Kral et al. (2026) investigate whether high-precision astrometry can reveal an exomoon around HD 206893 B, a brown-dwarf companion approximately 40.8 parsecs from Earth. A moon would cause its host to wobble around their shared center of mass, producing small deviations from the companion’s orbit around its star. The study presents the first dedicated application of this technique to a substellar companion. It also refines the orbits of HD 206893 B and the inner companion HD 206893 c, examines B’s atmosphere, and identifies promising targets for future searches. The authors used the VLTI/GRAVITY interferometer to measure relative positions with uncertainties typically of tens of microarcseconds. Their dataset spans 2019–2025 and includes eight closely spaced observing epochs between April and July 2024, designed to probe satellite periods ranging from days to months. They combined relative astrometry, Hipparcos–Gaia absolute astrometry, and stellar radial velocities to fit the orbits of B and c. After accounting for perturbations from c, they fitted a hierarchical model comprising the star, B, and a possible satellite. The satellite’s orbit was assumed to be circular, with no prior restriction on its orientation. Simultaneously acquired K-band spectra, with a resolving power of approximately 4,000, were combined and analyzed using atmospheric models and molecular cross-correlation. Within the adopted model and priors, the astrometry places a 95% upper limit of 0.8 Jupiter masses on a possible satellite. However, the posterior distribution for the satellite’s orbital separation shows a prominent peak near 0.22 astronomical units, corresponding to a period of approximately 0.76 years, or 275–278 days. Restricting the analysis to separations of 0.21–0.26 astronomical units yields a candidate mass near 0.5 Jupiter masses, approximately nine Neptune masses, although the abstract reports an approximate value of 0.4 Jupiter masses. This interpretation implies a wobble radius of roughly 0.14–0.15 milliarcseconds. The candidate’s orbital orientation is moderately constrained and would indicate substantial misalignment with B’s orbit around the star. The evidence remains suggestive rather than conclusive. The median Bayesian information criterion favors the moon model by 7.4, below the authors’ threshold of 10 for decisive preference. Other orbital-period peaks remain possible, and unmodeled instrumental systematics could produce similar residuals. Although known systematic effects were estimated to be smaller than 10 microarcseconds, the statistical comparison does not account for unknown systematics. An additional planet orbiting the star could also mimic a moon-induced signal, making continued stellar radial-velocity monitoring valuable for resolving these ambiguities. The revised orbital solution places B at 10.75 ± 0.08 astronomical units, with an eccentricity of 0.069 and a dynamical mass of 19.5 Jupiter masses, with uncertainties of approximately 1.3–1.4 Jupiter masses. Companion c has a mass of 11.1 Jupiter masses, an orbital separation of 3.74 astronomical units, and an eccentricity of 0.283. Their mutual inclination is approximately 1.8 degrees, indicating a nearly coplanar architecture that is more dynamically favorable than previous solutions. Spectral cross-correlation detects water at a signal-to-noise ratio of approximately 4.16 but reveals no significant carbon monoxide signal. Atmospheric interpretations remain model-dependent: Exo-REM and ATMO favor temperatures of approximately 1,100–1,160 kelvin and relatively large radii, whereas BT-Settl favors a temperature of approximately 1,580 kelvin and a smaller radius. The authors conclude that astrometry provides a valuable complement to transit and radial-velocity searches, particularly for satellites on wider orbits. Under optimistic precision assumptions, Neptune-mass or smaller satellites could be detectable in favorable systems; however, this sensitivity has not been achieved for the present candidate. Monthly GRAVITY follow-up is needed to test the candidate’s reality. AF Lep b and Beta Pictoris b are identified as the strongest near-term targets for GRAVITY+ searches.
- Scientists watched a faraway brown dwarf, an object smaller than a star, to look for a moon.
- Tiny changes in its position suggest that a moon about half as heavy as Jupiter might circle it every 275 days.
- Scientists are not sure the moon is real, because another planet or problems with their tools could cause these changes.
- They also studied another nearby companion and found signs of water in the brown dwarf’s atmosphere.
- Scientists need to keep watching to check for the moon, and better tools might help them find smaller moons someday.
Definitions
- brown dwarf: An object in space that is bigger than a planet but cannot shine like a star.
- star: A huge, hot ball of gas that makes its own light.
- moon: An object in space that travels around a planet or another object smaller than a star.
- position: The place where something is.
- Jupiter: The biggest planet in our solar system.
- planet: A large, round object that travels around a star.
- companion: An object in space that travels around a star or another object with it.
- atmosphere: The layer of gas around an object in space.
A Tiny Wobble Could Point to a Moon Around a Distant Brown Dwarf
Finding a moon beyond our solar system is an extraordinary observational challenge. A satellite can be too faint to see directly, yet its gravity may leave a measurable signature: a small wobble in the motion of its host. A new study explores whether that wobble could reveal a massive moon around HD 206893 B, a brown-dwarf companion roughly 40.8 parsecs, or 133 light-years, from Earth.
In “Exomoon search with VLTI/GRAVITY around the substellar companion HD 206893 B,” Q. Kral, J. Wang, J. Kammerer, S. Lacour, M. Malin, T. Winterhalder et al. present the first dedicated application of this astrometric technique to a substellar companion. Their results reveal an intriguing candidate signal—but not a confirmed exomoon.
How to Find a Moon Without Seeing It
A moon and its host orbit their shared center of mass. Consequently, the host does not follow a perfectly smooth path around its star: it also makes a smaller motion driven by the satellite. Astrometry, the precise measurement of positions in the sky, can potentially detect that extra movement.
The researchers used the VLTI/GRAVITY interferometer to measure HD 206893 B’s position relative to its star, typically with uncertainties of tens of microarcseconds. Their observations span 2019–2025, including eight closely spaced epochs between April and July 2024. That concentrated observing campaign was designed to investigate satellite periods ranging from days to months.
The challenge is separating a possible moon’s influence from everything else affecting B’s position. The system also contains an inner companion, HD 206893 c, whose gravitational perturbations must be accounted for.
Building a Model of the System
The team combined relative astrometry with Hipparcos–Gaia absolute astrometry and stellar radial-velocity measurements to refine the orbits of B and c. After accounting for c’s perturbations, they fitted a hierarchical model containing the star, B, and a possible satellite. They assumed a circular satellite orbit but imposed no prior restriction on its orientation.
An Intriguing Signal, With Important Caveats
Within the adopted model and priors, the observations place a 95% upper limit of 0.8 Jupiter masses on a possible satellite. More intriguingly, the inferred satellite separation shows a prominent probability peak near 0.22 astronomical units, corresponding to an orbital period of about 275–278 days.
Restricting the analysis to separations of 0.21–0.26 astronomical units produces a candidate mass near 0.5 Jupiter masses—approximately nine Neptune masses. The paper’s abstract gives a somewhat lower approximate value of 0.4 Jupiter masses. Either estimate describes a substantial satellite, rather than an analogue of the smaller moons familiar from our solar system.
This interpretation would produce a wobble radius of roughly 0.14–0.15 milliarcseconds. The candidate’s orbital orientation is moderately constrained and would imply substantial misalignment with B’s orbit around the star.
However, the statistical evidence remains suggestive. The median Bayesian information criterion favors the moon model by 7.4, below the authors’ threshold of 10 for decisive preference. Other orbital-period peaks remain possible, so the data do not uniquely identify the proposed orbit.
Instrumental uncertainty is another concern. Known systematic effects were estimated to be smaller than 10 microarcseconds, but the statistical comparison does not include unknown systematics. An additional planet orbiting the star could also mimic a moon-induced signal. Continued stellar radial-velocity monitoring would help test that alternative.
A Sharper View of Both Companions
Even without confirming a moon, the study improves our picture of the system. The revised orbital solution places B at 10.75 ± 0.08 astronomical units, with an eccentricity of 0.069. Its dynamical mass is 19.5 Jupiter masses, with uncertainties of approximately 1.3–1.4 Jupiter masses.
The inner companion c has a mass of 11.1 Jupiter masses, an orbital separation of 3.74 astronomical units, and an eccentricity of 0.283. The two companions’ mutual inclination is only about 1.8 degrees. Their nearly coplanar arrangement is more dynamically favorable than previous orbital solutions.
Water in a Model-Dependent Atmosphere
GRAVITY also collected K-band spectra at a resolving power of approximately 4,000. Combining these observations, the researchers detected water through molecular cross-correlation at a signal-to-noise ratio of about 4.16. They found no significant carbon monoxide signal.
The broader atmospheric interpretation depends strongly on the models used. Exo-REM and ATMO favor temperatures around 1,100–1,160 kelvin and relatively large radii. BT-Settl instead favors about 1,580 kelvin and a smaller radius. These differences highlight the remaining uncertainty in translating the spectrum into physical properties.
What Comes Next?
Monthly GRAVITY observations are needed to determine whether the candidate wobble repeats consistently. The study also identifies AF Lep b and Beta Pictoris b as the strongest near-term targets for GRAVITY+ satellite searches.
Under optimistic precision assumptions, favorable systems could reveal Neptune-mass or smaller satellites, although that sensitivity has not been achieved for this candidate. For now, HD 206893 B offers a promising test of astrometry as a complement to transit and radial-velocity searches—especially for satellites on wider orbits.
Direct astrometric detection of exomoons remains unexplored. This study presents the first application of high-precision astrometry to search for exomoons around substellar companions. We investigate whether the orbital motion of the companion HD 206893 B exhibits astrometric residuals consistent with the gravitational influence of an exomoon or binary planet. Using the VLTI/GRAVITY instrument, we monitored the astrometric positions of HD 206893 B and c across both short (days to months) and long (yearly) timescales. This enabled us to isolate potential residual wobbles in the motion of component B attributable to an orbiting moon. Our analysis reveals tentative astrometric residuals in the HD 206893 B orbit. If interpreted as an exomoon signature, these residuals correspond to a candidate (HD 206893 B I) with an orbital period of approximately 0.76 years and a mass of $\sim$0.4 Jupiter masses. However, the origin of these residuals remains ambiguous and could be due to systematics. Complementing the astrometry, our analysis of GRAVITY $R=4000$ spectroscopy for HD 206893 B confirms a clear detection of water, but no CO is found using cross-correlation. We also find that AF Lep b, and $β$ Pic b are the best short-term candidates to look for moons with GRAVITY+. Our observations demonstrate the transformative potential of high-precision astrometry in the search for exomoons, and proves the feasibility of the technique to detect moons with masses lower than Jupiter and potentially down to less than Neptune in optimistic cases. Crucially, further high-precision astrometric observations with VLTI/GRAVITY are essential to verify the reality and nature of this signal and attempt this technique on a variety of planetary systems.
