2.1.

Observing the Earth’s Atmosphere to Inform Exoplanet Studies

Direct imaging spectroscopy of Earth-sized exoplanets, a powerful tool with the ability to transform our view of our place in the cosmos, is at the heart of NASA studies for missions at the flagship and probe levels. These missions can fulfill their potential only if their optical and spectroscopic designs—and eventual observations—are grounded in comprehensive, analogous data sets. Specifically, we need spectral observations of Earth as a whole disk, observed over the course of several of the planet’s rotations and taken at different phase angles. Although prior efforts have obtained parts of this data set,^2–12 the necessary combination of observations has not been collected. EarthShine will fill this gap and allow us to better assess the ability of future telescopes to determine specific planetary properties, including rotation rate, presence/abundance of oceans and continents, degree of cloud cover, and presence of vegetation. The superior assessment EarthShine enables can be incorporated into technology development investments and architecture trades of the next flagship missions.

The combined 0.4 to $12.5\mu \mathrm{m}$ range of EarthShine’s instrumentation spans the wavelength range under consideration for remote detection and characterization of terrestrial exoplanets. The combined measurements will enable the testing of all planned methodologies and the direct comparison of their sensitivity to geological, atmospheric, and biogenic features. Astrobiology combines elements of all these drivers and how they impact habitability and life. Observations of Earth have already informed exoplanet models and future exoplanet observatories.^8,10,13–15 Exoplanets, meanwhile, allow us to study types of planets not found in our solar system and to derive planetary statistics and processes on a comprehensive scale.

EarthShine’s wavelength range also enables assessing a number of key biosignatures and habitability indicators in Earth’s spectrum.^16,17 Table 1 shows important gaseous species and surface reflectance features accessible to EarthShine, some of which can only be measured from space. EarthShine will be able to measure several habitability indicators: water (${\mathrm{H}}_2\mathrm{O}$), carbon dioxide (${\mathrm{CO}}_2$), and glints from ocean and cloud ice crystals (Fig. 3);¹⁸ as well as biosignatures (oxygen (${\mathrm{O}}_2$), ozone (${\mathrm{O}}_3$), methane (${\mathrm{CH}}_4$), and the vegetation red edge.

Table 1

EarthShine measures spectral features of astrobiological significance. Note some species overlap spectrally.

Fig. 3

Sun-glint as a detector of oceans and ice clouds. The specular reflection (glint) from the ocean and from ice crystals in cirrus clouds seen in this DSCOVR/EPIC image will contribute a greater fraction of visible light in crescent phase, seen by EarthShine. EarthShine imaging will test glint as a water indicator.

2.2.

Observing the Global Earth to Inform Earth Climate Studies

The planet’s global radiation budget is critical to interpret both exoplanet atmospheres and Earth’s climate, and thus it is remarkable that we do not yet have simultaneous globally distributed measurements of Earth’s radiated energy. EarthShine measurements will contribute to important Earth Science goals by measuring Earth’s radiation budget, accounting for energy exchanged between star (Sun), planet, and space.

The lunar surface offers unique observations not possible from LEO, GEO, or L1 satellites. Radiation budget measurements from the CERES instruments onboard—e.g., NOAA-20, Terra, Aqua, and Suomi National Polar-orbiting Partnership—observe only part of the Earth at specific times; satellites in geosynchronous orbits observe only parts of the Earth and never see the polar regions; GEO satellites ($R/r\sim 6$), where $R$ is the orbital distance from the center of the Earth, and $r$ is the radius of the Earth, capture phase angle variation but are insensitive to rotational modulation. EarthShine’s comparatively distant vantage on the Moon’s surface ($R/r\sim 60$) will enable both of these important components of characterizing unresolved exoplanet climate and surface conditions. The EPIC and NISTAR cameras on the DSCOVER spacecraft at L1 provide only specific, nearly back scattering “full-Earth” direction phase angles. To draw a direct link between L1 observations and Earth’s radiation budget, we must observe many scattering angles, not just back scattering. EarthShine observations will provide the missing phase angles. EarthShine will see the whole Earth—including its sunlit and shadowed parts: day, night, and twilight zone—at different phase angles, especially during crescent Earth phases. This complete viewing will help validate Earth (broadband and narrow band) albedo and provide comprehensive (both time and space, compared to LEO) whole-globe monitoring of volcanic and aerosol clouds (smoke and dust), including polar regions not covered by GEO. Further, EarthShine will uniquely inform:

Fig. 4

DSCOVER/EPIC-based wavelength dependence of enhancement caused by specular reflection over land (red) and ocean (blue). Note that 764 and 688 nm wavelengths are EPIC oxygen absorption A and B bands.

Further capitalizing on the complementary nature of EarthShine’s Earth Science and exoplanet goals, we will spatially integrate whole-Earth images to simulate spectral light reflection of the Earth as a point source, which will allow us to qualitatively analyze the influence of different surface types, clouds, aerosols, and gases on light signatures.^13,19 The disc integration involves an ensemble of Earth’s disc elements with specific illumination and emission geometries as viewed from a remote vantage. The observing geometries of LEO and GEO orbits are not well suited for selecting an appropriate near-real-time ensemble of these disc elements at a given time and of course spacecraft in such orbits cannot view the entire Earth disk in a single image. The lunar vantage with a range-to-diameter $\sim 60\mathrm{deg}$ and 6 deg FOV offers a better representation for this simulation.

EarthShine’s two cameras with broad spectral range from visible to IR and day and night coverage (including twilight zone) will allow us to follow dispersion of volcanic or dust clouds for the whole-Earth from daytime to nighttime and back. Since our wavelength coverage includes the ${\mathrm{O}}_2$ A-band ($\sim 760\mathrm{nm}$), we can distinguish between two layers of clouds as demonstrated in Gu et al.¹⁹ (originating from the combined effect of oxygen absorption and cloud heights).

2.3.

Measurements of Lunar Water in Unique Regions

Some surface regions such as lunar swirls may offer new options for long-term human presence on the Moon. Lunar swirls are characterized by higher albedo than the surrounding surface, as well as localized magnetic fields, which may have shielded swirls from solar winds and space weathering. Lunar swirls are thought to be of cometary origin (and thus potentially contain significant buried water). Syal and Schultz²⁰ stated that future missions should test this hypothesis “to fully unravel the secrets of the swirls.” Solar wind interaction with the surface is believed to produce the surficial $\mathrm{OH}/{\mathrm{H}}_2\mathrm{O}$ observed pervasively across the Moon. Observations of ${\mathrm{H}}_2\mathrm{O}$ at a lunar swirl with EarthShine’s WF-LHR will address whether the “mini-magnetosphere” inhibits the production of $\mathrm{OH}/{\mathrm{H}}_2\mathrm{O}$ at swirls.^21–24 Deployment in a lunar swirl is not required to achieve the general science goal of observing water vapor in the lunar exosphere from the lunar surface. However, observation of water in a lunar swirl would provide an interesting data point into the origin of these formations and if water is indeed detected, may provide a future landing habitat.

2.4.

Synergistic Benefits of EarthShine Suite

An essential goal of astrobiology is to detect and understand “other Earths.” Thus Earth’s own complexity presents the critical test to identify key features describing Earth’s climate, properties, and habitability that are detectable at interstellar distance. An “all hands-on deck” approach is essential to capture the complexity of Earth’s varied surface of continents, oceans, and ice fields under a transparent Rayleigh-scattering atmosphere containing water as gas, liquid, and solid simultaneously, hosting planetwide (technological) life, with a rocky satellite in the mix now known to show detectable hydration. The EarthShine suite will sample reflected light, as well as thermal self-emission, to characterize global spectroscopic chemical signatures, rotational modulation, phase curves, and energy balance of Earth, creating a rich phase space that can be explored to identify the most critical features for future reconnaissance, when space observatories will capture the first unresolved low-spectral resolution measurements of terrestrial exoplanets. EarthShine will also be sensitive to measuring water vapor at the lunar deployment site, characterizing a critical feature that may be a source of confusion in unresolved measurements of planet-moon systems. Although EarthShine data will include spatially resolved spectral imaging of Earth that would be impossible for the first-generation detection systems currently envisioned, these data will be needed to constrain what deductions can and cannot be substantiated from future spectroscopy and spectrophotometry. EarthShine will assess the discovery space of measurement targets that are not yet recognized, as well as features of habitability such as the vegetation red edge that have already been proposed, creating a laboratory to pursue investigating possible habitable worlds.

2.5.

EarthShine in Context

Previous studies of Earth as an exoplanet have measured “Earthshine,” defined as sunlight that has been reflected from the dayside Earth onto the dark (shadowed) side of the Moon. Earthshine measured from ground-based telescopes involves the additional step of measuring the light reflected from the dark side of the Moon back to observers on Earth.^14,25,26 Turnbull et al.,¹⁴ collected integrated Earth spectra reflected off the dark Moon and concluded that strong water signatures could be detected that would clearly imply habitability and further advocated that the NIR would be a powerful waveband for undertaking searches for habitability on exoplanets. A complementary transmission spectrum of Earth reflected off the Moon during a lunar eclipse²⁷ showed that several biologically relevant atmospheric features are significantly more prominent in transmission than in reflection. Measuring Earth’s spectrum from the Moon would provide far more robust measurements than measuring Earthshine from Earth’s surface. In ground observations, spectra pass through Earth’s atmosphere three times, complicating correction for telluric absorption and lunar reflectance.⁴ Further, the lunar vantage would allow us to collect spectral features inaccessible from the ground (Table 1). The idea of using the Earth as our best-studied habitable planet, to inform exoplanet research, is not new. From space, Galileo fly-bys in the early to mid 1990s collected images of Earth at a variety of wavelengths and illumination phases (Table 2). Sagan et al.² deduced that Earth was teeming with water and identified signatures that together are strongly suggestive of life, including abundant ${\mathrm{O}}_2$, the VRE, and atmospheric ${\mathrm{CH}}_4$ in thermodynamic equilibrium. The Mars Global Surveyor³ observed Earth during cruise and obtained the first known thermal IR spectra of the Earth’s whole disk, identifying key spectral features in Earth’s atmosphere. Spacecraft in near-Earth orbit have obtained a vast amount of Earth imaging data (e.g., Hearty et al.),¹⁵ although synthesizing the data to produce full-Earth images is challenging (and demonstrates the value of “distant-Earth” observations that can validate calibration and imaging techniques, rendering these data useful for exoplanet studies.)

Table 2

Space-based data on Earth as an exoplanet.

In 2008, as part of the Extrasolar Planet Observations and Characterization (EPOCh/EPOXI) program, the NASA Deep Impact mission observed the “unresolved” Earth as an exoplanet proxy, collecting hourly images in several Vis/NIR photometric bands for three separate 24-h periods at illumination phases from 66% to 77%, using both the high-resolution visible imager and the high-resolution imaging infrared spectrometer. The data imply that sophisticated observers in nearby extrasolar systems could differentiate between surface areas covered by oceans versus continents on Earth,^4,6–8 suggesting that we can pursue the same types of studies when characterizing small rocky exoplanets with future direct-imaging flagship missions. Similarly, the LCROSS mission collected shorter data sets on three separate occasions in 2009 at three widely separated Earth phases, detecting ocean glint and ozone signatures, in addition to further calibrating models.⁹ Further, Kane et al.¹⁰ used multispectral imaging data from EPIC/DSCOVR,¹¹ degraded to a pixel resolution comparable to that of our most optimistic future exoplanet observations and demonstrated that time variations in the data could be used to determine Earth’s rotation, obliquity, and atmospheric albedo—crucial observables expected to determine surface conditions on remote exoplanets. Numerous authors have proposed innovative uses of DISCOVR data to consider Earth as an exoplanet.^{14,15,25–32}

EarthShine will collect observations of Earth from space at illumination phases not yet observed (Fig. 5), with a consistent set of instruments, simultaneously spanning wavelength ranges of interest (Fig. 6). EarthShine’s comprehensive data set will enable us to identify distinctive characteristics of an Earth-like planet and life and to quantify how these features change with time.

Fig. 5

The power of EarthShine. Observations would begin at day 0, viewing the waning gibbous Earth at dusk. Lunar noon, day 7, views the “new Earth” at mission mid-point. The payload would go dormant near day 14, with a waxing gibbous Earth at dawn. By comparison, DSCOVR can view only “full Earth,” while EPOXI viewed only a modest range of phase angle on the dawn side.

Fig. 6

EarthShine tests our ability to detect and interpret terrestrial exoplanets. EarthShine will observe phase angles not well assessed before, measuring signatures of critical species (${\mathrm{H}}_2\mathrm{O}$, ${\mathrm{CO}}_2$, ${\mathrm{CH}}_4$, ${\mathrm{O}}_3$, and ${\mathrm{O}}_2$) in Earth’s reflectance spectrum from the visible to NIR range, over forward-scattering angles from 90 deg (dusk) to “new Earth” at 180 deg (midnight) to 90 deg (dawn), providing continuous coverage of Earth’s varying reflectance phase function for the first time. Horizontal lines represent wavelength ranges of the missions. EarthShine will be the first empirical test for IR methods to investigate exoplanet chemistry and climate.

To establish realistic, disc-integrated (or barely spatially resolved), spectroscopic representation of an Earth-like exoplanet, we need observations of Earth’s entire disk from a remote vantage. Such a database is critical in the interpretation of future observations of an exoEarth. This requires whole-disk spectroscopic observations of Earth acquired at various phase angles. Currently, our continuous whole-disk observations of Earth have primarily come from the DSCOVR and EPOXI missions. However, EPOXI had limited phase sampling and time duration. DSCOVR has observations over a longer timeline but only covers the Earth at full phase and neither transit spectroscopy nor direct imaging will access exoplanets at this phase. EarthShine will fill this critical void in understanding the spectroscopic presentation of an Earth-like exoplanet with observations of the Earth’s full disk at multiple phases, dramatically improving the database of observations and enabling future exoplanet spectroscopy mission design.

We will relate the observed spectroscopic signatures collected with EarthShine to physical processes acting on the planet using accurate radiative transfer modeling and retrieval methods. By employing these methods used to infer properties of exoplanets, we can better understand the relationship between measured signatures on our planet and the significance of actual bio/geo-markers on exoplanets. Over the last decades, modeling and retrieval capabilities for interpreting Earth’s data have matured tremendously,^33,34 with Robinson and Reinhard³⁵ summarizing the latest methods when interpreting Earth’s data as an exoplanet. For our forward models and retrieval analysis, we will employ the planetary spectrum generator (PSG,^36,37 developed to be broad in scope to capture the latest methods and techniques from astronomy, Earth, and Planetary Science communities. PSG integrates the latest advancements in 3D scattering radiative-transfer modeling³⁸ and has access to a comprehensive spectroscopy library, including billions of absorption lines from $>1200$ species, aerosol scattering models for 104 species, collision induced opacity sources for more than 20 pairs, and surface properties for 2000 materials.

Considering the whole-Earth disk nature of EarthShine observations, modeling and interpreting the data require accurately capturing 3D global information. The recently developed module for PSG, GlobES (Global Exoplanetary Spectra), is specifically tailored to allow ingest of 3D climatological data and retrieve/compute information from the emergent spectra. GlobES was recently demonstrated in an application to TESS’s first Earth-size Habitable-zone World, TOI-700d.³⁹ The comprehensive data collected with EarthShine will further advance and validate these methods, following an approach similar to Robinson et al.,⁸ by generating an ensemble of synthetic spectra with PSG to compare with the time-variable spectral and brightness variations collected across the wide EarthShine spectral bands. This will lead to a more robust understanding of model uncertainties, while also facilitating the investigation of seasonal and temporal variations in biosignature components.⁴⁰

2.6.

Unique Advances Enabled by EarthShine

EarthShine is the empirical test needed by the exoplanet community for modeling remote observations of terrestrial planets detected at other stars. EarthShine will cover a broad range of phase angle and wavelength in both reflected light and self-emission, covering the phase angle range of 90 deg through 180 deg, from dusk to midnight to dawn. No prior direct measurements exist of the whole Earth in this range of phase angle, which covers half the range of possible remote measurements. DSCOVR/EPIC has observed Earth extensively in UV, visible, and NIR wavelengths at 2 deg to 12 deg phase angle, and Deep Impact/EPOXI covered a limited range of 57 deg to 75 deg phase angle in visible to NIR wavelengths (Fig. 5). Combined with EPIC and EPOXI reflectance data, EarthShine will define Earth’s reflectance spectrum and phase curve over the full range from back-scatter to forward-scatter and compare dawn to dusk to search for any observable asymmetry. EarthShine will obtain the first ever thermal-IR global measurements over a range of phase angle and with sensitivity to rotational modulation, defining spectroscopic sensitivity required to detect biosignature gases (e.g., ${\mathrm{CH}}_4$ and ${\mathrm{O}}_3$) and habitable environments. These measurements will set requirements for exoplanet measurement capabilities in future flagships, whether using optical coronagraphy, IR interferometry, or direct imaging.

The EarthShine threshold mission is illustrated in Fig. 5, which is compatible with the currently available resources on today’s CLPS providers. Even a short 9- to 14-day mission would yield a significant improvement over the data available from previous missions listed in Table 2. As future power and temperature-control resources improve for the lunar surface payloads, it should be possible to extend the mission duration design to multiple lunar days, and survival through the lunar night, further enhancing the science reach of EarthShine by allowing for observations of Earth phases analogous to those future missions will collect to investigate potentially habitable exoplanets. In the near-term, alternative landing sites on the lunar near-side may shift the mission toward somewhat greater access to the dusk-side or dawn-side range of phase angle. For example, a site at $\sim 60\mathrm{deg}$ Eastern longitude on the Moon (toward the celestial West limb) would see sunrise $\sim 5$ days before gibbous-phase Earth on the dusk side and would see sunset $\sim 5$ days before reaching gibbous phase on the dawn side. Offsetting in the opposite direction in longitude would achieve later sunrise and later sunset, with more access to the dawn side.

2.6.1.

Transmission and reflection spectra

The Earth is the best-studied habitable planet to inform observations from future exoplanet missions. Its transmission spectrum is a proxy for Earth observations during a primary transit as seen from beyond the Solar System, while its reflection spectrum is a proxy for the observations of the Earth as an exoplanet by direct observation (after removal of the Sun’s spectral features).

As an illustration, Fig. 7(a) taken from Ref. 27 shows the Earth’s transmission spectrum, with some of the major atmospheric constituents marked. Below it, we show a comparison between the Earth’s transmission (black) and reflection (blue) spectra. Both spectra have been degraded to a resolution of $0.02\mu \mathrm{m}$ (to emulate spectra from exoplanets) and normalized at the same flux value at around $1.2\mu \mathrm{m}$. It is clear from this figure that the reflection spectrum shows increased Rayleigh reflectance in the blue. It is also striking how most of the molecular spectral bands are weaker, some even non-existent, in the reflection spectrum.

Fig. 7

Observations during lunar eclipse reveal fingerprints of a habitable world.²⁷ Visible-to-near-IR light reflecting off the fully eclipsed Moon passes through a ring of Earth’s atmosphere. (a), (b) The observed transmission spectrum (black) is analogous to what a distant observer watching Earth transit in front of the Sun would see, while measurements at other lunar phases are analogous to a directly imaged reflection spectrum (blue).

Variations due to large-scale exoplanetary features such as continents and oceans will be contained in perhaps a single pixel. Disentangling the various components that give rise to time variability in the signal will be a challenge. To assess the limits of this issue, our image data will be further deconvolved to $3\times 3\text{pixel}$ resolution following the approach used by Kane et al.¹⁰ on EPIC data (see Fig. 8).

Fig. 8

(a) Earth as seen from Earth/Sun L1, using the DSCOVR-EPIC camera. (b) The same image degraded to $3\times 3\text{pixel}$ resolution, analogous to expectations of future exoplanet observations. A single, or a few, pixels may contain all the available information on the habitability of an exoplanet. Credit: NOAA/NASA/Kane et al., 2017.¹⁰

2.7.

Lunar Site Selection

The science achievable with the EarthShine instrument suite depends very little on the choice of landing site, although longitude offset will affect the details of the beginning and ending Earth phase available for observation. Provided that the site is located on the Moon’s near side, and that EarthShine is deployed with an uninterrupted view of the Earth, our instruments can achieve all of the science goals described in this paper. As seen from the Moon, the Earth’s position in the sky follows an elliptical path with a semimajor axis of $\sim 15\mathrm{deg}$, passing north and south of the ecliptic by $\sim 5\mathrm{deg}$. Thus the requirement to have the Earth constantly in the EarthShine field of view rules out sites above $\sim 85\mathrm{deg}$ in lunar latitude.

As seen from the Moon, the Earth’s disk subtends an angle of $\sim 1.9\mathrm{deg}$, so the details of the eclipses and occultations described above vary in only minor ways regardless of whether the chosen site is equatorial or at high latitude. The thermal load on the payload will, of course, depend strongly on the lunar latitude, as described fully in Sec. 4.

2.8.

Additional Potential Secondary Science Objectives

2.8.2.

Inner solar system zodiacal dust observations in visible wavelengths

Although the innermost parts of the zodiacal cloud (ZC) are studied by solar oriented missions (spacecraft traveling toward the Sun;⁵⁸ spacecraft orbiting at 1 AU⁵⁹), the global shape of the inner parts of the ZC has been the subject of only a few studies. The global shape of the ZC is a key component of our understanding the structure and origin of the dust in the Solar System, thus providing important insight into Solar System space weather and exosolar dust debris disks. Using the lunar surface/horizon as an effective Sun shade, we can in better detail analyze the extent of the ZC in visible light using the EarthShine star tracker similar to that on the Clementine spacecraft⁶⁰ (see Fig. 9).

Fig. 9

Artist’s impression of the zodiacal light observation from the lunar surface. (a) Simulated zodiacal light observation from a lunar lander pointed to the west. The zodiacal light appears as a triangular white glow. The Sun is several degrees below the lunar horizon allowing us to probe the fine structure of the zodiacal light in the inner solar system. The white dashed line represents the ecliptic which represents an approximate line of symmetry of the zodiacal cloud in the inner solar system. (b) The sketch version of (a). The lunar horizon is represented by a solid black line while the dark lunar surface is shown as a shaded area. The Sun is located several degrees under the lunar horizon and is represented by a black symbol. The ecliptic is shown as a black dashed line. The zodiacal light is shown as a gradient of grayscale approximately resembling our zeroth-order estimate for the visual flux. Notes: (a) was created using the Stellarium program (zodiacal light and celestial sphere) and Apollo 17 panorama (lunar surface). (b) is a heavily processed version of (a).

The global shape of the ZC is driven by various processes: (1) the production of dust in asteroids and comets,⁶¹ (2) the dynamics of the dust driven by gravity and radiative processes,⁶² and (3) the elimination of dust in various sinks.⁶³ Each model of the ZC strives to match the global density profile of the dust, however, such observations are rare.⁶⁰ The ZC is composed of dust particles originating from vastly different sources: main-belt,⁶⁴ Jupiter-family comets,⁶⁵ and long-period comets.^66,67 Each of these sources provides a unique signature in the ZC and their mixing ratios are not well understood.⁶⁸ The Clementine spacecraft 6-week survey provides the only global inner solar system snapshot of the ZC in visible wavelengths.⁶¹ The Clementine mosaic suffers from several drawbacks: (1) it extends only from 3 deg to 30 deg away from the Sun, (2) its detector is rather primitive compared to today’s standards, and (3) due to various exposure times and only a six-week window, the data product suffers from poorly defined uncertainties and pollution from bright Venus’ signature.

We propose to reproduce the Clementine mosaic in much greater detail and longer data coverage using the EarthShine star tracker cameras. Thanks to our enhanced knowledge of the inner solar system dust dynamics⁴⁵ and previous observations by various spacecrafts and especially Clementine,⁵³ we can fine tune the star tracker exposure time to better survey the inner ZC. To overcome the limited data bandwidth, the data will be analyzed in situ and only the final data products will be sent back to Earth. The original Clementine data set consists of only seven observations, thus there is a vast room for improvement to be made by Earthshine.

3.1.

CHEILS

Developed at LLNL, CHEILS is an Earth-imaging grating spectrometer with a fused silica monolithic telescope fore-optic that is pointed at the Earth using a two-axis gimbal mechanism and ultimately focused on a COTS Sony IMX990 back-illuminated InGaAs detector housed in an Aval Imaging ABA-013VIR camera. A schematic of CHEILS is shown in Fig. 11 with key specifications summarized in Table 3.

Fig. 11

CHEILS schematic. Elements are cemented together and mounted in an Invar housing. Monolithic construction provides high environmental robustness, no postlaunch and landing alignment needed.

Table 3

CHEILS telescope, spectrometer, and detector descriptions.

The compact, high-performance, monolithic telescope [Fig. 12(a)] was originally developed for deployment on small satellites and aircraft where size and weight come at a premium. We refer to them as monolithic telescopes because they are constructed from a single block of high-purity fused silica, which provides intrinsic mechanical/thermal stability (negligible thermal expansion) when operating in extreme environments. The spectrometer disperses incoming light using an in-house ruled dual-blaze grating on a ZnS substrate [JWST near-infrared imager and slitless spectrograph (NIRISS) grating ruled in ZnSe shown in Fig. 12(b)].

Fig. 12

(a) Monolithic telescope mounted in Invar housing and (b) LLNL ruling JWST NIRISS grism.

Pointing of CHEILS is accomplished with a gimbal that is being designed by LLNL for an International Space Station (ISS) mission with a delivery set for late 2021 and a launch in January 2023. The mechanism utilizes two Empire Magnetics motors, each driving a pulley system.

During operation, CHEILS step-stares across the Earth’s disk at a rate of $7.0\mathrm{arc}\sec /\text{step}$ with a maximum scan time of $\sim 20\min$. Each scan produces a data cube with native spatial dimensions of 1280 cross-track × 1024 along-track with each spatial location having an associated 1024 bin spectrum covering a 400- to 1750-nm spectral region. Integration time is set to produce ½ to ⅔ maximum well-filling for high-albedo locations (Fig. 13). Depending on the scan frequency and uplink availability/capacity, onboard data processing algorithms will reduce data cube volume by rebinning spatially and/or spectrally, cropping data cube to pixel’s above detection threshold, and applying data compression for uplink.

Fig. 13

(a) Calculated well-filling and (b) SNR values for Lambertian reflection of solar top-of-the-atmosphere radiance and from Earth’s surface with 0.9 albedo. Integration time/frame = 1 s.

Observations of solar eclipses (Sec. 2.6.2) and stellar and planetary occultations by the Earth’s atmosphere (Sec. 2.6.3) would place additional requirements on the CHEILS hardware. CHEILS cannot directly observe the Sun without damage unless we place a neutral density filter in the optical path ahead of the focal plane array (FPA). This filter mechanism is not included in the instrument diagrams provided here to reduce complexity but can be included at a later date. Viewing the Earth while in total eclipse would be feasible with the current instrument configuration. Viewing partial eclipses, or at other times when the Sun is in or close to the FOV, would require the neutral density filter.

The CHEILS telescope will have a baffle in front to mitigate stray light. Measurements of similar baffles show a $1\times {10}^{-12}$ stray light background per Earthshine pixel when the light source is 5 deg outside the field, and $1\times {10}^{-14}$ at 30 deg. As a $1\times {10}^{-12}$ intensity reduction is equivalent to $\sim 30$ visual magnitudes and the brightness difference between the Sun and the Earth is $\sim 11$ magnitudes, a solar angle limit of 5 deg would likely be sufficient for operational purposes.

The feasibility of the stellar eclipses by the atmosphere will depend critically upon the details of the payload operations, which have yet to be specified for the Artemis program. CHEILS is designed to sweep over the $\sim 2\mathrm{deg}$-wide disk of the Earth with 0.25 deg of cross-track pointing error, accommodated by an oversized slit. For a stellar occultation measurement, the 2.5-deg slit projection should be more than adequate to accommodate any cross-track pointing error. The along-track accuracy, however, would present a challenge: the slit width will subtend $\sim 10\mathrm{arc}\sec$, so without dithering the required pointing accuracy would need to be $\sim 2\mathrm{arc}\sec$. This may be achievable later in the mission, once the sensor pointing offsets have been characterized well enough.

3.2.

WF-LHR

The WF-LHR is based on the GSFC team’s earlier laser heterodyne radiometer developments (shown in Fig. 14) that include an Earth-based ground instrument that measures ${\mathrm{CH}}_4$ and ${\mathrm{CO}}_2$ in the atmospheric column (mini-LHR), and an occultation-viewing CubeSat (MiniCarb) designed to observe ${\mathrm{CH}}_4$, ${\mathrm{CO}}_2$, and ${\mathrm{H}}_2\mathrm{O}$ from an LEO.^69–73 Similar in design, the WF-LHR will be solar pointing and observe ${\mathrm{H}}_2\mathrm{O}$ in the lunar exosphere at sunrise and sunset when the pathlength is closest to the lunar surface and has the highest density of molecules.

Fig. 14

LHR developments: (a) ground LHR in Alaska measuring ${\mathrm{CH}}_4$ over permafrost; (b) Hawai’i Space Exploration Analog and Simulation team members using a ground LHR during a long-duration Mars simulation program; (c) the joint GSFC/LLNL CubeSat MiniCarb at integration; and (d) MiniCarb on the ISS prior to deployment (payload at upper left, indicated in red).

A schematic of the WF-LHR is shown in Fig. 15. A fiber-coupled telescope mounted to a sun tracker delivers light to mix with a distributed feedback laser centered at $\sim 2.67\mu \mathrm{m}$. The interference between frequencies enables a resolving power of 160,000 with signal-to-noise ratio (SNR) $>300$ for 1 s integration time/data point. The laser scans across these features and the resulting beat signal is collected in the radio frequency and converted to mole fraction.

Fig. 15

The WF-LHR design leverages postage-stamp-sized distributive feedback lasers from the telecommunications industry to produce a compact and cost-effective package.

The WF-LHR uses a GSFC-developed tracker for solar pointing. The LHR collimator is housed within the tracker base as are two stepper motors for control of zenith and azimuth and control electronics. The pointing accuracy is better than 0.06 deg and it operates autonomously. The tracker dimensions are (508-mm height, $159\times 166\text{-}\mathrm{mm}$ width), and approximate mass is 4.5 kg.

Figure 16(a) shows sample data from column observations from one of our ground mini-LHR instruments operating at a permafrost site in Alaska with comparisons to a co-located flux tower that is observing gas concentrations at the surface. Data (blue) are analyzed using the retrieval tool of Goddard’s PSG.^38,74,75 The fit from this retrieval is shown in orange. Figure 16(b) shows projected observations of water vapor in the lunar exosphere. In this PSG simulation, the transmittance is modeled for a solar-pointing WF-LHR that is deployed in Reiner Gamma assuming a hydrostatic equilibrium atmosphere, a surface pressure of $1\times {10}^{-3}\mathrm{Pa}$, molecular weight of $40\mathrm{g}/\mathrm{mol}$, atmospheric temperature of 300 K, surface temperature of 130 K, albedo of 0.136, and emissivity of 0.864. During operations, the WF-LHR points at the sun and observes absorption from water vapor in the lunar exosphere at sunrise and sunset to maximize airmass. The WF-LHR scans continuously from 2.6669 to $2.6674\mu \mathrm{m}$ with 60 points per scan, each scan taking approximately 1 min assuming a 1-s integration time per data point.

Fig. 16

(a) Heritage from mini-LHR ground observations on Earth. Half-hour column data products show good agreement with surface observations at the site. Data are shown in blue and the PSG retrieval fit is shown in orange. (b) Simulated transmittance of ${\mathrm{H}}_2\mathrm{O}$ vapor lines in the lunar exosphere modeled with the PSG for the WF-LHR (Villanueva et al. 2018,³⁶ Edwards 1992³⁴).

3.3.

EPyC-IR

The EPyC-IR is a GSFC-developed compact hyperspectral imaging camera spanning the 2.5- to $12.5\text{-}\mu \mathrm{m}$ MWIR to LWIR spectral range. This spectral range contains habitability diagnostic spectral features of an exoEarth (${\mathrm{H}}_2\mathrm{O}$, ${\mathrm{CH}}_4$, ${\mathrm{CO}}_2$, ${\mathrm{O}}_3$, ${\mathrm{N}}_2\mathrm{O}$, CO, ${\mathrm{SO}}_2$, etc.), existence of atmospheric stratification, radiative balance between parent star insolation and thermal emission, surface properties regulating emissivity, and thermal structure (Fig. 17). EPyC-IR consists of the type-II strained layer superlattice (SLS) sensor, linear variable filter (LVF) assembly, Ricor cryocooler, and a catadioptric telescope. The heart of the focal plane imaging system is the sensor chip assembly consisting of the $1024\times 1024$ format $18\mu \mathrm{m}$ pixel pitch SLS InAsSb/InAs sensor chip hybridized to an FLIR Indigo ISC0404 readout-integrated circuit. The pixel spectral response of the SLS technology is extremely uniform and temporally stable. The focal plane assembly is cooled to $\sim 70\mathrm{K}$ by a miniature Ricor K508N Stirling tactical cryocooler.

Fig. 17

EPyC-IR spectroscopy complements CHEILS reflected light measurement. Earth’s nadir-viewed spectrum from 2.5 to $12.5\mu \mathrm{m}$ (modeled) is bracketed by EPyC-IR and includes reflected light for $<3\mu \mathrm{m}$ (opaque due to water opacity) transitioning to thermal emission at longer wavelengths. Critical gases diagnostic of the habitable Earth-like atmosphere are clear in this range—${\mathrm{H}}_2\mathrm{O}$, ${\mathrm{CO}}_2$, ${\mathrm{O}}_3$, and ${\mathrm{CH}}_4$. These signatures can be measured in exoplanets only from space, and thus methods can be tested only in space observations of Earth. Balance between pure blackbody functions (dotted lines for 260, 270, 300, and 310 K) and calculated atmospheric emission (snow, forest, desert, ocean, and cloud) constrain composition and thermal structure. A cloudy planet presents a colder scene for 3.5 to $4.2\mu \mathrm{m}$ and $>10\mu \mathrm{m}$; the 300-K vegetation spectrum is dominated by lower emissivity than a 300-K ocean or desert in the LWIR region and is manifest as the dimmest of these three scenarios in this range. The global (disc integrated) spectrum includes contributions from terrains with unique spectral signatures such as visible suppression by savanna vegetation, low-albedo over ocean, and reduced emission due to ice absorption at 3.5 to $4\mu \mathrm{m}$ over snow. Models for crescent-phase spectroscopy change the weighting of features and must be tested empirically.

A commercial 200 mm focal length, $f/2$ catadioptric telescope from StingRay Optics also brackets the same spectral range and matches the $18\mu \mathrm{m}$ pixel scale. EPyC-IR is supported by an azimuthal scan and elevation-positioning mechanism leveraged from another instrument.

Compact thermal imager (CTI),⁷⁶ the heritage instrument, successfully observed Earth system scenes, including ocean, forest canopy, clouds, savanna and deserts, cryosphere regions, and transients such as biomass combustion (e.g., Fig. 17) from the ISS in 2019. EPyC-IR leverages design features from prior developments: electronics and imaging core (CTI), LVFs (OSIRIS-REx/OVIRS), and optics and scan mechanism (other mission instruments).

To acquire the hyperspectral data cube of Earth disc for a given phase, EPyC-IR will use a gimbal mechanism, which will enable step-and-stare frame acquisition bracketing the disc with all spectral bands. The LVF results in spectral columns sampling different spatial regions; the mechanism then steps the 3-deg EPyC-IR field of view by 0.03 deg (corresponding to the 1% filter FWHM or 102 pixels of the 1024 span).

During step-stare operations, EPyC-IR frames are captured at a regular 50 Hz rate with an integration time of $\sim 1\mathrm{ms}/\text{frame}$, acquiring 1-ms exposures at 20 ms frame intervals. Fifty frames are coadded and stored in the instrument control electronics (Fig. 18). Each column of 102 pixels projected to sky plane samples a spatial slice in the spectral band of the column, with adjacent column sampling a different spatial region at the next wavelength; the mechanism then steps EPyC-IR by 0.03 deg (corresponding to the 1% filter FWHM or 102 pixels of the 1024 span). Repeated step-and-stare operations capture the Earth’s disk (100 steps of 0.03 deg) with a set of 100 spectral bands as the EPyC-IR FOV transitions between the starting and ending FOV configuration (Fig. 18).

Fig. 18

EPyC-IR hyperspectral maps across Earth. EPyC-IR sweeps an LVF in step-and-stare operation to acquire a hyperspectral data cube. Three contiguous $3\mathrm{deg}\times 3\mathrm{deg}$ fields-of-view are required to sample the 2-deg diameter Earth (a) in all wavelength bands to reconstruct wavelength band-sequenced imaging frames. (b) Each colored column of the $3\mathrm{deg}\times 3\mathrm{deg}$ FOV represents a wavelength band on the LVF and each column samples a different spatial region. In the step-and-stare operation, each column will gradually sweep adjacent spatial regions. Thus a sweep between the starting and ending field-of-view positions (a) will produce one data cube of the center field of view with complete spectral sampling. Spectral resolving power and/or spatial resolution are selectable to meet data transmission limitations.

EPyC-IR is currently optimized for Earth disc measurements rather than stellar occultation observations through the Earth’s atmosphere, but such observations should still be possible. Efficacy will be established through models to establish limiting stellar brightness as a function of lander heat rejection capabilities. For example, additional cooling of telescope optics elements will allow for reduction in thermal emission with a resulting improvement in the detection sensitivity of stellar point sources.

For the case of a total solar eclipse, a fully unilluminated Earth allows for greater sensitivity. Since totality within the umbral region will be longer than that typical of solar eclipses on Earth, we envision pointing the telescope to observe the Earth’s limb to observe stellar occultations. An additional mitigation under consideration is to include a long-pass optical filter ($>2.4\mu \mathrm{m}$) to block the visible-to-SWIR solar insolation which may cause damage to optical coatings. The optimum position for such a filter is in front of the primary mirror (catadioptric telescope design) and requires a broadband antireflection coating (2.5 to $12.5\mu \mathrm{m}$) to minimize optical losses.

The ability to discriminate molecular vibrational signatures of water vapor from OH is unambiguous in the MWIR/LWIR spectral region. The $2.7\text{-}\mu \mathrm{m}$ vapor phase stretch mode ($\sim 3\mu \mathrm{m}$ in water ice and hydroxyl species) is common to both ${\mathrm{H}}_2\mathrm{O}$ and OH; in comparison, the $\sim 6\text{-}\mu \mathrm{m}$ vibration bending mode is unique to molecular ${\mathrm{H}}_2\mathrm{O}$ and is a prime water proxy in a surface water rich planet such as Earth. In addition to thermometry of the planet, other signatures relevant to an exoEarth analog in this spectral region include ${\mathrm{CO}}_2$, ${\mathrm{O}}_3$, and ${\mathrm{CH}}_4$. Infrared self-emission is an effective way to characterize terrestrial exoplanets due to the detectability of these diagnostic atmospheric species at habitable temperatures as well as constraining theoretical estimates for equilibrium temperature. The relatively favorable contrast ratio between the primary star and planetary thermal emission, $\sim 1:{10}^4$ compared to $\sim 1:{10}^9$ at visible wavelength, has previously led to developing designs for large-scale infrared interferometers that may yet be selected for deployment. Exploring this spectral range is essential to evaluate the viability of the various spectral intervals available.

Characterization by EarthShine in MWIR and LWIR will enable identification of spectroscopic wavelengths and optimum bandwidths for high spatial resolution chronographic systems optimized for capturing images of an exoEarth in biosignature-indicative spectral bands. Other applications of the characterization include transit studies of exoplanets in diagnostic wavelengths identified by EPyC-IR to identify and quantify composition and thermal structure.