香京julia种子在线播放

One of the most exciting scientific endeavors of this century is the quest to find a habitable, Earth-like exoplanet. It would be most exciting to find a habitable exoplanet that is close enough to the Earth that one could imagine sending a probe to explore it, or maybe one day sending human explorers. Even more exciting would be the discovery of a planet that has oxygen in its atmosphere, which would make it potentially habitable for us and also likely to support life of its own. Direct imaging of at least 25 habitable exoplanets and finding biosignatures such as oxygen and methane in their atmospheres is in fact the primary science driver for the Habitable Worlds Observatory (HWO), a conceptual future NASA flagship mission that is thought to require a significant technology maturation process (National Academies of Sciences, Engineering, and Medicine, 2021) before it can be selected.

Most current and planned exoplanet discovery missions use the transit method, which relies on the small decrease in starlight observed when an exoplanet passes in front of a star, to find exoplanets. TESS (Ricker et al., 2014) is the current NASA transit mission, following the highly successful Kepler mission (Borucki et al., 2003), which discovered 2,806 of the 5,638 currently known exoplanets according to the NASA Exoplanet Archive (Akeson et al., 2013; NASA Exoplanet Archive, 2024). Both ESA (PLATO, Rauer et al., 2014) and China (Earth 2.0, Ye, 2022) are building new missions to detect Earth-like planets using the transit method and aim to launch in 2026 and 2027 (Zang et al., 2024), respectively. However, only a small fraction of exoplanet orbits are edge-on as observed from the Solar System, so a large volume of the local Galactic neighborhood must be searched to have a high probability of finding an Earth-like planet.

Instruments such as ESPRESSO (Pepe et al., 2021), EXPRES (Jurgenson et al., 2016), and NEID (Schwab et al., 2016) could reach a radial velocity precision of 10 cm/s (compare with the Sun’s speed of 9 cm/s around the Earth-Sun center of mass), if difficulties with velocity variation of the material on the surface of the host star can be overcome. This raises the possibility that habitable exoplanets with a wider range of orbital inclinations could be identified. Even so, the radial velocity method misses planets that orbit face-on.

The only way to find all of the local Earth-like exoplanets is to resolve exoplanets from their host stars, for example, by direct imaging. Resolving exoplanets from their host stars also vastly improves the signal-to-noise for spectroscopy of exoplanet atmospheres. However, to detect Earth, which is 1 AU from the Sun, from a distance of 10 pc requires an angular resolution of 0.1 arcseconds (0.1″). The Sun is more than ten billion times brighter than the Earth in visible (reflected) light, but the luminosity contrast can be reduced by observing in the infrared (optimally ∼ 10 μ m) where the Earth peaks in emission and the Sun is only a million times brighter than the Earth. To separate two objects that are 0.1″ apart at a wavelength of 10 μ m, the diffraction limit ( θ ∼ λ / D ) requires optics with a physical length scale of at least 21 m. Even then, coronagraphic techniques are required to reduce the light from the host star enough that the exoplanet can be detected. The National Academy of Sciences (NAS) Exoplanet Science Strategy (National Academies of Sciences, Engineering, and Medicine, 2018) document recognized the desire to resolve Earth-like exoplanets from their host stars in the infrared, where habitable planets are brightest. However, they concluded that: “Constructing a diffraction-limited 20 m class space telescope that would be cold enough to operate at 10 microns is considered exceptionally difficult.”

Two recent proposals for NASA Great Observatories that could find habitable exoplanets were the Habitable Exoplanet Observatory (HabEx, Gaudi et al., 2020) and the Large UV/Optical/IR Surveyor (LUVOIR, The LUVOIR Team, 2019). HabEx combined a 4 m telescope with a separate starshade flying ∼ 1 0 5 km away. LUVOIR explored the deployment of an 8 m or 15 m mirror paired with a coronagraph with a contrast of order 1 0 − 10 . Neither design planned to detect the infrared ( ∼ 10 μ m) emission spectrum from an Earth-like planet, so high resolution and extreme coronagraphy in visible/ultraviolet were required. Because of the difficulty in developing and launching the James Webb space telescope (JWST, Gardner et al., 2006), the NAS review recommended that while “Pathways to Habitable Worlds” was a top priority, it would require significant technology maturation (National Academies of Sciences, Engineering, and Medicine, 2021) before a final design could be chosen. A scaled-back Great Observatory consisting of a “ ∼ 6 m diameter Infrared/Optical/Ultraviolet space telescope with high-contrast imaging and spectroscopy,” and that combined elements of HabEx and LUVOIR, is currently under consideration; NASA has established the HWO Technology Maturation Project Office (HTMPO) to address critical technical hurdles to the success of the HWO.

Our calculations suggest that local exoplanets could be identified and characterized in the infrared using a telescope that has a similar collecting area to JWST (and the proposed HWO), and using Achromatic Interfero Coronagraph (AIC, Rabbia et al., 2007) technology that is already mature, as long as the primary mirror is rectangular rather than roughly circular. The rectangular mirror should be ∼ 20 m long in one direction, which makes it possible to resolve the Earth and the Sun (1 AU apart) from a distance of 10 pc, as long as the long axis of the telescope is aligned with the line from the Sun to the Earth. To detect exoplanets at all orientations, the telescope must take (at least) two images of the sky; one should be taken with the primary mirror rotated 90 ° from the other. High-aspect-ratio rectangular mirrors require more observation time, but a much smaller telescope mirror to reach a given resolution.

A less ambitious mission based on an infrared rectangular mirror concept could provide a much more efficient pathway to identifying habitable planets that have promising aspects, in anticipation of more complex missions that have enhanced characterization capabilities.

We describe here our suggestion that current technology could achieve the HWO primary goals if a rectangular telescope design is adopted. This discovery has arisen from our work in understanding the properties of Dittoscopes (Ditto, 2003), which collect light with a large Primary Objective Grating (POG); the grating is then “observed” at grazing exodus with a focusing element akin to a conventional telescope. The grazing exodus is required for the secondary telescope to collect light from the entire surface of the grating. As shown in our previous work (Swordy et al., 2023), the rate of photon collection from this system is determined primarily by the secondary telescope, but the diffraction limit (in one direction) is set by the length of the grating. The fact that Dittoscopes could achieve high angular resolution inspired us to use them for exoplanet detection and spectroscopy of their atmospheres. We attempted to design a large space telescope for direct detection of habitable exoplanets using a primary objective diffraction grating design.

In Swordy et al. (2025), we outline the concept of the Dispersion Leverage Coronagraph (DLC) and the application for which DLC was originally designed: the Diffractive Interfero Coronagraph Exoplanet Resolver (DICER), a notional space-based Dittoscope designed to detect Earth-like planets orbiting Sun-like stars within 10 parsecs from Earth and identify ozone ( O 3 ) in their atmosphere. DICER collects infrared ( ∼ 10 μ m) light with two large primary objective gratings that are each 10 m by 1 m in size, and uses DLC interferometry to extinguish the light from the exoplanet’s host star. We showed that this optical design is capable of resolving Earth-like exoplanets from Sun-like stars. However, it has difficulty resolving the exoplanet from the background zodiacal light in our solar system and also any zodiacal light around the host star. This problem was partially solved by adding a technically challenging, high resolution spectrograph to the optical train. With this component, we showed that DICER could plausibly find and characterize ∼ 4 nearby, habitable exoplanets around Sun-like stars in a 7 year mission. This represents about 30% of the habitable exoplanets within 8 pc that were in our simulation.

In this paper, we show that if we instead collect the light with mirrors that are the same size and shape as the primary objective gratings that were imagined for DICER, a larger number of exoplanets could be discovered. In addition, the optical design is much less complicated because we would not require a high resolution spectrograph. This suggests that the important innovation for finding exoplanets was not the grating, but the shape of the light-collecting optical element.

First, we describe the proposed rectangular infrared telescope and coronagraph. We then identify stars closer than 10 pc to the Earth, and run a simulation to generate mock habitable exoplanets around those stars in 1,000 universes. Considering the properties of the stars and the simulated exoplanets, we estimate how many of these would be discovered with the rectangular infrared telescope, and how much exposure time would be required. We then determine the amount of time required to detect the 9.6 μ m absorption band of ozone; the calculation assumes we need a five sigma detection of the dip in the spectrum.

We show that using technology that has already been developed for JWST, we could plausibly build an infrared telescope that could achieve the HWO goal of detecting 25 habitable, Earth-like exoplanets and determine whether there is ozone in their atmospheres within a 3.5 year mission. There are many assumptions in our calculations, and many design choices that could increase or decrease this yield and the time-to-completion.

Our goal is a telescope design that could detect all of the habitable, Earth-like planets around Sun-like stars that are closer than 10 pc. Because planets with habitable temperatures have a peak blackbody emission wavelength near 10 μ m, we start with a telescope design that is optimized in the infrared. At 10 μ m, the diffraction limit requires that at least one dimension of the primary collector be about 20 m across to achieve 0.1″ resolution; and that dimension must be aligned with the direction in the sky from the host star to the exoplanet. Because a round mirror of diameter 20 m is not feasible, we suggest making a rectangular mirror that has a long dimension of 20 m.

If we do not know the position angle of a planet around its host star (the normal situation when attempting to discover an exoplanet), then for optimal resolution the long axis of the mirror needs to be rotated, while staring straight at the host star, to try all possible exoplanet position angles. However, most of the exoplanets can be identified from two images taken with the primary mirror rotated by 90 ° between exposures. The technique for finding exoplanets and the implications for the observation time will be covered in the next section.

In this paper, we imagine the capabilities of a 20 m × 1 m version of JWST, equipped with a modern coronagraph that is optimized for discovery of nearby exoplanets and a mid-infrared instrument with a fraction of the capabilities of JWST’s Mid-Infrared Instrument (MIRI, Rieke et al., 2015). The mirror would most likely be made in segments, as shown in Figure 1. For calculation purposes, we imagine an optical design that is a 1 m-wide slice of JWST that has been scaled up by a factor of 20/6.5. In this system, the secondary mirror is ∼ 2.3 m × 1 m, and is placed to the side of the primary mirror so that it does not obscure the primary mirror. The expectation is that, like JWST, this telescope would be deployed at L2, and the mirrors would be passively cooled with the aid of a sunshield. Also like JWST, the primary mirror segments would be actively aligned to achieve an optimal point spread function in the focal plane.

The sunshield concept presented in Figure 1 extends 2 m past the ends of the primary mirror, and curls behind the secondary support and up to the secondary mirror. When pointed within 25 ° of the antisolar point (away from the Sun), the portion of the sunshade behind the mirror will shield all of the optical elements at all rotation angles of the primary mirror. When pointed perpendicular to the plane of the ecliptic (the worst case), the triangular portion will shield the optical elements for 120 ° of primary mirror position angle, which would easily allow the spacecraft to obtain two images separated by a 90 ° , as required for our baseline exoplanet search. For launch, we imagine that two 10 m portions of the primary mirror are folded with their reflective surfaces towards each other, the secondary mirror support is compressed to about the same length of 10 m, and the triangular portion of the sunshield is folded up. Note that both the primary mirror and the sunshield would require far less folding than JWST to fit in a launch vehicle, and require far fewer steps to unfold after launch. We estimate the size of the folded spacecraft as 11 m long (10 m of mirrors plus the thickness of the secondary mirror, the thickness of the sunshield, and possibly additional room for power, communications, and instrument packages), 2.5 m wide (slightly larger than the width of the secondary mirror), and 2.5 m deep (the thickness of two primary mirror segments folded together, plus the sunshield behind them, and imagining the power, communications and instrument packages will also need some space). This is of course a very rough estimate.

Although there are many coronagraph designs that work with conventional telescopes, we have assumed a coronagraph of type AIC. With AIC, the focal plane contains two images of the exoplanet, separated by twice the angular separation of the exoplanet and the host star. The PSF of each exoplanet image has an elongated profile determined by the diffraction limit of the rectangular mirror: 0.1″ × 2″. AIC is capable of nulling over a 25% bandwidth (Gappinger et al., 2009), which covers the 1.8 μ m range of the MIRI F1000W filter we are assuming for our observations, but it can be tuned to other wavelength ranges. The coronagraph would need both active cooling and active correction of the path length. An optical schematic is given in Figure 2.

We are also assuming that the detectors are state-of-the-art single photon detectors: transition edge sensors (TES, Nagler et al., 2018; Höpker et al., 2019; Nagler et al., 2021), superconducting nanowire single-photon detectors (SNSPD, Wollman et al., 2021; Verma et al., 2021; Lita et al., 2022), or mid-infrared kinetic inductance detectors (MKID, Ras et al., 2024). With these detectors, all photons are collected without the introduction of noise; the only noise comes from zodiacal light in the Solar System and the exoplanet host system.

The development of an optimal survey strategy and detailed throughput information for each observation is beyond the scope of this work. For example, it would require determining the best strategy to observe each known star so that the local zodiacal light is a minimum. Here, we present only a simple calculation to elucidate the capabilities of this system. Note that this study closely follows the feasibility simulation done by Swordy et al. (2025) for DICER, but we repeat the steps here for clarity.

For comparison, we evaluated the exoplanet-finding capability of a square telescope with an equivalent collecting area. This telescope has a “diameter” of 4.47 m. The results for the square telescope will be similar to the typical circular footprint for a telescope collecting area, and have the advantage that we can run exactly the same code with only two numbers (the length and width of the telescope) changed.

The PSF is spread over the same number of pixels either way; one resolution element in the rectangular case is λ 2 / ( L W ) and in the square case is λ 2 / ( L W ) 2 . Therefore, the detection of isolated point sources is about the same in either case. If the goal is to measure the detailed shape of an extended object, the rectangular telescope would require multiple pointings at different rotation angles, and a much more complex data analysis pipeline. Where the rectangular telescope design really shines is in the detection of point sources that are closer to each other than can be detected with the equivalent circular telescope. Note, however, that engineering capabilities such as mirror alignment, pointing and jitter, residual optical path length, etc., will need to be at the level (at least in one dimension) that is required for a circular telescope of diameter L .

With a diameter of 4.47 m, the diffraction limit is 0.46″, which means we will not be able to resolve planets that are closer than 0.23″ to the host star. This restriction makes it impossible to find most of the habitable planets we are trying to identify. For the 46 stars with 45,652 simulated exoplanets, we found 2,563 exoplanets with an exposure time of 2 days, 2,795 exoplanets with an exposure time of 4 days, 2,903 exoplanets with an exposure time of 6 days, 3,044 exoplanets with an exposure time of 8 days, and 3,144 with an exposure time of 10 days. Increasing the exposure time will not allow us to reach our goal for finding Earth-like habitable exoplanets, because most of the exoplanets we wish to find cannot be resolved from the host star with the square design and are therefore not possible to detect.

The left panel of Figure 9 shows the exoplanet temperatures and radii that we can find with the equivalent area square telescope. Note that we are preferentially losing the lower mass and warmer exoplanets, including those that are most similar to the Earth. Only 5% (14 of 278) of the simulated exoplanets with radii, masses and temperatures similar to the Earth ( 0.9 < R p / R ⊕ < 1.1 , 0.9 < M p / M ⊕ < 1.1 , and 250 K < T p < 300 K ) were detected. The right panel of Figure 9 shows the distribution of angular separation between the host star and the detected exoplanets, highlighting the much larger inner working angle of this telescope footprint.

There is nothing about the idea of making a rectangular mirror, or the AIC coronagraph, that precludes its use at any wavelength. Any current design for HWO or any other flagship mission could be adapted to work with a rectangular primary optical element. For example, if the 20 m × 1 m mirror proposed here was used in the visible at 500 nm, the images could in principle be 0.005″ × 0.05″. This would allow for the detection of Earth-like exoplanets to distances of about 200 pc (twenty times farther, because the wavelength of the light is 20 times shorter). We note, however, that there would be considerably more stellar leakage, and the constraints on pointing and stability would make construction much more difficult.

We also point out that the proposed infrared survey for exoplanets favors the detection of habitable exoplanets, with only modest coronagraph performance. Planets that are much cooler than the Earth will be very faint at 10 μ m, and therefore much more difficult to detect. Planets that are much warmer than the Earth will be closer to the host stars, and therefore less likely to be resolved. The extremely faint, habitable exoplanets are detected at their most luminous wavelength.

In general the design of a rectangular telescope can be adapted to favor discovery of any particular exoplanet type, at any distance. It can also be designed to enable detailed followup for discovered exoplanets, in which case one would be able to orient the long axis of the mirror in the direction of the current position angle of the exoplanet for maximum resolution. Resolving an exoplanet from its host star is the first step to any exoplanet observation, and for fixed mirror area the highest resolution can be obtained with a rectangular shape. It is instructive to compare the rectangular mirror infrared telescope with other concepts that are being considered for habitable exoplanet discovery.

The Large Interferometer For Exoplanets (LIFE, Quanz et al., 2022; Dannert et al., 2022) collaboration is exploring the use of an array of smaller space telescopes that fly in formation to very high precision. LIFE benefits from two past interferometric, planet-finding mission concepts, ESA’s Darwin (Cockell et al., 2009) and the NASA’s Terrestrial Planet Finder Interferometer (TPF-I, Beichman et al., 1999), that were canceled due to limitations of our technology and knowledge of exoplanets. Conceptually, LIFE includes four collector spacecraft that formation-fly in a rectangular array, sending light to a beam combiner spacecraft in the center. The whole array rotates to modulate the signal from an exoplanet. LIFE uses a Bracewell interferometer that is enormously expensive in terms of conops; changing the boresight direction is prohibitively expensive for use as a survey instrument. Because of this, it is better utilized for assaying promising planets that have already been detected. Our concept is more feasible because the fixed 20 m long mirror alleviates the need for formation flying, but it does not allow for baselines in the 25–250 m range imagined by LIFE.

The LUVOIR Decadal Survey Mission Concept Study (The LUVOIR Team, 2019) imagined finding and characterizing exoplanets with an 8–15 m space telescope with a coronagraph that is capable of extremely high contrast nulling. The HabEx Decadal Survey Mission Concept Study (Gaudi et al., 2020) also aimed to study exoplanets, but with a 4 m diameter telescope and a 52 m starshade that flew 76,600 km away from the telescope to enable a 0.07″ inner working angle coronagraph. These ideas are the basis of the technology development for HWO.

The Chinese Academy of Sciences (CAS) has funded a concept study for the Tianlin mission (Wang et al., 2023), a ∼ 6 m UV/Opt/NIR space telescope that is primarily designed to discover and measure biosignatures for nearby G/K stars with the direct imaging method. They aim to start operation of this 5+ year survey within the next 10–15 years. They imagine looking for water, O 3 , O 2 , CH 4 and chlorophyll in the planetary atmospheres. For example, a 6 m telescope fitted with a coronagraph that can suppress starlight by a factor of 1 0 9 can detect water vapor in a nearby Earth-like planet orbiting a Sun-like star with about 40 days of exposure time.

The LUVOIR, HabEx, Tianlin and LIFE concepts grew out of the strong need for an exoplanet observatory with a large baseline and an efficient coronagraph with small inner working angles. The rectangular mirror design proposed in this paper would be easier to launch into space and requires a dramatically lower contrast coronagraph than LUVOIR or Tianlin. It does not require a starshade to fly thousands of miles to point from one host star to another like HabEx, and does not require formation flying like LIFE. It has somewhat more modest goals in that it targets a smaller list of nearby exoplanets and as proposed here does not really deliver spectra. However, discovering habitable planets around the closest stars to Earth arguably promises the highest reward. Our proposed design might be a simpler way to identify interesting habitable planets for later follow-up with HWO or LIFE.

The rectangular mirror telescope is not without its own engineering challenges. One such challenge that requires further study is the structural stability of the rectangular primary mirror. Thermal stability will also need to be considered. While we did not specifically address structural and thermal stability of the optics, they are design concerns because vibration in the mirrors and their support structure will cause the mirror to bend and contribute to jitter (Hyde et al., 2004) and optical thermal stability is required to control distortion. Infrared observing in particular requires a high degree of temperature control to reduce thermal backgrounds. However, it seems plausible that careful design, informed by structural models and confirmed by environmental testing (Kimble et al., 2018), will ameliorate any concern about launch loads or geometric changes in the primary mirror due to thermal gradients (Arenberg et al., 2006). We note the extensive modeling and design effort conducted by NASA to assure that JWST’s 18 hexagonal beryllium mirrors, mounted on individual hexapods, were capable of maintaining their shape in the presence of launch loads and, during operations, external thermal variability. The design we propose for the primary grating is well within NASA’s high-fidelity modeling and environmental test (cryovac and vibration) capabilities (Feinberg et al., 2024).

Although we arrived at the idea of a high aspect ratio mirror independently, we point out that we were not the first to suggest this. The Terrestrial Planet Finder (TPF) project considered two space mission design concepts: an infrared astronomical interferometer (TPF-I; Lawson et al., 2006), which imagined several small telescopes flying in formation, and a visible light coronagraph (TPF-C; Traub et al., 2006), which studied a design with a large oval telescope with a major axis of 8 m and a minor axis of 3.5 m. The oblong mirror design was proposed to reach small inner working angles while keeping the mirror small enough to be easily deployed in space. A variation of TPF-C that used a 50 m diameter disk called an occulter (Cash, 2006), flying in formation with TPF-C at a distance of 50,000 to 70,000 km, was called TPF-O, and predated the HabEx starshade. Our simple design uses a mirror that is more elongated than that of TPF-C, but operates in the infrared like TPF-I. By operating in the infrared, the requirements on the coronagraph are more relaxed, eliminating the need for an occulter.

TPF-C concentrated much more on exoplanet spectroscopy and the identification of biosignatures than we have. However, our proposed mission operations could be enhanced to include identification of other biosignature molecules (for example, Levine et al., 2006, Appendix 1.A), most notably carbon dioxide (15.0 μ m, 10.4 μ m, and 9.3 μ m), methane (7.7 μ m and 8.0 μ m), and water (7.0 μ m and 20.5 μ m). Because AIC only nulls over a 25% bandwidth, we are limited to a spectral bandwith of ∼ 1 − 5 μ m in the mid-infrared. While a few of these absorption bands could be probed in one pointing, others would require a pointing just for that line. Details of the science case might also require measurements of the continuum, for instance to determine the temperature of the exoplanet. Alternatively, this mission could be viewed as a less expensive way to identify the most promising targets for future spectroscopic missions such as HWO and LIFE. Slewing HWO or LIFE from one pointing to the next will be enormously expensive; certainly, the LIFE conops is such that moving the rotating constellation will be very complex and require a significant expenditure of propellant.

Refinements of the science plan should also include varying the exposure time for each star, based on the star’s distance and temperature; closer, hotter stars will need less exposure time to observe exoplanets in the habitable zone. The required exposure time will also evolve as more information on exozodiacal dust, including exozodiacal dust around these stars in particular, becomes available.

In our proof-of-concept calculations, we have not considered repeat observations of exoplanets to determine their orbits. It takes observations of at least three orbital positions to determine an elliptical path of the planet in the sky, but more could be required depending on the positional accuracy of the measurements.

Figure 10 of Stark et al. (2019) shows the number of exoplanets that could be discovered in 2 years of telescope time (1 year of exposure time) as a function of telescope diameter for several popular exoplanet telescope/coronagraph/starshade designs. While our calculation uses different assumptions and is in many aspects less rigorous, it is interesting to note that our estimate of three discovered exoplanets with 1.3 years of telescope exposure with a ∼ 4.5 m telescope is similar to the number of exoplanets that Stark et al. (2019) estimated for a segmented on-axis telescope of the same diameter with an apodized pupil Lyot coronagraph. Looking at the curves for yield as a function of diameter, our yield for a 20 m × 1 m telescope is roughly equivalent to the yield for a 10 m diameter telescope. This suggests that if building, deploying, and operating a 20 m × 1 m telescope is easier and less costly than a 10 m diameter telescope, the rectangular design is preferred.

Our proof-of-concept calculations consider only optics and not the very real engineering considerations for operating a space telescope. For example, we calculated the throughput for two images taken with the mirror rotated 90 ° between the two observations. In practice, rotating a telescope–and in particular a telescope with mass distributed to large radii–could consume a considerable amount of fuel. It might be more advantageous to continuously rotate the telescope through 90 ° . This will cause smearing over the distance that a point source moves in the image plane during the length of an observation. The rate at which an exoplanet will move is v = 2 π r / T , where r is the angular distance from the host star to the exoplanet and T is the rotation period, which is four times the exposure time. For a 10 day exposure time, and a large separation angle of 2″, the exoplanet moves 0.013″ per hour. If the image is read out every hour, or even every few hours, the smearing from rotation will be considerably smaller than the point spread function. This does complicate the data analysis, but if anything increases exoplanet detectability; exoplanets will be excluded inside a circular region with radius equal to half of the diffraction limit for the long axis of the telescope rather than a square with a side length of the diffraction limit for the long axis of the telescope. Exoplanets that would have landed in a place with unfavorable transmission function in one of two images will instead have a much longer effective exposure time because they will be excluded less than half of the time; in exchange, many more exoplanets will fall in the unfavorable region for a small fraction of their exposure.

In this paper, we suggest that the search for exoplanets would benefit from changing the shape of the primary collecting element to a rectangle. We show that using technology that has already been developed for JWST, we could plausibly build an infrared telescope that could achieve the HWO goal of detecting 25 habitable, Earth-like exoplanets and determine whether there is ozone in their atmospheres. Since HWO does not aim to image exoplanets in the mid-infrared, the proposed mission would not only be a precursor mission, but would also provide complementary data for future HWO observations.

The rectangular mirror telescope outlined here is an infrared space observatory with a 20 m × 1 m primary mirror and an AIC coronagraph. This observatory can find an estimated 11 habitable exoplanets orbiting the closest 15 stars with 5100 K < T eff < 6600 K, and determine whether they have ozone in their atmospheres, in a mission of 1 year. If the mission is extended to ∼ 3.5 years, an estimated 27 habitable exoplanets are expected from a sample of 46 F, G, and K stars within 10 pc.

These nearby exoplanets are the most precious planets to find because we will be able to study them in the most detail. The closer the exoplanet is, the more likely we could send a probe to investigate, establish communication with its residents, or possibly one day visit.

A fixed rectangular mirror allows for high resolution measurements only along the long axis of the rectangle. By taking half of the exposure time at one mirror position and half of the exposure time with the mirror rotated by 90 ° (or alternatively by smoothly rotating the telescope through 90 ° during the exposure), we can achieve most of the resolution for finding exoplanets that a more conventional circular telescope with a diameter equal to the long axis of the rectangular mirror does. A wider rectangular mirror will increase the light collecting area of the rectangular telescope, but does not change the inner working angle of the system.

The calculations presented in this paper are intended as proof-of-concept only. There are an enormous number of parameters that can be optimized in the design of such a telescope. The total required exposure time can be reduced by making the mirror slightly wider. The wavelengths at which the planet’s spectrum is observed can be changed. The central wavelength of the observations could be adjusted. The number of habitable exoplanets that were used in our simulation or the zodiacal light estimates could be adjusted, leading to a different yield than was calculated here. Additional instrumentation can be added to enable additional science programs. We suggest here only the idea that a smaller, simpler, and therefore less expensive telescope could be deployed to achieve the primary science driver of the Habitable Worlds Observatory. A more careful mission design would certainly use different exposure times for each target star and a more nuanced plan for identifying biomarkers in the identified exoplanet spectra.

We suggest that a rectangular mirror infrared telescope could identify the most interesting nearby habitable planets for follow-up with HWO or LIFE. But the concept of a rectangular mirror is adaptable to any mission concept that requires high resolution, particularly the ability to distinguish between two point sources. For this purpose, we trade longer exposure times (due to a smaller primary mirror) and more complicated data analysis (the telescope must be rotated through different position angles on the sky if the position angle of the second source is not known) for the ease and lower cost of deploying a much smaller primary mirror. The rectangular mirror concept can be implemented at any wavelength range and in combination with any standard astronomical instrument.