Beta Pictoris d: a new spectroscopic insight into planetary formation

If two teams find the same planet in the same week using completely different methods, whose numbers do you trust?

Welcome to FreeAstroScience. Two papers landed in the same astrophysics journal claiming the same newly found planet, so we went back through both to see where the numbers actually line up and where they don’t.

A third giant planet around the young star Beta Pictoris was found independently by two research teams within the same week of 2026: Aidan Gibbs’s team using infrared spectroscopy, and Ben Sutlieff, Markus Bonse and Jayne Birkby’s team using direct imaging plus archival data. The two papers put the planet’s mass at roughly 2 to 4 Jupiter masses and its orbit at 24 to 26 astronomical units, figures that agree in broad strokes but not in every detail. Neither paper confirms the other: each team reached the same object independently, through its own evidence.

Most of the coverage since has told this as one find with an assist. A team spots something, and a second team checks their work. That is not quite what happened. Aidan Gibbs and his co-authors at UC San Diego picked up a faint spectroscopic signal while studying an entirely different planet in this system, using the James Webb Space Telescope’s NIRSpec instrument. In Edinburgh and Oxford, Ben Sutlieff, Markus Bonse and Jayne Birkby were independently combing eleven years of archived telescope images for exactly this kind of object, using direct imaging instead. Neither team knew about the other’s result until both papers landed together in the same ApJL issue. Gibbs et al.’s own paper says as much, in its introduction: “An independent and contemporaneous discovery of this planet using VLT/ERIS and archival observations is separately reported by B. J. Sutlieff et al. (2026).” That’s a genuinely different story than confirmation. It’s also a slightly humbling correction for anyone, including us, who read “confirmed” too quickly the first time.

Updated August 2026 to cover both of the papers now published, not just one team’s account.

Two teams, two instruments, one planet

Beta Pictoris sits about 63 light-years away in the constellation Pictor. Astronomers have used the system as one of the field’s most useful test beds for watching planets form, mostly because it is still so young. Gibbs et al.’s own paper adopts 23 million years for the system, taken from a 23-million-year isochrone in the ATMO 2020 evolutionary models, and the Max Planck Institute for Astronomy’s release on this discovery agrees. Sky and Telescope’s coverage says 20 million, three million years younger. We can’t referee the isochrone fitting ourselves, but when the primary paper and its own press release land on the same number, that’s the one worth trusting over an outlying secondary account.

Two giant planets, Beta Pictoris b and c, were already confirmed in this system before 2026. We’re leaving c’s own detection history out of this account on purpose: it doesn’t bear on how d was found, and it deserves its own piece rather than a footnote here. Our earlier look at how Beta Pictoris b itself likely formed already noted the system hosts at least three giant gas planets, without naming the third one. That piece drew on interferometry data pointing toward an inner-disk origin for b, revising the outer-disk story once read from its atmosphere’s carbon isotopes. Now it has a name. Beta Pictoris d is the widest-orbiting of the three, sitting near the inner edge of the ring of leftover dust and rubble that still surrounds the star.

How does a spectrograph find a planet nobody had photographed?

Gibbs and his team were not hunting for a third planet at all. They were pointing the James Webb Space Telescope’s NIRSpec instrument, tuned across wavelengths from 2.87 to 5.14 microns, at the already-known planet Beta Pictoris b. A second signal turned up in the same field. Data from a follow-up epoch, including the MIRI instrument’s mid-infrared channel, confirmed it was not noise. Gibbs et al.’s extracted spectrum shows methane absorption at 3.3 microns, evenly spaced carbon monoxide lines between 4.45 and 5.0 microns, and water absorption further into the infrared. Together, they form a chemical fingerprint, read directly off the planet’s own light rather than inferred from a picture of it.

Turning that spectrum into a mass took real interpretive work, and the paper is candid about the trouble. An initial chi-square fit favored a scorching 900 to 1,200 kelvin. The authors say plainly why they don’t trust it. “We caution that this temperature range is simply the most consistent with the shape of the continuum-subtracted spectra, but it is very unlikely to bound the true effective temperature,” they write. That hot a fit would also mean a mass of roughly 5 to 8 Jupiter masses. A planet that massive would probably have shown up already in earlier ground-based imaging and JWST coronagraphy searches that came back empty. So the authors argue instead for a cooler 600 to 800 kelvin, on physical grounds. That part isn’t hand-waving. The radii come out consistent within roughly 20 percent of evolutionary-model predictions, the fit matches the spectrum’s own line-depth diagnostics better, and it agrees with those same non-detection limits. Picture a wood-fired pizza oven running flat out: that is roughly what 600 to 800 kelvin feels like, in case the number alone means nothing. It is also why the mass comes out as a range, 2 to 4 Jupiter masses, rather than a single figure the way a scale would give one.

Reading a planet’s chemistry out of its spectrum rather than its picture is not unique to this system. Our coverage of GJ 504 b, a much colder directly imaged planet with salt-laced clouds, found in the same young stellar moving group as Beta Pictoris, shows the same trick working from the opposite direction. Nobody argues about the orbit there, only the atmosphere.

A decade of archived light nobody had matched

Sutlieff, Bonse and Birkby’s team took the harder, slower route. Beta Pictoris b already outshines any smaller companion nearby. The newly found planet turned out to be about 100 times fainter still, by some measures the faintest exoplanet ever imaged directly from Earth. Such light is faint and fickle, easy to mistake for a cosmic-ray hit or a scattering artifact if you only have one exposure to judge it by. So the team went looking through more than eleven years of observations instead of waiting for a cleaner shot. They worked with the VLT’s ERIS instrument, cross-checked against the older SPHERE instrument on the same telescope, and pulled in archival James Webb NIRCam images. Leyla Iskandarli spotted what she described as an extra, fuzzy blob in one observing run. Markus Bonse, working independently with a machine-learning approach to the same archive, found the same signal on his own. Only once the two lines of evidence matched did the team go back and confirm the planet in images taken as far back as 2014.

Sutlieff and colleagues also give the planet’s orbital period: 91 years, from the fitted orbit. Ninety-one years is a clean way to size up the age question named above. At 91 years per lap, the system’s better-sourced age of 23 million years puts the planet through roughly 253,000 orbits; even Sky and Telescope’s lower estimate only trims that to about 220,000. Either way, this is a planet whose single orbit outlasts human memory, and whose full history is still a small fraction of its star’s short life.

One thing is worth pausing on before the numbers get complicated. Its light had been sitting in observatory archives the entire time. Call it the equivalent of a photograph that took over a decade to get developed. “Planet d, it seems, has been playing a game of hide-and-seek with us for over a decade and only now can we say ‘found you!'” Jayne Birkby said that, in the University of Oxford’s release announcing it. Ben Sutlieff called it “a serendipitous discovery” in the same release. He added a line worth keeping. “Systems with multiple directly imaged exoplanets are the ‘holy grails’ of discoveries, because they can teach us a lot about what different exoplanets are like in the same formation environment.” Direct imaging is this hard for a reason. Our look at new photonic techniques being developed to suppress a star’s glare for future NASA telescopes lays out exactly the problem this team spent eleven years working around.

So which orbit is right, 24 AU, 26, or beyond 30?

Put the two papers’ numbers side by side and the picture gets more interesting, not less. Gibbs et al.’s direct fit to the planet’s motion gives a semi-major axis of 24 astronomical units, with an uncertainty spanning roughly 19 to 32 AU. Sutlieff, Bonse, Birkby and their collaborators get 26.0 AU, with a tighter but lopsided uncertainty of about 19.9 to 28.2 AU.

Table 1: Beta Pictoris d’s mass and orbit, as each independent discovery paper reports it (Gibbs et al. 2026 and Sutlieff, Bonse, Birkby et al. 2026).

Overlap those two ranges and there is a wide band, 19.9 to 28.2 AU, where both fits agree. Neither team is actually fighting over where this planet sits.

There’s one number that doesn’t fit that tidy overlap.

Gibbs et al. separately argue, from orbital stability and from what it would take to carve the debris disk’s inner edge into its observed shape, that the planet’s true orbit is probably beyond 30 astronomical units. Thirty AU sits outside Sutlieff and colleagues’ own error bar entirely. Gibbs et al. are unusually candid about why their own direct fit might be running low: “given the small amount of orbital coverage, which can bias the posteriors obtained in the orbit fit.” In plain terms, nobody has watched this planet complete enough of its 91-year lap to pin the ellipse down from motion alone. Their fit leans on limited data. The stability case leans on a different kind of evidence entirely.

So are the two teams actually fighting over the orbit? Not really.

On temperature, the two papers land closer together than the orbit numbers might suggest. Sutlieff and colleagues report an equilibrium temperature of 600 kelvin, plus 45 or minus 60. Gibbs et al., after walking back their own initial best fit as implausible, settle on 600 to 800 kelvin. Two teams, two instruments, and a number that lands in roughly the same place either way.

Two systematics that could still move these numbers

Some limits are worth naming plainly rather than letting a reader stumble onto them later. Gibbs et al. flag the first one themselves: the mass and effective temperature both come from a continuum-subtracted spectrum. Their own paper says outright that this “makes accurate determination of the planetary effective temperature and mass difficult due to the degeneracy of line depths and shapes with temperature and other atmospheric physics.” Call it precision, not vagueness.

A second limitation follows from the same method, even with the age question settled. Turning an effective temperature into a mass requires an evolutionary model, and evolutionary models need the system’s age as an input. Gibbs et al.’s own paper used 23 million years, the ATMO 2020 isochrone named earlier, to reach its mass estimate. Whether Sutlieff and colleagues assumed the same age for their own mass estimate, their paper doesn’t say. As of this writing, no third paper has tried to settle the exact orbit either, so 24 AU, 26 AU and the beyond-30-AU stability case all remain live.

Where two methods leave this planet’s numbers

Two independently reviewed papers, published side by side in The Astrophysical Journal Letters, put a third giant planet around Beta Pictoris. One found it through the infrared spectrum Gibbs and colleagues pulled from James Webb data. The other found it through eleven years of archival imaging by Sutlieff, Bonse, Birkby and their collaborators. Their orbits overlap between roughly 19.9 and 28.2 AU. Temperature estimates agree even more closely, near 600 to 800 kelvin. What neither number resolves is Gibbs et al.’s separate case for an orbit beyond 30 AU. The age question is no longer split: Gibbs et al.’s own paper adopts 23 million years, and that is the figure the primary literature carries.

We wrote this one out in full because two good papers deserve to be read together, not filed under whichever headline got there first. Trace it or drop it. We hold every number here to that standard. Later observations will stretch the orbital arc both fits depend on, and we will come back when they do.

. Denise Meloni.

Sources

• Gibbs, A., Ruffio, J.-B., Bidot, A., Barman, T. S., Do Ó, C. R., Konopacky, Q. M., Perrin, M. D., Baburaj, A., Dacus, B., Macintosh, B., Madurowicz, A., and Xuan, J. W. (2026). Discovery of an Exterior Third Planet Orbiting Beta Pictoris. The Astrophysical Journal Letters, 1006, L11. Published 2026 July 15. https://doi.org/10.3847/2041-8213/ae801b
• Sutlieff, B., Bonse, M., Birkby, J., Christiaens, V., Biller, B., Parker, L., Iskandarli, L., et al. (2026). Direct Imaging Discovery of Giant Exoplanet Beta Pictoris d: A Decade-Long Game of Hide-and-Seek. The Astrophysical Journal Letters, 1006, L10. https://doi.org/10.3847/2041-8213/ae80a0
• Max Planck Institute for Astronomy, press release on the discovery of Beta Pictoris d (2026).
• University of Oxford, physics news release on the discovery of Beta Pictoris d (2026).
• Sky and Telescope, coverage of the Beta Pictoris d discovery (2026).
• exoplanet.eu catalog entry for Beta Pictoris d, orbital and physical parameters (2026).