIEWER

FIG. 8.

PanCam and ISEM calibration target and its location on the rover. Larger circles (Ø30 mm) are for both ISEM and PanCam. The smaller circles (Ø18 mm) and the shadow posts will be used by PanCam only.

The target includes eight stained glass diffuse reflectance calibration patches with different spectral reflectance properties. Two of these calibration patches will be used by ISEM—the “white,” which has a reflectance near 100% in the 0.4–3.0 μm spectral range, and the multiband patch, which has distinct spectral features. The white patch is manufactured from Pyroceram provided by the Vavilov State Optical Institute in St. Petersburg, and the multiband patch is manufactured from WCT-2065, a rare earth doped glass manufactured by Schott and supplied by Avian Technologies in the United States. The calibration patches will be calibrated for absolute total hemispherical reflectance and bidirectional reflectance distribution function, and all measurements will be traceable to photometric standards.

Dust deposition on the radiometric calibration target during the ExoMars mission will be accounted for in the data processing by developing a model of the calibration target and dust system, which will build on the results of previous missions, measurements of settling rates on the rover panels, solar arrays, and so on (Kinch et al., 2007, 2015), and from PanCam calibration results.

4.1. The experiment cycle

The experiment cycle consists of one spectrum measurement by ISEM. It is explained in Figure 9.

OPEN IN VIEWER

FIG. 9.

A diagram showing timing and breakdown of one spectrum measurement.

To eliminate the dark signal of the detector and any possible stray light signal, the instrument measures photocurrents when the AOTF RF is OFF (the dark signal) and when it is ON (the full signal). A subtraction of the two values gives the true signal corresponding to the spectrally filtered radiation on the detector.

During both the ON and OFF states, the main ADC continuously samples signals at a 20 kHz rate. The MC stacks all the measured values and makes the ON–OFF subtraction. This elementary measurement can be repeated up to 256 times (“accumulation factor” parameter), effectively increasing the integration time. All resulting values are stacked as well. Each elementary measurement includes a dead time of ∼0.6 ms needed to turn the AOTF ON or OFF. These overheads are mostly the result of remote commanding of the AOM in the OB, but also include two 0.12 ms transients in the AOTF. An additional dead time of 0.45 ms is needed to change the RF. The integration time, variable between 1 and 256 ms, and defining the ON/OFF timing, and the number of repeated ON/OFF cycles for a given RF value (the accumulation factor), are set by a command.

As discussed above, the measurement time is a critical factor. To achieve flexibility in spectral sampling, up to eight “windows” can be defined for one measurement of a spectrum. Every window can be placed anywhere in the AOM frequency range, and it has its own number of spectrum points, frequency steps, and accumulation factors.

During an elementary measurement, the 16-bit ADC readings add up to a 24-bit value, and further accumulation, up to 256 times, results in a 32-bit value for each spectrum point. Only 16 bits out of it are used such that three options are available as follows: (1) transfer merely the 16 LSBs, (2) transfer 16 bits starting from a certain bit to remove noise, or (3) normalize within one spectrum so that the maximum fits exactly within 16 bits. In the latter case, the risk is that an errant spike distorts the entire spectrum.

The exchange of commands between ISEM and the OBC proceeds as follows: The OBC transmits a command (TC) and rover elapsed time (RET). The TC defines ISEM operation parameters and triggers the spectrum measurement. The RET command synchronizes ISEM internal time. In the course of operation, the ISEM sends scientific and housekeeping packets to the OBC, and these data are stored into the OBC memory.

4.2. Operations on the surface

ISEM operations over the course of the Reference Surface Mission are intimately connected to those of the mast and PanCam. All operations are to be guided by the mast's pointing. Three basic operating regimes have been identified, which include the following:

(1) Operations to support sampling

ISEM will work in parallel with PanCam, as shown in Figure 2. First, the WAC observes an area panoramically, and then, the HRC is pointed to the most interesting spots. The spot must be illuminated by direct sunlight. At least one ISEM spectrum is planned to be acquired together with every HRC image, and ISEM measurements may also be conducted separately. The ISEM data will be used to characterize the mineral composition. Sometimes ISEM will be used together with the WAC “geology” filters; however, full analysis with the PanCam WAC geological filters is slower and produces much more data, so this capability cannot be used regularly.

(2) Operations to assess the geological value of targets

A number of geologically interesting scenes may be distant from the planned rover track or inaccessible for rover sampling, such as crater rims and outcrops. Remote characterization is the only possibility in this case. ISEM will be operated in the same manner as in the first case. There is no theoretical limit on the sounding range provided the illumination conditions (including the phase angle) are acceptable. For instance, illuminated hill slopes may be observed from dozens of km. This mode could be used for rover orbiter coordinated measurements and validations. For a favorable scene, ISEM may provide cross-validation over the area covering multiple CRISM pixels. A practical limit for scene observations is determined by the ISEM FOV, an ∼1-m spot is observed from a 44-m distance.

(3) Dedicated environmental operations

ISEM data may be used to characterize atmospheric humidity and aerosol loading. Some information will be obtained as a by-product of geological measurements intermitted with observations of the calibration target. The information about the gaseous atmospheric absorption will be assured by dense spectral sampling of the appropriate spectral intervals. Aerosols are best characterized by cross-sky observations. Plans for the PanCam WAC include direct Sun observations and cross-sky viewing atmospheric campaigns. As discussed in Section 2.5, the ISEM capability to observe the direct Sun is not confirmed. At the same time, following preliminary laboratory characterization, we do not expect any operational constraints placed by ISEM on PanCam solar imaging.

In the case of cross-sky imaging, the ISEM line of sight will be oriented above the horizon, and a special ISEM data acquisition sequence will be implemented. The HRC measurements are not needed, and ISEM will observe in parallel with WAC. The elevation and phase angle for these measurements need to be carefully planned, targeting as much as possible a range of phase angles for WAC and detectable scattered signal in the IR for ISEM.

4.3. Resources required

ISEM data volume is relatively small. One spectrum is only 2 kB in size (Table 2 and Section 3). The factor most affecting the rover operations is the measurement time. Fixed surface platforms are better suited for lengthy operations, generally allowing significant accumulation of measurements. However, sequential acquisition of spectra effectively increases the recording time, and as a result, required measurement times become an important factor for the rover operation cycle.

The allocated duration of measurements is from 2.5 min (minimum case) to 8.0 min (optimal case) for a single target. During the whole sequence, ISEM will consume ∼12 W, requiring an energy of 0.5–4.4 W per hour. As described in Section 3.4, the length of the spectrum, or the number of points in the spectrum, can be selected. For a “geology” target, we will always measure across the full spectral range of ISEM, which may be divided into 1–8 intervals with different sampling. The total number of points in one spectrum will vary between 256 and 1024. Within the spectral range, some portions of the spectrum will be measured more accurately, where, for example, a narrow mineral absorption feature or important atmospheric absorption is expected, while other portions of smooth continuous character can be undersampled.

4.4. Measurement performance, examples, comparison with state of the art

As the full working prototype of ISEM has not been built yet, the performance of the instrument was verified by using the closest prototype, the qualification model of LIS for the Luna 25 lander. The ISEM hardware involves some improvements with respect to LIS, including better optical throughput (and a more sensitive detector featuring the three-stage Peltier cooler). The spectral properties of the AOTF remain the same.

4.4.1. Spectral range and spectral resolution

The bounds of the spectral range were verified by using He-Ne 1.152 and 3.39 μm laser lines. The theoretical shape of an AOTF spectral instrument function is close to sinc² x. Side lobes of this function modify the measured spectrum, but this could be readily accounted for once the bandpass is properly characterized. More dangerous are distant side lobes, which manifest themselves as a sort of stray light and may reduce the apparent depth of spectral features. More discussion on this aspect of AOTF characterization may be found in the work of Korablev et al. (2013). The measured bandpass transmission function of the ISEM prototype is shown in Figure 10a. The evolution of the resolving power λ/Δλ through the spectral range measured at multiple wavelengths when using light sources with sharp emission features (different He-Ne laser, single-mode laser diodes, and low-pressure gas-filled lamps) is shown in Figure 10b.

OPEN IN VIEWER

FIG. 10.

(a) Normalized transmission function of the instrument measured at 1.523 μm laser line. The FWHM is 19 cm⁻¹ (2.7 nm). (b) The resolving power measures through the full spectral range. FWHM, full width at half maximum.

A simulation measurement of a mineral spectrum of a Mars analogue by LIS prototype is presented in Figure 11. Crystalline gypsum was manually ground and sieved to obtain ∼1-mm grains, and the resulting sample was illuminated with a 150-W halogen lamp from a distance of 18 cm. The full time to measure this spectrum was 30 min, and the effective exposure to measure one spectral point was 1 s in the long-wave AOTF subrange and 0.6 s in the short-wave subrange. The spectrum shown is normalized by a spectrum of a Lambertian screen (metal surface covered by special reflective coating). The differences between the library and measured spectra might be attributed to different origins of the samples; the difference between the laboratory Fourier-transform spectrometer (FTS) and ISEM spectra is likely due to different observation geometries.

OPEN IN VIEWER

FIG. 11.

A reflectance spectrum of crystalline gypsum measured by the closest equivalent of ISEM: LIS qualification prototype and a laboratory FTS spectrometer (Bruker), compared to a CRISM library spectrum (LASF41A). FTS, Fourier-transform spectrometer; LIS, lunar infrared spectrometer.

4.4.3. Estimation of detection capabilities

Starting from the S/N estimation above, we may estimate minimum detectable abundance for minerals discussed in Section 2.4 (Fig. 1). Precaution factors in the measurement time (2–3) and in sensitivity (∼3) allow sufficient margin to account for, for example, an unfavorable phase angle or other factors. We used spectra of different minerals of interest (Fig. 1a–f) on the background spectra of “bright” and “dark” regions of Mars (Erard, 2001) to simulate the reflectance of an admixture to the bulk spectrum of the martian surface. The preliminary sensitivity estimates to detect the fraction of the selected mineral are presented in Table 3. Given the relatively high ISEM S/N ratio deduced above, these estimates based on the relative depth of individual spectral features are very conservative. A better sensitivity analysis is planned in the future, involving analysis of martian mineral analogues and multicomponent analysis.

OPEN IN VIEWER

Table 3.

Preliminary Estimation of the Infrared Spectrometer for ExoMars Sensitivity to Detect Different Minerals and Mineral Groups From CRISM Spectral Library (Fig. 1) on the Background of Bright and Dark Region Mars Spectra (Erard, 2001)

Mineral or mineral group	Bright region, %	Dark region, %	Comment
Kaolinite (Fig. 1a)	<5	5–10	1.4, 2.2, 2.4 μm
Montmorillonite (Fig. 1b)	5–10	10–15	1.4, 1.9, 2.2–2.3 μm
Serpentine (Fig. 1c)	5–10	5–10	1.4, 2, 2.35 μm
Carbonates (Fig. 1d)	5–10	∼10	2.3, 2.5, 3.4 μm
Gypsum (Fig. 1e)	∼5	5–10	1.4–1.6, 1.8, 2.2 μm
Kieserite (Fig. 1e)	<5	∼10	Broad structure 1.4–2.4 μm
Jarosite (Fig. 1f)	<5	5–10	1.48, 1.75, 2.25 etc. μm

Mineral features used to estimate the sensitivity are listed in the last column.

The issue of detection capabilities can be divided into two broad categories: (1) characterization of a target in the presence of windblown dust and (2) detection of a phase or phases of interest in multicomponent assemblages. There are a number of studies that have examined detection limits for various components that are relevant to this issue, and these are supplemented by ongoing studies by our team that are designed to specifically address this issue. The results of these studies are summarized in Table 4. Some selected examples are shown in Figures 12 –14.

OPEN IN VIEWER

FIG. 12.

Reflectance spectra of palagonitic dust sprinkled on powdered carbonate (magnesite). The most diagnostic absorption band, near 2.33 μm, is indicated by an arrow. Note that the band is clearly present with a 260 μm palagonitic dust coating. It is still present, but barely discernible for a 525-μm-thick dust coating.

OPEN IN VIEWER

FIG. 13.

Reflectance spectra of intimate mixtures of nontronite (an Fe-rich clay) and adenine (a component of DNA). The brown arrows indicate spectral regions where adenine absorption bands are still detectable when only 1 wt.% of adenine is present in the mixtures.

OPEN IN VIEWER

FIG. 14.

Reflectance spectra of intimate mixtures of hematite (an Fe-oxide) and adenine (a component of DNA). The brown arrows indicate spectral regions where adenine absorption bands are still detectable when only 1.3 wt.% of adenine is present in the mixtures.

OPEN IN VIEWER

Table 4.

Summary of Some Studies Concerning Detection Limits and Discrimination of Different Phases Relevant to the Capabilities of Infrared Spectrometer for ExoMars

Mixture type	Detection limit	Characteristic spectral features	Source
Orthopyroxene+clinopyroxene	∼15% of either phase	2 μm region	Cloutis and Gaffey (1991)
Olivine+orthopyroxene	∼20% olivine, 5% pyroxene	2 μm region	Cloutis et al. (1986)
Pyroxene+palagonitic dust (spectral equivalent of Mars dust) or hematite	∼10% pyroxene	2 μm region	Cloutis and Bell (2004)
Carbonate+basalt	∼20% carbonate	2.3–2.6 μm region	Palomba et al. (2009) and this study
Carbonate+palagonitic dust (Fig. 12)	∼15% carbonate	2.3–2.6 μm region	Palomba et al. (2009) and this study
Palagonitic dust on carbonate (Fig. 12)	<∼500-μm-thick palagonite coating required	2.3, 2.5 μm regions	This study
Palagonitic dust on basalt	<∼250-μm-thick palagonite coating required		This study
Kaolinite+illite	∼20% of either phase	2.2–2.5 μm region	This study
Gypsum+basalt	∼5% gypsum	1.4, 1.9 μm regions	This study
Gypsum+palagonitic dust	∼5% gypsum	1.4, 1.9 μm regions	This study
Pyroxene+volcanic glass	∼15% pyroxene	2 μm region	This study
Palagonitic dust+carbonate	∼20% carbonate	2.3, 2.5 μm regions	This study
Nontronite+adenine (Fig. 13)	∼1% adenine	1.65, 2.2 μm regions	This study
Hematite+adenine (Fig. 14)	∼1% adenine	1.65, 2.2 μm regions	This study

Detection limits depend on a wide variety of factors, beside end member abundances. These include factors such as relative grain size, the availability of diagnostic absorption bands, the relative intensity of diagnostic absorption bands, and physical properties such as whether phases of interest are physically mixed or whether an obscuring phase, such as dust cover, if present.

We have found that, at least for the variety of mixtures examined, detection limits are on the order of 10–25 wt.%. For some phases, such as certain organic compounds, detection limits are much lower (e.g., 1 wt.% adenine is detectable in the presence of 99 wt.% nontronite (an iron-bearing clay) or hematite (an iron oxide) (Figs. 13 and 14)). Further extensive tests of mineral mixtures and detection limits are ongoing and planned.

4.5. Environmental requirements and characterization

The placement of the ISEM optical head on the rover mast with no possibility of thermal control imposes stringent requirements on survival temperature. The AOTF is a critical device involving bonding of highly anisotropic materials, such as TeO₂ and LiNbO₃. Anisotropic thermal expansion makes the AOTF devices vulnerable to large temperature excursion, in particular at low temperatures. AOTF technology has demonstrated survival and even operation down to −130°C and below (Leroi et al., 2009; Mantsevich et al., 2015). Several nonoperational thermal cycles attaining −130°C lower bound were also performed on two dedicated ISEM acousto-optic components; more testing to fully validate the technology is to be completed.

An important aspect of the AOTF spectrometer characterization is the thermal calibration, because the optical heads are intended for operations across a relatively wide temperature range. Some results of thermal characterization of the AOM were described by Mantsevich et al. (2015). The operation of the AO filters was tested in the range of −50° to +40°C. The temperature affects the elastic properties of TeO₂, changing the ultrasound velocity and birefringence. The variation of the TeO₂ refraction coefficient as a function of temperature is negligible. In turn, the change of the slow acoustic wave velocity causes a noticeable shift of the dispersion curve, comparable with the AO filter pass band in the operating temperature range. The characterization of the flight instrument will therefore include calibration of the dispersion curve within the operational temperature range.