This illustration shows Barnard’s star, with the correct size and temperature/color, as orbited by the four recently confirmed exoplanets around it. All four exoplanets are close in, with orbits ranging from 2.3 to 6.7 days, and small in mass: between 0.17 and 0.34 Earth masses. (Credit: International Gemini Observatory/NOIRLab/NSF/AURA/R. Proctor/J. Pollard)
Barnard’s star, the closest singlet star system to ours, has long been a target for planet-hunters. We’ve finally confirmed it: they exist!
Since we first realized that Earth was just another planet orbiting our Sun — like Mercury, Venus, Mars, Jupiter, or Saturn — we’ve been compelled to wonder whether the other stars in our night sky possessed planets like we do. This question went wholly unanswered, from a scientific perspective, until the definitive detection of exoplanets first arrived: back in 1992. In the time since, we’ve discovered and confirmed more than 5000 exoplanets, including:
rocky exoplanets smaller than Earth,
Earth-sized exoplanets,
super-Earth exoplanets that are likely to be rocky,
mini-Neptune exoplanets that are likely to have thick gas envelopes,
Neptune-sized worlds that are more than a dozen times as massive as Earth,
Jupiter-like exoplanets that are enormous, massive, and puffy,
and even super-Jupiter worlds that approach the mass of brown dwarfs.
However, perhaps due to our bias of having just one star in our own Solar System, we’ve long favored to hunt for planets around nearby, singlet stars, as opposed to planets that might exist in binary, trinary, or richer multi-star systems. The closest three stars to the Sun are Proxima Centauri, Alpha Centauri A, and Alpha Centauri B: a trinary system. However, the next-closest star is Barnard’s star: a red dwarf located a mere 5.96 light-years away. Known as the “Greyhound of the Skies,” it’s one of the fastest-moving stars relative to the background of more distance objects that tend to remain fixed. After more than a century of searching, and a couple of prominent false positives, we’ve finally discovered that it does have planets of its own, after all. Here’s the story behind the discovery.
In the early 21st-century, we’ve successfully mapped out practically all the stars in our neighborhood in three-dimensional space. The closest stars to us don’t always align with the stars we can see, as what’s visible is determined by a combination of distance and intrinsic brightness, but all stars beyond the Sun are at a much, much greater distance than anything within our Solar System. The Alpha/Proxima Centauri system is a trinary, and has the three closest stars to our Sun at present; Barnard’s star is the fourth closest, and is the nearest singlet star system to our own. (Credit: Andrew Z. Colvin)
You might think that being less than six light-years away, and the fifth closest star to Earth (after our own Sun and the Proxima/Alpha Centauri system), Barnard’s star would have been a target of astronomers for millennia. But the star itself has only been known since 1916, when astronomer E.E. Barnard catalogued it and measured its proper motion, or its motion relative to the other, more distant stars in the sky. It turned out that the star moves at 10.3 arcseconds-per-year relative to the Sun: the fastest motion of any star as seen from our point-of-view.
It’s a remarkably interesting star, as well. It’s a red dwarf star: the most common type of star in the Universe but also the faintest and coolest. It has a mass of only 16% our Sun’s mass, and a radius that’s just 19% of our Sun’s, or only about double the radius of planet Jupiter. It’s cool, radiating at a surface temperature of around 3200 K. The star itself is has only 28% of the heavy elements that our Sun has, suggesting that it was created long before our Sun was. It’s also a very slow rotator, taking 142 days to spin just once around its axis. All of this, combined, suggests that Barnard’s star is a relatively old star: formed an estimated 10 billion years ago.
Just as one would expect for a star that’s this:
old,
cool,
and low in mass,
it does exhibit flaring activity, but very little of it. Whereas red dwarfs that are under a billion years old flare at exceedingly high rates, Barnard’s star shows relatively little stellar/flaring activity.
This five-year timelapse shows the relative motion of Barnard’s star, the one star with the fastest proper motion in terms of angle as viewed from our Solar System, relative to the background of other stars in its vicinity. Although it will still take around 350 years for the star to move even 1 degree across the sky, that’s incredibly fast for an object located several light-years away. (Credit: Paul Mortfield & Stefano Cancelli)
Barnard’s star was also the alleged site of the very first claimed exoplanet detection: all the way back in the 1960s. Peter van de Kamp was a Dutch astronomer who was the longtime director of Sproul Observatory at Swarthmore College, and for decades — from 1937 through the late 1960s — painstakingly collected data on a number of stars, including (and especially) for Barnard’s star. He thought that because the star was moving so quickly, any planets orbiting it would cause its apparent position to “wobble” over time, revealing the presence of any massive enough worlds in its planetary system.
After many years, he began to find evidence of what he called rhythmic undulations, and by 1963, he claimed to have collected enough evidence to show the existence of a planet: a Jupiter-mass world that orbited the star with a period of 24 Earth years. By 1969, at the height of the space race (nearly contemporaneous with the Apollo 11 Moon landing), van de Kamp updated his report to include evidence of a second planet. If there were planets around this star, located only a mere 6 light-years away, perhaps that should be our ultimate target. After landing on the Moon and reaching for Mars, perhaps Barnard’s star should be exactly where we set our sights on as our first destination beyond the Solar System?
The extreme temperature differences between a world in full sunlight at its closest approach to its parent star versus nighttime and/or its farthest point in orbit on that same world can lead to extreme temperature differences, with extreme consequences for phase changes in matter. It can also lead to a misidentification of exoplanets, as stellar rotation and starspots must be taken into full account. (Credit: T. Mikal-Evans (MPIA) / T. Müller (MPIA/HdA))
The key to proving the existence of these planets, however, would demand independent confirmation. The aging van de Kamp began looking for a successor, and settled on Wulff Heintz, a careful astronomer who was incredibly excited by van de Kamp’s claims. With full access to all of van de Kamp’s records, Heintz’s initial enthusiasm was swiftly replaced by skepticism. Data from the 1930s and 1940s appeared inconsistent with later data from the 1950s and 1960s, and Heintz wanted to know why. He noted that the telescope was:
disassembled,
cleaned,
and then reassembled,
all in 1949. This was coupled with his noticing that many plate images were blurry, and therefore were overexposed.
When he began analyzing the data itself, he took care to record the errors and uncertainties in defining a position, at any moment in time, for Barnard’s star based on van de Kamp’s data. When the errors were included, Heintz reached a startling conclusion: the data was insufficient to determine whether or not such planets existed. The jumps that van de Kamp had recorded fell well within the range of the errors provided by the measurements, and so the evidence supporting the existence of these planets evaporated. Although van de Kamp objected, no credible exoplanet discoveries would occur until 1992.
The discovery of the first 5000 exoplanets, as recorded by year and by method. For the first ~15 years or so, the radial velocity method was the dominant method of discovery, later superseded by the transit method beginning with NASA’s now-defunct Kepler mission. In the future, microlensing may surpass them all, as microlensing will be sensitive to low-mass (i.e., Earth-mass and below) exoplanets in a way that the prior two main methods have not been with current instrumentation. These confirmed planets represent only a fraction of the total planetary candidates. (Credit: NASA/JPL-Caltech/NASA Exoplanet Archive)
Interestingly, the first slew of exoplanets were indeed found with van de Kamp’s method: the stellar wobble method, where the star itself appears to move both side-to-side (which isn’t easily observable) as well as forward-and-back along our line-of-sight (which is observable), and was especially sensitive to planets that were very close to their parent stars. In the 21st century, the Kepler and TESS missions led to an explosion of exoplanets using a different method: the transit method, which only works for planets that happen to be aligned in such a fashion that the planet’s disk actually passes in front of the star’s disk from our perspective.
The transit method has brought us the majority of the known exoplanets as of today, with the stellar wobble method sitting in second place. While Barnard’s star doesn’t exhibit any evidence for planetary transits, that doesn’t mean that there are no planets orbiting it; it only means that any planets that do orbit it haven’t been caught crossing in front of the star’s disk from our perspective. However, the stellar wobble method would still be sensitive to any such planets that orbited it.
At last, in 2018, a credible-seeming claim of exoplanets around Barnard’s star arrived: claiming that there was good evidence for an exoplanet of about three Earth masses orbiting Barnard’s star with about the same orbital period of Venus around the Sun. With over 20 years of high-quality data associated with Barnard’s star, this claim seemed much more robust than van de Kamp’s earlier ones.
The idea of the radial velocity (or stellar wobble) method is that if a star has an unseen, massive companion, whether an exoplanet or a black hole, observing its motion and position over time, if possible, should reveal the companion and its properties. This remains true, even if there’s no detectable light emitted from the companion itself; it cannot have its properties studied the way a transit can allow. (Credit: E. Pécontal)
Again, though, what we need is independent confirmation. One study that shows evidence for a periodic signal where the star’s light becomes redder (allegedly, because it’s moving away from us) and bluer (because it’s moving towards us) is a start, but we have to take care to rule out all other possible causes. After all, stars and star systems are messy astrophysical places, where lots of different processes play a role. Stars vary, have starspots, and exhibit flares, and it’s possible that long-term stellar activity could mimic this exact type of signal.
In 2021, a follow-up study was conducted on Barnard’s star with a different instrument: the Habitable-zone Planet Finder. They didn’t find evidence for a planet, but instead found evidence for stellar activity: mild flares and starspots, where the star’s rotation (with a period of around 145 days, with an uncertainty of around 9 days) and Earth’s revolving around the Sun (with a period of around 365 days) leads to a data artifact: a false, transitory signal (known as an alias) with a period that’s in-between, obeying the formula (1/rotation period — 1/Earth’s revolutionary period) = (1/the false signal period). If the rotation period turns out to be 142 days instead of 145 days, which is well within the margin of error, this reproduces the 233 day period for the suspected planet exactly.
In other words, there is no planet, and furthermore, emphasizes the fact that finding non-transiting planets around even nearby red dwarfs is hard!
Our notion of a habitable zone is defined by the propensity of an Earth-sized planet with an Earth-like atmosphere at that particular distance from its parent star to have the capacity for liquid water, without a cover of ice, on its surface. Although this describes the conditions that Earth possesses, it is unknown whether this is a requirement, or even a preference, of life. Many worlds assumed to be good candidates for life will likely be uninhabited; others not presently considered will likely surprise us down the line. (Credit: Chester Harman; NASA/JPL, PHL at UPR Arecibo)
But we shouldn’t be discouraged by “false detections” in our search for bona fide planets around other stars; just because a scientific endeavor like planet-finding is difficult doesn’t mean we should assume that there aren’t any planets at all! Instead, we should demand that we get better data, and use that data to actually determine whether there are planets present or not, and if so, what their properties are.
This is exactly what happened last year: in 2024. In a very brave paper by Jonay I. González Hernández and others, they used the ESPRESSO instrument aboard an 8.2 meter ground based telescope, the ESO’s VLT, to acquire a large number of observations (a total of 156) over a four year period. This time, they already knew to model the long-term stellar activity cycle (with a period of ~3200 days, or about 8 years and 9 months) and the rotation-driven activity of the star (with a period of ~140 days), fully accounting for these intrinsic variations that can mimic a stellar wobble-type signal.
After collecting and calibrating all the data, and accounting for the known intrinsic variations of the star, they published their results. They wound up finding four short-period, low-mass candidate planets, with periods of 2.34 days, 3.15 days, 4.12 days, and 6.74 days, respectively. The strongest evidence emerged for one of those planets in particular: a ~0.37 Earth mass exoplanet with an orbital period of 3.15 days.
This figure shows, at left, a table of the four candidate exoplanets around Barnard’s star identified by ESPRESSO data, along with the mass-period relationship of Barnard’s star’s planets (top right) and a model of the planetary system relative to the star’s (green) habitable zone at lower right. (Credit: J.I. González Hernández et al., Astronomy & Astrophysics, 2024)
This would represent a fascinating find, if confirmed. First off, these four planets would be just a little interior to the so-called habitable zone of its star: where a planet with an Earth-like atmosphere would have the right temperatures for liquid water on its surface. The habitable zone for Barnard’s star would represent any planets that had an orbital period of between 10 and 42 days; the equilibrium temperature of the strongest planetary candidate from that study, with a mass of ~0.37 Earth masses and a period of 3.15 days, would be around ~400 K: just a little bit too hot for Earth-like conditions.
But, as always, we cannot trust just one study with one instrument to declare that these exoplanetary candidates are indeed exoplanets; we’d need independent confirmation.
That’s exactly where the new MAROON-X instrument, mounted upon the NSF’s Gemini telescope, comes into play. MAROON-X can provide complementary, independent data to ESPRESSO for any star that both instruments look at over the same relevant time period. That’s the ideal situation for confirmation: looking at the same objects at the same time with different observatories and different instruments. This eliminates calibration errors, errors specific to one instrument or one observatory, or errors that arise due to a particular method of data analysis that only one such team uses.
Instrument scientist Luke Gers, in a clean suit, participates in the assembly of the MAROON-X instrument at Gemini Observatory North, on the summit of Mauna Kea, in 2019. MAROON-X is one of the pre-eminent planet-finding instruments from the ground using the radial velocity (or stellar wobble) technique. (Credit: International Gemini Observatory/NOIRLab/NSF/AURA/A. Peck)
In a very exciting paper, led by Ritvik Basant and published on March 11, 2025, a total of 112 radial velocity measurements were taken independently with the MAROON-X instrument, and they confirmed the three strongest exoplanet candidates identified by ESPRESSO. That means they’ve confirmed:
exoplanet Barnard’s b, with a period of 3.154 days and an estimated mass of 0.32 Earth masses,
exoplanet Barnard’s c, with a period of 4.124 days and an estimated mass of 0.31 Earth masses,
and exoplanet Barnard’s d, with a period of 2.340 days and an estimated mass of 0.22 Earth masses.
Furthermore, they are sufficiently good that, when the MAROON-X data is combined with the ESPRESSO data, they can at last confirm a fourth likely exoplanet: Barnard’s e, with a period of 6.738 days and a very small mass of just 0.17 Earth masses. For this last planet, its equilibrium temperature is estimated to be around 310 K, which could mean it’s got the potential to have liquid water on its surface, which would be a fascinating scenario.
Importantly, these four now-identified planets are all lower in mass than Earth is, represent a very compact system. Even though there are reasons to disfavor the notion that these planets might have Earth-like atmospheres, or any substantial atmosphere at all, it’s a remarkable feat to detect them at all. These planets, even in orbit around such a low-mass star, represent something the smallest radial velocity (or stellar wobble) signals detected at the present time.
These curves show the radial velocity data, stacked to correspond to the period of each planet, for Barnard’s star. The four suspected exoplanets, Barnard’s b, c, d, and e, are all shown, with the evidence from MAROON-X and ESPRESSO overwhelmingly confirming the innermost three planets, with the fourth being announced but far less certain. (Credit: R. Basant et al., Astrophysical Journal Letters, 2025)
The data confirming the inner three exoplanets is much stronger, however, than the data in favor of the fourth; it may yet be that a combination of the inner two planets, orbiting together, is mimicking the signal of the suspected fourth, as three times the period of the innermost planet (7.02 days) and twice the period of the next-inner planet (6.31 days) would, if averaged together (6.76 days) almost perfectly match the signal of the suspected fourth planet. It may yet turn out not to be real! In addition, there’s another big find that comes along with combining the MAROON-X data with the ESPRESSO data: it’s sufficient to rule out the presence of any Earth-mass planets (or any planet with more than 57% of Earth’s mass) within the entire Habitable Zone around Barnard’s star: with periods of between 10 and 42 days.
Even though this data is very compelling and these four exoplanets have been announced as confirmed exoplanets, we must keep in mind that some of them could still be spurious, with the outermost exoplanet being the most suspicious. But, for the first time in history, we can accurately claim that the closest singlet star system to our own — Barnard’s star — most definitely has exoplanets, and that the exoplanets we’ve found are close in, in compact orbits, and even lower in mass than the Earth is.
Do any of these worlds have atmospheres, liquid water on their surface, and even possibly life on them? Are all four of these planets real? Are there any more worlds out beyond the innermost ones discovered thus far? It’s an exciting time to be an exoplanet scientist, as not only do we hope to find out the answers to these questions, but with the plans for NASA’s upcoming Habitable Worlds Observatory, we may soon be able to even directly image some of these planets on their own.
Starts With A Bang is written byEthan Siegel, Ph.D., author of (affiliate links following)Beyond The Galaxy,Treknology, andThe Littlest Girl Goes Inside An Atom. His first National Geographic book,Infinite Cosmos, is out now!