TMT/MICHI will be a transformative and unique exoplanet science instrument1. The current generation of 10 m class telescopes (e.g. Keck, Subaru, Gemini, VLT) provided the first direct images and spectral characterizations of young Jovian exoplanets at 1-5 μm. Continuous development of instruments on these telescopes is maturing key exoplanet imaging methods and technologies, which MICHI will benefit from at longer wavelengths. While extremely sensitive at TIR, JWST’s limited angular resolution prevents it from directly imaging and characterizing exoplanets at solar system-like scales (~1-25 AU) for all but the very nearest stars. With MICHI, TMT will leverage advances made by observatory partners that (1) open up both new exoplanet discovery space and new phase space for characterization and (2) complement exoplanet science with the possible PSI (Planetary System Imager). We are in close contact with the PSI science team and are aware of the commonality in this science area between PSI and MICHI; we intend to make detailed shared studies at a future time should these WPs be approved.

TIR Imaging & Characterization of Solar Neighborhood Jovians to Earth-Size Exoplanets

TMT/MICHI provides a dramatic increase in angular resolution and orders of magnitude better sensitivity in the TIR (especially at 10 μm) compared to current ground-based telescopes. Current exoplanet imaging systems can probe young (1-100 Myr) Jovian planets, and possibly, in the near future with next generation or upgraded instruments, a few Jovians in reflected light. In both cases, however, direct detection of the many Jovian exoplanets indirectly identified by radial-velocity (RV) methods (r > 1 AU, orbiting older stars) and exo-Earths around Sun-like stars remain out of reach. TMT/PSI expects to image rocky exoplanets in reflected visible/near-IR (NIR) light around the nearest low-mass stars.

In comparison, MICHI will be sensitive to cooler, Gyr-old Jovians orbiting nearby (d < 10 pc) stars, dramatically increasing the number of systems where planets can be detected, providing a substantial overlap with the RV found planet population. TIR characterization using both low- and high-resolution spectroscopy of Jovian exoplanets (including free-floating objects or similar mass) will probe non-equilibrium chemistry, with ammonia expected to play a central role at 10 μm, and atmospheric metallicity which probes the formation process. For old and cold RV exoplanets in nearby systems, determining the system’s inclination by direct imaging will yield the planet’s dynamical mass.

Most excitingly, MICHI’s spatial resolution could enable imaging of the first Earth twin in the habitable zone of nearby Sun-like stars. While some Earth-size exoplanets have been found using transit and radial velocity techniques, no Earth-like exoplanets have yet been discovered. Planned NASA missions will not afford this capability; JWST enables limited transit spectroscopy characterization of fortuitously aligned nearby transiting M dwarf systems and WFIRST CGI will likely not reach Earth-mass sensitivity. Far-future (2040+) proposed NASA missions (i.e. HabEx, LUVOIR) rely on detecting exo-Earths in visible/NIR reflected light and hence MICHI’s 10 μm coverage affords highly complementary capabilities. TIR imaging is sensitive to planetary emission, where the exoplanet-to-star contrast at 10 μm is 103 times more favorable (contrast ratio (CR) requirement ~107) than at visible light (CR requirement ~1010) for a Sun-like star. Compared to 8 m-class telescopes, TMT is ~200 times more efficient at reaching the required background-limited contrast at 10 μm, requiring < 1 hour for a detection of the closest systems instead of 100’s of hours using an 8 m. The habitable zone is resolved at 10 μm at distance of up to 5 pc for Sun-like stars. Primary northern hemisphere targets include Tau Ceti (multiple candidate rocky habitable zone planets), Epsilon Eridani, and 61 Cygni A/B.

TMT’s sensitivity enables spectral characterization of the nearest/brightest Earth-like exoplanets (Fig. 1). Low 10 μm spectral resolution characterization (calibrated to the star’s spectrum) could enable biomarker detection, such as O3, H2O, O2, CH4 and CO2, as well as estimates of the planet’s surface temperature via blackbody fitting. The transformative science potential enabled by MICHI at high angular resolution and at high contrast far exceed current 8 m class telescopes, and that possible with the 6.5 m JWST.

Figure 1. TMT/MICHI 10 μm imaging and characterization capabilities. Left: Simulated image of Alpha Centauri A with MICHI. A coronagraph and advanced post-processing techniques are used to remove the central star (NB. although this object is difficult to observe from MKO, it is used as a demonstrator of the concept). Right: Earth N-band spectrum showing various biomarkers in the 10 μm window (adapted from Hanel et al. 1972).

High Spectral Resolution Spectroscopy of Exoplanets; Detailed Characterization

A TIR high-dispersion spectrometer (HDS) affords high-resolution spectroscopy of exoplanets and offers unique insights into the atmospheric dynamics, composition, and chemistry of these objects; and in a few cases, into their spin and orbital angular momenta, and aspects of their global weather patterns.

MICHI will become TMT’s workhorse instrument for estimating atmospheric metallicity through O and C measurements and the ratio of these two elements (C/O). The L band accesses strong h3O absorption, and this wavelength range has provided the bulk of the high-resolution detections of this molecule in short-period exoplanets (Birkby et al. 2013, 2017, Lockwood et al. 2014, Piskorz et al. 2016, 2017). Furthermore, in Solar-abundance atmospheres h3O is expected to be the most abundant species after h3 (Heng et al. 2017), suggesting that L band measurements of h3O will form a large part of MICHI’s exoplanet work. The L band also hosts the strongest CH4 feature easily accessible from the ground, a molecule that is yet to be detected in any short-period transiting planet. When L band h3O and CH4 measurements are combined with the strong CO and CO2 absorption lines in the M band, MICHI will be poised to measure the abundances of all major O- and C-bearing molecules; if higher-order hydrocarbons such as C2h3 are present, MICHI will also be able to measure their abundance (de Kok et al. 2014), constraining models of hydrocarbon “soot” haze formation in exoplanet atmosphere (Morley et al. 2015, Kawashima & Ikoma 2017). Both metallicity and C/O provide strong constraints on the formation method and location of the observed planets (e.g., Mordasini et al. 2016).

One of the most exciting prospects with TMT/MICHI is advancing photon constrained observations of the rotation of exoplanets (Fig. 2), pioneered on the VLT. The β Pic b planet rotation rate of 8.1±1 hours was determined by observations of CO emission lines. Whilst this type of measurement was repeated for ~10 objects, TMT/MICHI will extend this to ~several 10s of objects, enabling a detailed sample of the rotational properties of exoplanets and how they obtained their angular momentum during formation. For numerous highly irradiated planets, detailed line profiles will probe upper atmosphere wind speeds and circulation patterns, advancing high spectral resolution results that previously measured the day-to-night winds speeds on hot Jupiters HD209458 b (-2 1 kms-1; Snellen et al. 2010) and HD 189733 b (-1.7  1.2 kms-1; Brogi et al. 2016). Such properties are uniquely determined at high-spectral resolution.

Figure 2. First measurement of an exoplanet rotation (left) showing the rotationally-broadened cross-correlation profile of the giant exoplanet β Pic b (Snellen et al. 2014). The dotted line shows the expected profile without planetary rotation. Examples (right) of the predicted profile of rotation (green line) and winds (purple line) on the cross-correlation profile of exoplanet atmospheres (Kempton & Rauscher 2012; Showman et al. 2013).

More broadly, MICHI high spectral resolution observations will accomplish groundbreaking science on multiple fronts, exemplified by these selected highlights:

  • Detect atmospheres of the nearest neighboring rocky worlds via high-dispersion spectroscopy and medium-resolution IFU observations (Snellen et al. 2015, Lovis et al. 2017, Hoeijmakers et al. 2018)
  • Detect giant storms on the surfaces of widely-separated giant planets via Doppler imaging (Crossfield et al. 2014);
  • Atmospheric chemistry of exoplanets, from (super-)Earths to hot Jupiters, and production of 1D photochemistry maps as a function of longitude. This will lead to the first measurements of low-abundance species, e.g. isotopic ratios connected to evolutionary history (e.g. HDO, 13CO);
  • Measure true masses of non-transiting (super-)Earths around nearby bright stars.

TIR Characterization of Protoplanetary Disks

The prospects for the formation of planetary systems and the development of life are pre-determined in protoplanetary disks. Whether planets have the necessary chemical ingredients to generate and nurture life depends on the chemical and physical disk structures from which they form. Time-dependent theoretical models of the physical and chemical evolution of the disk (e.g. Aikawa et al. 2002) have been tested through observations of disk chemistry and structure (Fig. 3). Particularly the snow lines of water and other volatile molecules (e.g. Blevins et al. 2016, Notsu et al. 2017) play a vital role in planetary formation, and subsequently the chances for life. As noted above, such observations are optimally performed at TIR wavelengths due to the greatly improved CR. TMT/MICHI will be capable of spatially and spectrally resolving the snow lines of these species by making use of gas emission lines and ice absorption features (e.g. Honda et al. 2016). Disk thermal structure evolution can also be measured by tracing the spatial distribution of crystalline silicate material in the disk. TMT/MICHI affords a unique opportunity to uncover both mineralogical evolution and transport of solid material within a disk by spatially-resolved MIR spectroscopy (e.g. Okamoto et al. 2004, and Fig. 3). For the same reasons as those described in §2, existing 8m’s and JWST do not afford the combination of high sensitivity and spatial/spectral resolution, leaving a large and crucial parameter space to be filled by TMT/MICHI.

TMT/MICHI will also be a powerful tool to obtain emission/absorption line profiles of a number of molecular (e.g. CO, OH, h3O, h3, CH4, C2h3), atomic (e.g. H lines), and ionic (e.g. [Ne II], [Fe II]) lines in the TIR. Molecular lines are excellent tracers of temperature and abundance, and their spatial location obtained by MICHI will help our understanding of disk chemical/physical structures and evolution (e.g. Salyk et al. 2008; Banzatti et al. 2017). CH4 and C2h3 lines are useful to trace the distribution of the C/O ratio, are only available in the TIR. Atomic and ionic lines in the TIR provide powerful diagnostics of mass inflows, outflows and physical processes (e.g. Neufeld & Hollenbach 1994, Pascucci et al 2011, Edwards et al. 2013, Watson et al. 2016). As multiple processes can produce these lines (e.g. Herczeg et al. 2012) it is critical to have both high spatial (< 0.1”) and spectral (R > 60,000) resolution observations to disentangle contributions from infall, accretion, and outflows to assess mass flow rates. JWST/MIRI will not resolve these lines spectrally nor spatially, but TMT/MICHI will. This makes it possible for the first time to audit mass flow rates in protostars & disks, providing fundamental insights into the physical processes that assemble protoplanetary disks.

Growing protoplanets are also important targets to fill in the evolutional gap between the disk and planets. MICHI will be able to detect these objects in two ways with IFU spectral images. First, spiral waves created by accreting planets should be detectable in CO and other molecules (Regaly et al. 2015). Second, a growing giant planet may have a circumplanetary/proto-lunar disk, with a chemical and physical structure analogous to the much larger parent circumstellar disk. Evidence for a planet has already been seen in CO spectroastrometry in a ~10 AU planet around HD 100546 (Brittain et al. 2014); TMT/MICHI will push these detections to within an AU, where planets are more prevalent.

Figure 3. Cartoon of the chemical structure, including the h3O snow line and planet formation in a typical protoplanetary disk around a solar-mass star (from Henning & Semenov 2013).

Unravelling AGN & SMBH Astrophysics

It is observationally well known that supermassive black holes (SMBHs, ≥ 106 M) are ubiquitous in galaxy centers, and the masses of SMBHs and galaxy stellar components are tightly correlated. This suggests that SMBHs and galaxies co-evolved and that SMBHs play an important role in galaxy formation. AGN (active galactic nuclei) are the primary objects in these studies as they are accreting surrounding gas and exchanging gravitational potential energy to produce copious amounts of heat and thermal radiation. The empirically-based Unified Model posits that a central accreting SMBH is surrounded by a geometrically and optically thick torus shaped dusty structure that obscures the central engine from some lines of sight. Although the torus can naturally explain various observational characteristics of AGN, little is constrained about its physical/morphological properties due to its compactness (< 10-20 pc, < 0.1-0.2" at z > 0.005), nor crucially its interaction with the host galaxy, creation, nor how it is sustained; this is despite the torus being the cornerstone of the theory. The torus re-radiates the central engine’s energy, peaking at > 30 μm (e.g. Fuller et al. 2017), but detailed MIR properties remain elusive. Recent MIR interferometry (i.e. Hönig et al. 2013) suggests the presence of ‘polar dust’ at scales of < 20 mas, possibly related to feedback mechanisms, further complicating the picture. High spatial resolution observations at TIR wavelengths are the most powerful way to unravel the properties of the enigmatic AGN torus and discriminate it from host galaxy contaminating radiation. Finally, the source of excitation of PAH emission in galaxy centers remains controversial. There are now many objects for which PAH emission is observed within a few 10’s pc from the AGN (i.e. Hönig et al. 2010; Esquej et al. 2014) and hence it is unclear if stellar or AGN photoexcitation is dominantly responsible for the PAH emission. TIR long-slit or IFU observations permit simultaneous probes of the torus, polar dust, and nuclear PAH emission, uniquely available through the combination of high spatial-resolution and high sensitivity that TMT/MICHI offers, unavailable at the spatial resolution of the JWST.

Torus dust emission at TIR wavelengths will be spatially imaged for the first time using TMT/MICHI (Fig. 4, left) for up to ~10 sources. By dramatically reducing host galaxy contamination, high fidelity spectroscopy (Fig 4., right) will be available for ~100 sources, providing a statistically significant sample. This will deliver fundamental information about the size, inclination angle, and dust mass in the torus when combined with existing and under-development models. Such observations will break model degeneracies that pervade current work and allow distinguishing between the plethora of torus models and finally permit a true characterization of the genesis and maintenance of the torus. Observationally it is clear that the torus is geometrically thick but simple torus models gravitationally collapse it to a disk. Possible ‘inflating’ mechanisms are modeled to be AGN radiation pressure or supernovae explosions associated with nuclear star formation inside the torus and/or the surrounding circumnuclear disk to ~10-20 pc. Others have suggested that the torus genesis and inflation is due to a magneto-hydrodynamic wind (Lopez-Rodriguez et al. 2015, Chan & Krolik 2017). If so, AGN polarimetry could significantly aid in estimating the magnetic field direction and properties. However, although much NIR AGN polarimetry data exists, only a handful have TIR polarimetric data (Lopez-Rodriguez et al. 2018), primarily due to the photon hungry nature of polarimetry and requirement for high spatial-resolution. TMT/MICHI will greatly alleviate both limitations and thus permit such observations. TIR IFU spectroscopy and TMT’s spatial and sensitivity gains will allow details of the feedback of material between the host galaxy and AGN central engine to be explored with unprecedented detail, linking the host galaxy to the AGN dominated regions. Detection of nuclear outflows will be compelling evidence of the torus origin by AGN radiation pressure. MICHI will examine the possible link to the large scale (~1-10 kpc) ionized outflows thought to be sweeping material away in the host galaxies. [SIV]10.5 μm and [NeII]12.8μm are key lines due to their high ionization potential. IFU observations with TMT/MICHI in the relatively low extinction wavebands of L and N bands are ideal to address this dichotomy, again impossible at JWST spatial resolution.

Figure 4. Clumpy model simulation of the torus in NGC1068 (left) at ‘infinite’ resolution (1st row), the TMT PSF (2nd row), model TMT/MICHI observations (3rd row), and Richardson-Lucy deconvolution (30 iterations, 4th row) showing extended emission along the torus ‘funnel’ at 3-5μm. Torus model dust distribution is shown (center). The criticality of isolating torus emission from diffuse galaxy contamination is shown for NGC5135 (right, Diaz-Santos et al. 2010). This exemplifies host galaxy contamination in diffraction-limited spectra from Spitzer (0.8m, solid black, galaxy dominated) vs. Gemini (8 m, red = AGN dominated, blue = AGN + nucleus). A similar situation will be repeated in the JWST/TMT era, where JWST data are significantly contaminated, whereas TMT’s superior spatial resolution will minimize host galaxy contamination. Definitively resolving the torus (10 sources) and minimizing contamination (~100 sources) is essential to advance knowledge of the torus.

Supplemental Science Cases

Our driving science cases summarized in §2-4 provides unique input to the instrument requirements. These requirements permit numerous supplemental projects to be pursued from a variety of fields. Below are two (of the many) exemplary cases our science team has provided.

The Interstellar Dust Budget Over Cosmic Time

Heavy elements synthesized through stellar evolution have chemically enriched the universe. Massive stars are dominantly the first stellar sources that provided metals and dust in the interstellar medium of galaxies at high-z and in the early universe. Supernovae (SNe) are regarded as excellent candidates for the dust budget in high-z galaxies, but the amount of dust supplied by SNe remains poorly constrained. The paucity of our knowledge of the properties and the geometry of the pre-existing circumstellar medium prevents us from extracting the IR emission from SNe dust and accurately estimating the density and temperature environment of SNe dust. Although JWST will greatly increase the number of observed dusty SNe, the properties of SNe dust will remain poorly determined without knowledge of the circumstellar environment of core-collapse SNe progenitors (e.g., Wolf-Rayet stars and LBVs). TMT/MICHI will resolve circumstellar dust structures around Galactic dusty Wolf-Rayet (WR) stars and LBVs and play a crucial role in understanding how massive stars supply dust into the interstellar medium, particularly in the early universe. Additionally, some recurrent novae (e.g. T Coronae Boreallis, V745 Sco, RS Oph and V445 Pup), whose white dwarf mass is close to the Chandraskhar mass, are regarded as potential progenitors of Type-Ia SNe. With TMT/MICHI, those targets afford an invaluable opportunity to probe circumstellar environment present around Type Ia SNe progenitors.

Galactic novae afford a unique chance to explore dust formation processes in ejecta in ‘human-scale’ observing timescales. Although novae produce a relatively minor contribution of dust in the interstellar medium, several studies suggest a link between novae dust and pre-solar grains in meteorites. With TMT, expanding ejecta of nearby (< a few kpc) dusty nova are resolvable in the TIR (e.g. Chesneau et al. 2012; Sakon et al. 2016), and combined with LSST optical light curves offer unique opportunities to observe the dust condensation process and possibly to understand the origin of minerals and organics. We note the optical identification of the GW 2017 event and the detection of very heavy element production. As shown by Telesco et al (2015), TIR at the highest spatial resolution can estimate the mass and element production of SNe. TMT/MICHI will afford the chance to follow-up future novae and GW events to trace detailed element production.

Solar System Observations

Solar System science will greatly benefit from MICHI on the TMT, as it has with TEXES (e.g. Lacy et al. 2002), mounted on both Gemini North (Greathouse et al. 2011; Tsang et al. 2016) and NASA IRTF, (e.g. Encrenaz et al. 2016; Fletcher et al. 2016). With much higher spectral and spatial resolution than the JWST, MICHI can map temperatures and atmospheric trace species on Mars, Venus, gas giants and their moons, tracking spatial and temporal evolution, to expand our understanding of these diverse worlds, and crucially support data-poor studies of exoplanets. Additionally, with such high spectral resolution TIR capabilities currently unachievable in space, these observations will serve as mission support for NASA and ESA missions (e.g. the Europa Clipper, JUICE (targeted to Ganymede) or Psyche (asteroid belt)).

In 2017 the first known interstellar visitor to the Solar System was identified. “Oumuamua” was observed by numerous telescopes, where the light curve indicated a highgly elongated morphology, and the SED implies the presence of surface organic compounds (i.e. Meech et al. 2017). In the LSST era detection of such objects is likely to become common, optimally exploited in the TIR where objects directly radiate. MICHI/TMT will be an outstanding facility to observe their SED, spectra, and light curves of these objects.


1. The 2018 papers submitted in response to the call from the National Academy of Sciences Exoplanet Science Strategy group by Currie et al. ( and Myers et al. ( explore the 30m "TIR advantage".
2. The improved Strehl ratio that MICHI (~80%) will enjoy at 10μm compared to the 8m's (≤30%) will actually make this differential ratio even greater (with the exception of the adaptive secondary mirror (ASM) fed mid-IR (MIR) imager NOMIC on the LBT).