Detectability of Molecular Signatures on TRAPPIST-1e through Transmission Spectroscopy Simulated for Future Space-Based Observatories

As discoveries of terrestrial, Earth-sized exoplanets that lie within the habitable zone of their host stars continue to occur at increasing rates, efforts have began to shift from the detection of these worlds to the characterization of their atmospheres through transit spectroscopy. While the detection of molecular signatures can provide an indication of the presence of an atmosphere, Earth-like exoplanets create an exciting opportunity to further characterize these atmospheres by searching for biosignatures that may indicate evidence of past or present life. To date, detection methods have focused on promising targets that orbit M-dwarf stars, such as TRAPPIST-1e, that have a rocky composition and lie within the habitable zone of their host star. While JWST will provide new insights on the atmospheric compositions of these exoplanets, terrestrial planets that fall in the habitable zone of close-in systems will continue to pose challenges in spectroscopy. Herein, we use a Global Climate Model (GCM), a photochemical model, and a radiative transfer suite to simulate an atmosphere on TRAPPIST-1e that assumes the boundary conditions of modern Earth. The detectability of biosignatures on such an atmosphere via transmission spectroscopy is modeled for JWST, where mission concepts such as LUVOIR, HabEx, and Origins are used to compare potential capabilities for the distant future. Despite the drastic increase of aperture size and instrument sensitivity for future observatories, we show that only CO2 would be detectable in transmission spectroscopy for such an atmosphere on these planets, as the presence of clouds and their impacts on scale height strongly limits their molecular detectability. In such a case, the synergy between space- and ground-based spectroscopy may be essential in order to overcome these difficulties.


INTRODUCTION
The search for small, Earth-like rocky exoplanets has made significant progress since the launch of Kepler in 2009. To date, there are over 4,000 confirmed exoplanets orbiting around a multitude of stars, with additional candidates being selected almost daily. As we continue to detect terrestrial exoplanets that resemble Earth in size, some of the most exciting discoveries point towards planets that fall within the habitable zone (HZ), the region around a star where liquid water may be present Transmission Spectroscopy of TRAPPIST-1e 3 under the correct conditions (Kasting et al. 1993;Kopparapu et al. 2013). Such detections have thus far been biased to low mass stars (late-K and M-dwarf stars), as their small size and compact HZs give way to planets with short orbital periods that provide a high frequency of transits. While both radial velocity and transit techniques have revealed the first rocky exoplanets orbiting within the habitable zone of their low-mass host stars (Mayor & Queloz 1995;Anglada-Escudé et al. 2016;Gillon et al. 2016Gillon et al. , 2017, the majority of planets discovered through the transit method, specifically, orbit extremely close to their host star (Kaltenegger et al. 2012). This bias has influenced the discovery of many rocky planets orbiting M-type stars which are of particular interest as their high planet/star contrast ratio offers a strong possibility of having their atmospheres observed in the future. (Pallé 2018). It has been estimated that 0.16 % of M-dwarf stars contain terrestrial-sized planets orbiting within the habitable zone (Dressing & Charbonneau 2015).
For some exoplanets, transit spectroscopy and/or secondary eclipse measurements (primarily done from space with the Hubble Space T elescope (HST ) and the Spitzer Space T elescope) have provided empirical details on their atmospheric compositions (e.g. Seager & Deming 2010;Sing et al. 2016).
With a few exceptions (e.g. Kreidberg et al. 2014;de Wit et al. 2018), these investigations have primarily targeted so-called hot Jupiters, gas-giant planets with orbital periods of only a few days.
However, as discoveries of rocky exoplanets both within and outside of the HZ continue to increase, the first attempts to put constraints on their atmospheric properties have begun (de Wit et al. 2016(de Wit et al. , 2018Delrez et al. 2018). Within this assortment of planets, one of the most exciting and nearby exoplanetary systems that is a target for future observations is the TRAPPIST-1 system (Gillon et al. , 2017. Bearing seven Earth-sized exoplanets (Gillon et al. 2017) orbiting an ultra-cool late-type M-dwarf star ( M8V Liebert & Gizis (2006)) located 12.4 parsec from Earth (Kane 2018), the TRAPPIST-1 planets are similar in size and irradiation to the rocky planets within our Solar System (Gillon et al. 2017). Their ultra-cool, low-mass parent star signifies that the evolution of their existence and the pathways they undertook to form are potentially much different than what our Solar System planets experienced (Turbet et al. 2018). This leaves us with the ideal laboratory to study how the atmospheric evolution of a planet orbiting an M-dwarf star can impact its habitability (Wolf 2017;Lincowski et al. 2018;Turbet et al. 2018). Among the seven planets, 3-D climate simulations have shown that TRAPPIST-1e could be the most habitable planet of the system, being able to maintain liquid water on its surface across a large range of atmospheric compositions (Wolf 2017;Turbet et al. 2018;Fauchez et al. 2019a;Fauchez et al. 2019b). This makes it an ideal target to search for the presence of biosignatures, molecular features that may indicate evidence of past or present life.
For the TRAPPIST-1 system, data obtained by HST provided initial constraints on the extent and composition of the planet's atmospheres, suggesting that the four innermost planets do not have a cloud/haze-free H 2 -dominated atmosphere (de Wit et al. 2018). However, follow up work by Moran et al. (2018) has shown that HST data can be fit to a cloudy/hazy H 2 -dominated atmosphere.
Complementary to HST , NASA's Spitzer Space T elescope has fostered notable breakthroughs in exoplanet detection, including the discovery of four of the seven TRAPPIST-1 planets (TRAPPIST-1d, e, f, and g), and has been used to constrain their orbital and physical parameters (Gillon et al. 2017). Spitzer has also allowed us to put additional constraints on the atmospheric composition of TRAPPIST-1 b, where Delrez et al. (2018) has found a +208 ± 110 ppm difference between the 3.6 and 4.2 µm Spitzer bands, suggesting CO 2 absorption. The ability to determine whether the TRAPPIST-1 planets have high molecular weight atmospheres or no atmospheres at all requires additional observations with future facilities.
The next generation of observatories will allow for far more in-depth explorations of atmospheric properties of the TRAPPIST-1 planets. In particular, data from the James W ebb Space T elescope (JW ST ) could provide strong constraints on atmospheric temperatures and on the abundances of molecules with large absorption bands . JW ST houses two science instruments capable of using transit spectroscopy to detect light from planets and their host stars: The Near-Infrared Spectrograph (NIRSpec), and Mid-Infrared Instrument (MIRI). NIRSpec intends to analyze the spectrum of over 100 objects observed simultaneously, covering the infrared wavelength range Transmission Spectroscopy of TRAPPIST-1e 5 from (0.6-5 µm). The Mid-Infrared Instrument (MIRI) has both a camera and a spectrograph that perform between the range of 5-28 µm. Only the low resolution spectroscopy (LRS) mode allows for time series observations with MIRI.
Thus far, many studies have been done to evaluate the potential of JW ST to characterize the TRAPPIST-1 planets. Morley et al. (2017) determined that less than 20 transits are needed to rule out a flat line for a 5σ detection of spectral features in a CO 2 -dominated atmosphere on six of the seven TRAPPIST-1 planets, while its ability to characterize individual molecular features using transit spectroscopy will be much more limited. Transit spectroscopy measurements from JW ST will be severely impacted by the presence of clouds on terrestrial exoplanets, which we would expect to see on a potentially habitable or inhabited planet (Fauchez et al. 2019a). Upon placing clouds in the atmosphere through the use of a 3D global climate model (GCM), a terrestrial planet such as TRAPPIST-1e will only allow for the detection of CO 2 if it contains an atmosphere similar to that of modern or Archean Earth (Fauchez et al. 2019a). Although this finding allows us to gain a deeper understanding of the requirements to detect an atmosphere on rocky exoplanets around M-dwarfs, HabEx: -While HabEx proposes a multitude of architectures, this work uses the 4-m monolithic, off-axis telescope concept with a wavelength range of 0.2-1.8 µm. It is equipped with a suite of four proposed instruments that demonstrate various science capabilities, but the most relevant instrument for this work is the HabEx Workhorse Camera (HWC, 0.2-1.8 µm) (Gaudi et al. 2018). HabEx intends to include starlight suppression technologies such as a starshade, and/or a coronagraph.
Origins: -The current design concept for Origins is a 5.9 m on-axis telescope with a Spitzer-like structure that allows for minimal deployment while having a collecting area equivalent in size to that of JW ST (Battersby et al. 2018). Observations with Origins intend to have a high sensitivity that covers a broad wavelength range (3-600 µm).
Origins proposes multiple science instruments, but the most appropriate for conducting transmission spectroscopy measurements is the Mid-Infrared Spectrometer Camera-Transit Spectrometer (MISC-T) that operates across the 2.85-20.5 µ wavelength range.
The objective of this work is to cross-compare the capability of each of these future space-based missions to characterize TRAPPIST-1e (or an equivalent potentially habitable exoplanet) via transmission spectroscopy. The paper is structured as follows: Section 2 discusses the method and the tools used in this study to simulate both the climate and the transmission spectra of TRAPPIST-1e. Details on the LMD-G GCM can be found in Turbet et al. (2018); Fauchez et al. (2019a). In this work, we have performed climate simulations of TRAPPIST-1e using the planet parameters from (Gillon et al. 2017;Grimm et al. 2018). Herein, TRAPPIST-1e is assumed to be fully covered by a 100 m deep ocean (aqua-planet) with a thermal inertia of 12000 J · m −2 · K −1 · s −2 without ocean heat transport (OHT). TRAPPIST-1e is also assumed to be in synchronous rotation. The horizontal resolution of the model is 64×48 coordinates in longitude × latitude (e.g., 5.6 • ×3.8 • ). In the vertical direction, the atmosphere is discretized in 26 distinct layers using the hybrid σ coordinates (with the top of the model at 10 −5 bar) while the ocean is discretized in 18 layers. The stellar TRAPPIST-1 emission spectrum was computed using the synthetic BT-Settl spectrum (Rajpurohit et al. 2013) assuming a temperature of 2500 K, a surface gravity of 10 3 m · s −2 and a metallicity of 0 dex. Figure 1 shows the surface temperature map for TRAPPIST-1e with a modern Earth-like atmosphere (1 bar of N 2 and 376 ppm of CO 2 ) with a surface pressure of 1 bar. H 2 O vapor is brought up to the atmosphere via evaporation of the ocean's surface. The black line delimits the area where the surface temperature is above the freezing point of water and therefore represents the HZ where liquid water can be present on the surface. TRAPPIST-1e is therefore locally habitable when considering an atmospheric composition with modern Earth-like boundary conditions (Turbet et al. 2018;Fauchez et al. 2019a;Fauchez et al. 2019b).

Photochemistry simulations with the Atmos Model
The LMD-G GCM, like most GCMs used in exoplanet research, does not include photochemistry prognostically. Therefore, in order to simulate an atmospheric composition more complex than the 8 Pidhorodetska et al. one provided by the GCM, the addition of off-line 1-D photochemistry is implemented using the Atmos code. Atmos is a 1-D radiative-convective climate model, coupled with a 1-D photochemistry model used to simulate various exoplanet atmospheres (Arney et al. 2016(Arney et al. , 2017Lincowski et al. 2018;Meadows et al. 2018). The boundary conditions for the modern Earth-like atmosphere are described in (Fauchez et al. 2019a) Table 2, adapted from Lincowski et al. (2018) in Table 8 PSG is a spectroscopic suite that integrates the latest radiative transfer methods and spectroscopic parameterizations while including a realistic treatment of multiple scattering in layer-by-layer pseudospherical geometry (Villanueva et al. 2018). PSG permits the ingestion of billions of spectral lines of over 1,000 molecular species from several spectroscopic repositories (e.g., HITRAN, JPL, CDMS, GSFC-Fluor). For this investigation, the molecular spectroscopy is based on the latest HITRAN database (Gordon et al. 2017), which is complemented by UV/optical data from the MPI database (Keller-Rudek et al. 2013). For moderate spectral resolutions (λ/∆λ < 5000) as those presented here, PSG applies the correlated-k technique for the radiative transfer portion, while multiple scattering from aerosols is performed by PSG using the discrete ordinates method, in which the radiation field is approximated by a discrete number of streams distributed in an angle with respect to the planeparallel normal.
In order to properly capture the diversity of atmospheric conditions at the terminator as computed by the GCM, the transit spectra presented in this work were computed by running PSG at each lat-lon bin at the terminator of the planet. Information about temperature, pressure and abundance profiles at each lat-lon gridpoint from the GCM were ported into the input parameters for the spectroscopic simulations performed with PSG. These individual transit spectra were then averaged to compute the total planetary transit spectra. Considering that the spacing of the latitudinal points is constant in the GCM, the integration weights for each spectrum were assumed to be the equal, and a simple average of the transit spectra was performed.  while contributing to the grey area beneath the black line that corresponds to the total transmission spectrum. We see that in the UV and visible, O 3 and N 2 (Rayleigh scattering) are the main contributors to the spectrum. In the near and mid-infrared, many wide H 2 O absorption bands are present, along with some weaker CH 4 bands. The CO 2 features have the strongest relative transit depth comparable to the O 3 feature at 9.6 µm. Note that two collision-induced absorption (CIA) Transmission Spectroscopy of TRAPPIST-1e 11 features are particularly notable on the spectrum: the N 2 -N 2 CIA at 4.3 µm (Schwieterman et al. 2015) and the O 2 -O 2 CIA at 6.4 µm (Fauchez et al. 2019c). The former overlaps the strong CO 2 feature and will be detectable only in the absence of CO 2 . Panel (C) is similar to panel (B), aside from the presence of clouds whose location is predicted by the LMD-G GCM that are included within the radiative transfer calculations. Here, we see a significant decrease in the relative transit depth of each line. It is noted that clouds are strongly opaque to the visible and infrared transmitted radiations. As a result, the spectral continuum is raised above the cloud deck where the atmosphere is semi-transparent (Fauchez et al. 2019a;Suissa et al. 2019a,b). Because the relative transit depth corresponds to the transit depth in the continuum subtracted from the transit depth in the line, a higher continuum reduces the relative transit depth. Origins would require the use of transmission spectroscopy to characterize the atmosphere of planets orbiting in the HZ of ultra-cool M dwarfs such as TRAPPIST-1. Direct imaging would not be possible for such close-in systems because of their inner working angle (IWA) and temperatures that are too cold to be characterizable in emission spectroscopy (Lincowski et al. 2018;Lustig-Yaeger et al. 2019;Fauchez et al. 2019a). Fortunately, each of these future observatories would have at least one instrument with transmission spectroscopy capabilities. The characteristics of these instruments are summarized in Table 1. Eff. aperture (m) 5.6 5.9 4.0 8 15 Figure 3. Same than Fig. 2 panel (C) but with wavelength ranges of the instruments over-plotted. Figure 3 shows the same transmission spectrum as that in the bottom panel of Fig. 2, but with the addition of the wavelength range covered by the instruments. We see that depending on the telescopes and/or instruments, different spectral lines would be detectable. For instance, while LUVOIR has the largest aperture, its wavelength coverage (cf. Table 1) does not include the strongest CO 2 bands at 2.7 or 4.3 µm, and it operates in the spectral region where cloud opacity is the most prominent.

Identification of Spectral Lines for a Modern Earth-like Atmosphere on TRAPPIST-1e
As a result, the water and O 2 lines in this region are far too shallow to be detectable even with the largest aperture size.

13
When instrument performances are compared, several parameters would be at play to detect specific molecular species: • Wavelength coverage: Different spectral lines are accessible depending on the wavelength coverage of the instrument.
• Resolving power (R): λ/δλ with λ the wavelength and δλ the spectral resolution. Reducing R allows to increase the number of photons per spectral bands, reducing the noise. However, R should be high enough to spectrally resolve the the width of the spectral feature. In this work we have optimized the resolving power by finding the lowest R to maximize the S/N.
• Aperture size: A larger aperture size collects more photons improving the S/N and therefore reducing the integration time needed to detect a given spectral feature.   square root of the number n of collected photons ( √ n). Yet, the increase in the number n of photons collected depends on the ratio between the radius squared of the two mirrors (R A /R B ) 2 (assumed to be perfect disks), with R A the radius of LUVOIR-A (7.5 m) and R B the radius of LUVOIR-B (4 m).
The S/N therefore improves by a factor R A /R B =7.5/4=1.875 between LUVOIR-B and LUVOIR-A.  (2018)) and could therefore be difficult to disentangle. Also, the host star is so dim that the number of photons transmitted through the planet's atmosphere is orders of magnitudes lower than for planets orbiting G-dwarfs. The S/N therefore improves slowly while acquiring more transits. The increase of the aperture size from the current 2.5 m with HST to 15 m with the LU V OIR − A mission concept, along with improvements in instrument performances, would probably not be enough to significantly reduce the number of transits required to detect gaseous spectral lines. Note that in this study we have assumed a photon-limited noise scenario where the total noise "n" is represented by a "white noise" decreasing while acquiring more photons. However, instrument systematics and/or background (astrophysical) noise would be added to the noise that will decrease more slowly and eventually reach a noise floor. (Greene et al. 2016) has estimated a conservative noise floor of 25 ppm and 50 ppm for a 1 σ detection with both JW ST 's NIRSpec Prism and MIRI, although optimistic estimations mention half of these values (Fauchez et al. 2019a). According to Fig. 3 only CO 2 and O 3 could produce relative transit depths higher than those noise floors. However, Origins MISC-T intends to use a technology that allows the noise floor to reach 5 ppm (Meixner et al. 2019).
Synergies between instruments may be crucial in order to combine observations within various wavelength ranges and accumulate transits over an extended period of time. For example, observations with future extremely large telescopes such as the ELT , GM T or T M T using cross-correlation techniques (Snellen et al. 2013) are promising and should be use in conjunction with transit observations from space.

CONCLUSIONS AND PERSPECTIVES
In this work, we have used TRAPPIST-1e, potentially the most promising target for atmospheric characterization of a planet in the HZ of a nearby M-dwarf, as a benchmark to compare transmission spectroscopy performances of future space-based observatories. This study does not aim to investigate the detectability of each gaseous species under various habitable conditions, such as those of Earth through time. Instead, we focus on the most well-known habitable atmospheric composition, that of modern Earth, and compare a variety of instrument capabilities to characterize individual molecular species. Our study shows that, despite the anticipation of tremendous future improvements in terms of aperture size and instrument performance, these factors would not be enough to characterize such planets via transmission spectroscopy. Indeed, most spectral lines from the gaseous species of a modern Earth-like atmosphere produce a relatively small transit depth and clouds drastically reduce their amplitude. Even for the largest aperture size of 15 m for LUVOIR-A, hundreds or thousands of observed transits would be required to detect molecular species at a 3 or 5 σ confidence level.
Only CO 2 and its strongest feature