The target set

The mission-duration rule

Six times the distance in light-years.

Total mission time is transit time plus the light-time for the data to come home. At 0.2c the transit is five years per light-year and the return is one year per light-year, so total mission time in years is simply six times the distance in light-years. At 0.1c it is 11 times. At 0.05c it is 21.

Now impose an institutional filter. Call a 40-year total the career horizon: an engineer who joins the program at 30 sees the data at 70. Call an 80-year total the institutional horizon, roughly the interval from the opening of the Panama Canal to the present, and about the outer limit of a program a single organization has ever sustained with continuous purpose.

At 0.2c, the career horizon admits targets within 6.7 light-years: Alpha Centauri, Proxima, and Barnard’s Star. Nothing else. At 0.1c, the career horizon admits targets within 3.6 light-years, and there are none.

This is why every serious beamed-sail study converges on 0.2c, and the convergence is not aspiration. It is the threshold velocity at which the nearest star system falls inside a single professional lifetime. Below it, no target qualifies. The number was never chosen for its physics.

The candidates

Nine systems, and what each costs in time.

Distances and stellar parameters are Measured. Planet masses are radial-velocity minimum masses, which carry the sin i ambiguity — the true mass is the minimum divided by the sine of an unknown orbital inclination. Mission time is transit plus one-way light time for data return, and excludes any deceleration phase.

Proxima Centauri

The highest readiness and the lowest expectations.

Proxima b is the closest known exoplanet, at 4.2465 light-years, with a minimum mass of about 1.07 Earth masses, an 11.19-day orbit at 0.0485 AU, and an equilibrium temperature placing it inside the conventionally defined habitable zone. It is almost certainly tidally locked.

It also orbits a violently active M dwarf. A naked-eye superflare was observed in 2018 with a bolometric energy around 10³³⋅⁵ erg, roughly a thousandfold brightness increase over 10 seconds. TESS photometry implies about three superflares per year at 10³³ erg and roughly one event every two years at 10³⁴. A Sun-like star produces a 10³⁴–10³⁵ erg event about once per 1,000 years. A planet at 0.0485 AU has been sitting inside that for on the order of five billion years. The 2025 ALMA millimeter campaign added the first constraints on flare rates at millimeter wavelengths and did nothing to improve the picture.

The consensus expectation is that Proxima b has no atmosphere worth the name. That does not disqualify it. It sharpens the mission.

A flyby probe does not need the planet to be habitable. It needs the planet to be measurable, and the single most consequential measurement available anywhere in the target set may be a negative one: does an old rocky planet in the habitable zone of an active M dwarf retain any atmosphere at all?

Roughly three quarters of the stars in the galaxy are M dwarfs. If their habitable zones are systematically stripped, the overwhelming majority of nominally habitable real estate in the universe is dead. If Proxima b has an atmosphere despite everything, the opposite follows. Either result resolves a question about the entire galaxy, and Proxima b is the nearest place to ask it.

You do not go to the most habitable planet. You go to the planet whose answer resolves the largest class of planets.

Alpha Centauri A and B

The only Sun-like stars inside a career.

Alpha Centauri A is a G2V star of 1.1 solar masses and 1.52 solar luminosities; B is a K1V of 0.9 solar masses and 0.5. The pair orbits with a semi-major axis of 23.5 AU and an eccentricity of 0.52, giving a periastron of about 11.2 AU and a period of 79.9 years. Dynamical studies place the stable region for terrestrial planets inside roughly 2.5 to 3 AU of each star, with habitable zones at about 1.2 AU for A and 0.7 AU for B — comfortably inside that limit. Earth analogs are permitted.

None has been detected, and the reason is instrumental rather than physical. The binary is the problem: even with a coronagraph masking A, light from B floods the field.

The strongest current lead is S1, a 3.5 millijansky point source detected at 15.5 microns by JWST MIRI in August 2024, 1.5 arcseconds from Alpha Centauri A, corresponding to about 2 AU. It was not recovered in February or April 2025. If it is the same object as C1, seen by VLT NEAR in 2019, then 52 percent of the dynamically stable orbits consistent with the combined astrometry are also consistent with those two non-detections, at periods of two to three years, eccentricity near 0.4, and mutual inclination near 50 degrees to the AB orbital plane. It would be a Saturn-mass giant in the habitable zone of the nearest Sun-like star. It is a candidate, not a planet.

One piece of geometry is worth noting. Proxima lies about 0.21 light-years from Alpha Centauri AB, roughly 13,000 AU. At 0.2c that is about 13 months of additional flight. A single launch corridor serves three stars, one of them Sun-like, one of them host to the nearest known rocky planet. No other direction in the sky offers that, and it is the basis of the beachhead argument on arrival and deceleration.

Barnard’s Star

Already checked, and it came back empty.

Barnard’s Star is the nearest single star at 5.963 light-years, an M4V of 0.0035 solar luminosities, with a century of retracted planet claims behind it — most famously van de Kamp’s astrometric planets of the 1960s.

That changed in 2024 and 2025. ESPRESSO identified four sub-Earth candidates; MAROON-X independently confirmed planet b and candidates c and d from 112 radial velocities taken contemporaneously and without coordination between the teams; the combined dataset confirmed the fourth. The system is four sub-Earths with periods of 2.34, 3.15, 4.12 and 6.74 days and minimum masses between 0.19 and 0.34 Earth masses. Barnard e may be the lowest-mass planet ever detected by radial velocity. All four are far too hot.

The finding that matters is the non-detection. The combined data rule out planets above 0.57 Earth masses in the habitable zone, at periods of 10 to 42 days, with 99 percent confidence.

Barnard’s Star is therefore not a target that might improve with better evidence. It is one that has already been checked. It retains real value as the second-nearest system, as a career-horizon-accessible test of compact M-dwarf system architecture, and as an in-situ measurement of a very old, very quiet, very low-metallicity stellar environment. It is no longer a habitability target, and any treatment that lists it as one is working from pre-2025 information.

Tau Ceti, Teegarden, Luyten

And three that belong on the list.

Tau Ceti is the nearest single Sun-like star at 11.9 light-years, a G8V of 0.52 solar luminosities with notably low metallicity. Its planet candidates e and f have been proposed and contested repeatedly, with radial-velocity signals close to the activity noise floor. The load-bearing fact is not its planets but its debris disk, inferred from far-infrared excess to hold roughly 10 times the dust of the solar system. That cuts twice: dust indicates a dynamically active system with ongoing collisions, which is a habitability discount because anything in the habitable zone is under sustained bombardment, and simultaneously a survivability discount for the probe, because a flyby at 0.2c through a dense circumstellar disk is a materially different risk problem. The hazard and the signal are the same physical fact.

Teegarden’s Star is an M7.0V ultracool dwarf of only 0.097 solar masses at 12.5 light-years. With TRAPPIST-1 it is one of only two known planet hosts at or below a tenth of a solar mass. Planets b and c have periods of 4.90 and 11.42 days and minimum masses of 1.16 and 1.05 Earth masses; a third, d, was announced in 2024 at 26.13 days and 0.82 Earth masses, with a fourth candidate at 172 days. The star is quiescent for its class, and M4 to M10 dwarfs appear least likely to erode secondary atmospheres. If Proxima b tests whether M-dwarf habitable zones survive violence, Teegarden b and c test whether they work at all when the violence is absent. Recent general circulation modeling puts Teegarden b inside the habitable zone but close to the inner edge, at about 1,481 watts per square meter against Earth’s 1,361 — near the runaway greenhouse threshold rather than comfortably inside.

Luyten’s Star (GJ 273) is an M3.5V at 12.36 light-years hosting GJ 273 b, a minimum 2.89 Earth-mass planet on an 18.65-day orbit inside the habitable zone. It carries one distinction no other target has: in October 2017 the Sonar Calling GJ273b project transmitted a deliberate message from the EISCAT facility in Norway. It arrives in 2030, and a reply would arrive around 2043.

Three absent from most short lists deserve mention. Ross 128, at 11.0 light-years, hosts a 1.4 Earth-mass temperate planet around an unusually quiet M4 dwarf, and on the combination of proximity, stellar quiescence and planet properties arguably outranks both Luyten and Teegarden while being closer than either. GJ 1002, at 15.8 light-years, hosts two temperate Earth-mass planets around a very quiet star — too far for this century, but it belongs on a 2150 list. And Sirius A and B, at 8.60 light-years, has no habitability case whatever and belongs on the list for a completely different reason: at 25.4 solar luminosities it is by an enormous margin the best photon-braking target within reach.

You cannot choose what you cannot characterize

The observatory is the targeting system.

The detection limits that currently hide any terrestrial planets at Alpha Centauri are exactly what the 2050s and 2060s observatory program is built to lift. Optical space interferometry, direct imaging, and above all the solar gravitational lens nodes are not a parallel scientific effort running alongside a launch program. They are the targeting system for it. If the observatories are late, the launch program has nowhere useful to point.

China’s contribution to that layer is the Closeby Habitable Exoplanet Survey (CHES), an astrometry mission proposed under the Chinese Academy of Sciences Strategic Priority Program on Space Science. It would use a 1.2-meter coaxial three-mirror telescope with a 36-meter focal length and a 0.44 by 0.44 degree field, working to one microarcsecond of positional accuracy — which would make it the most precise astrometric instrument ever flown — against roughly 100 FGK-type stars within 10 parsecs. It expects to find on the order of 50 Earth-like planets and, by measuring true masses and three-dimensional orbits rather than radial-velocity minimum masses, would remove the sin i ambiguity that limits nearly every entry in the table above. CHES remains in assessment alongside the Earth 2.0 proposal, with no confirmed approval or launch date, and it is the first space mission of any nation dedicated exclusively to terrestrial planets in nearby habitable zones.