A 3.5-Meter Space Coronagraph Could Open Nearby Giant Planets—but Earth-Sun Twins Stay Out of Reach

A Telescope That Turns Off the Star

Directly imaging an exoplanet is like photographing a firefly next to a searchlight. The firefly is the planet; the searchlight is its star. The proposed 3.5-meter Segmented-Mirror Robotic Space Telescope would tackle that problem with a coronagraph, an instrument that blocks starlight and suppresses the diffraction pattern that would otherwise swamp a faint companion.

The paper, posted to arXiv in September 2026, is a mission white paper, not a detection announcement. It lays out what the exoplanet program could do if the telescope is built. The baseline design includes an 18-segment primary mirror, a high-contrast coronagraph, wavefront sensing and control, pointing stability, and operations software needed for coronagraphy from the outset.

How Faint Is a Planet?

Contrast is the planet-to-star brightness ratio. A Jupiter twin in reflected visible light is roughly 10^-9 as bright as its star. An Earth-Sun twin is about 10^-10. The Roman Space Telescope's coronagraph is a 10^-8-class technology demonstration; its best-effort models reach a few times 10^-9. The 3.5-meter white paper treats a post-processed contrast near 10^-9 as a development goal, not a demonstrated number. At that level, mature giant planets become accessible. Earth-Sun analogs remain a stretch.

The Geometry Sets a Distance Horizon

The coronagraph's inner working angle is the smallest separation from the star where it can deliver high contrast. At 550 nm, a 3 lambda over D inner working angle is about 0.097 arcseconds. That geometry creates a reflected-light horizon. A planet at 1 AU can stay outside the inner working angle only for stars closer than about 10.3 pc at 3 lambda over D, or 15.4 pc at a more forgiving 2 lambda over D. A Jupiter analog at 5.2 AU is accessible out to about 53.5 pc. At longer infrared wavelengths the diffraction limit worsens: at 1.6 micrometers the Earth-twin horizon shrinks to about 3.5 pc. The nearest stars are therefore not just convenient targets; they are the only ones that make some measurements possible.

The Nearest 20 Stars, Then the Nearest 20 FGK Stars

The observing strategy is distance-first. The primary exploration sample is the 20 nearest stellar components, belonging to 13 systems and lying within about 3.5 pc. The team chose them strictly by distance, without filtering out active, multiple, or planet-hosting stars. That reduces selection bias but also creates hard cases: close binaries such as GJ 65 and the triple EZ Aquarii may need specialized wavefront control or restricted search zones. The paper keeps all 20 rather than dropping difficult systems.

A second tier targets the 20 nearest F-, G-, and K-type main-sequence stars whose habitable zones could fall outside the coronagraph's inner working angle. These are main-sequence F, G, and K stars, ranging from hotter than the Sun to cooler, where liquid water could exist at habitable-zone distances.

Giants First, Earth Twins Later

At 10^-9 contrast, the mission's most robust reflected-light targets are mature giant planets, not small rocky worlds. The paper also notes that young giant planets, still glowing from formation, can be 10^-6 as bright as their stars in the near infrared. Those are easier targets and could be observed early. The team identifies six known planets as first-light candidates because their maximum angular separations fall between the coronagraph's inner and outer working angles and their modeled reflected-light contrast clears 10^-9.

A Narrow Window Around Cool Dwarfs

Earth-like planets are harder. One workaround is to look at cooler K stars. Because their habitable zones lie closer to the star, an Earth-like planet there receives the same instellation at a smaller orbital separation. That boosts its reflected-light contrast. But the same closeness shrinks the angular separation, pushing the planet near the coronagraph's inner working angle.

The paper examines 61 Cygni A and Epsilon Indi A, two nearby K5 stars. For an Earth-sized planet at the Earth-equivalent insolation distance, the modeled contrasts are about 1.2 x 10^-9 and 6.9 x 10^-10, with angular separations of about 108 and 137 milliarcseconds. The nominal inner working angle is about 97 milliarcseconds. Both systems sit in a narrow, difficult regime. The paper flags 61 Cygni A as conditional because companion light from 61 Cygni B must be modeled, while Epsilon Indi A is retained but its predicted contrast falls below the nominal 10^-9 goal. The team's optimistic photon-noise estimates for a 5-sigma molecular-band detection are 8.7 and 12 hours, respectively. Those are lower bounds, not scheduled exposures.

What the Calculations Do Not Include

The paper is explicit about its limits. The 10^-9 contrast is a development goal, not an achieved performance. The exposure-time estimates assume idealized conditions and omit correlated residual speckles, detector noise, template mismatch, and the loss of coronagraph throughput near the inner working angle. The target catalog is preliminary: some candidates are early-M stars, IV–V luminosity-class objects, or binaries that need further review. The six first-light planets are known systems; recovering them would test characterization, not blind discovery. Zodiacal and exozodiacal light add noise, and exozodiacal clumps can mimic planets; background galaxies and other discrete sources can create false positives. Multi-epoch astrometry and spectroscopy can remove discrete contaminants.

A Bridge Between Roman and the Habitable Worlds Observatory

The 3.5-meter concept sits between the Roman Space Telescope's 2.4-meter coronagraph pathfinder and the proposed Habitable Worlds Observatory, a roughly 6-meter mission designed to reach 10^-10 contrast and characterize Earth analogs. The white paper argues that a smaller robotic telescope could mature segmented-aperture wavefront control, build a reference atlas of reflected-light spectra from nearby giants, and provide a lasting census of the Sun's nearest neighbors. Its wide-field imager would complement the coronagraph with transit photometry, occurrence statistics, and long-term monitoring of stellar activity.

A five-year reference schedule in the paper assigns 6,350 hours to the Volume III coronagraph program, about 14.5 percent of the wall clock. But that schedule is a feasibility exercise, not a promise.

What It Would Mean to Look Next Door

If built, this telescope would not by itself answer whether we are alone. It would do something more tractable and still profound: turn the Sun's immediate neighborhood into a known planetary landscape. It could image and read the atmospheres of giant planets, probe the habitable zones of the nearest Sun-like stars, and set the observational foundation for the larger telescopes that might one day find an Earth twin.