Jupiter-Mass Binary Objects (JuMBOs)

• JuMBOs are candidate free-floating binary systems comprising planetary-mass components with separations of 28–384 au, discovered in Orion’s Trapezium Cluster.
• A lower-limit multiplicity fraction of at least 9% and unusually wide separations contrast with typical substellar binaries, questioning standard star-formation models.
• Proposed formation channels—dynamical ejection, photoerosion, and vortical collapse—highlight the need for spectroscopic confirmation to resolve their true nature.

Jupiter-mass Binary Objects (JuMBOs) are a proposed population of very low-mass, free-floating binary candidates in the inner Orion Nebula and Trapezium Cluster, identified in JWST near-infrared imaging as part of a broader sample of planetary-mass candidates extending the mass function down to $0.6\,M_{\rm Jup}$. Their defining peculiarity is the combination of planetary-mass components and wide projected separations of about $28$–$384$ au, together with a lower-limit multiplicity fraction of at least $9\%$, a regime that is difficult to accommodate within standard star-formation, brown-dwarf, or planet-ejection frameworks. At the same time, the status of the objects as bona fide planetary-mass binaries remains unsettled, and spectroscopy is still required to distinguish that interpretation from alternatives such as reddened background stars (Pearson et al., 2023, Rodriguez et al., 22 Apr 2025).

1. Discovery in the inner Orion Nebula

The observational basis for JuMBOs is a JWST/NIRCam survey of the inner Orion Nebula, specifically the Trapezium Cluster, covering $11' \times 7.5'$ on the sky, or about $1.2 \times 0.8$ pc at an adopted distance of $390$ pc. The observations were obtained in JWST Cycle 1 GTO programme 1256 over $34.9$ hours between 26 September and 2 October 2022, using 12 filters from F115W through F470N. The survey identified $540$ candidate planetary-mass objects (PMOs) with best-fit masses $\le 13\,M_{\rm Jup}$, including $28$0 showing CH$28$1 absorption and best-fit masses $28$2; the lightest candidate has a mass of $28$3 (Pearson et al., 2023).

Identification relied on photometric spectral-energy-distribution fitting rather than spectroscopy. The modeling assumed a fixed age of $28$4 Myr, a distance of $28$5 pc, and extinction over $28$6 to $28$7, and compared the photometry to the CBPD22/ATMO2020 CEQ, NEQ$28$8, and NEQ$28$9 grids, together with Baraffe et al. (2015) models for higher masses. The selection emphasized strong H$384$0O and CH$384$1 absorption signatures, exploiting the fact that very low-mass planetary objects have strongly non-blackbody near-infrared SEDs (Pearson et al., 2023).

Within that PMO sample, the original survey identified $384$2 systems with a binary companion within $384$3 arcsec, plus two visual triple systems. Later papers frequently refer to $384$4 JuMBOs, reflecting the same underlying sample framed as a count of binary systems rather than “40 systems plus two triples” (Pearson et al., 2023, Parker et al., 1 May 2025).

2. Population properties and why they are surprising

JuMBOs were initially interpreted as systems not associated with stars in which both components have giant-planet-like masses. The reported component masses span roughly $384$5 to $384$6 in one summary and $384$7 to $384$8 in another, while the projected separations lie between about $384$9 and $9\%$0 au, or $9\%$1 and $9\%$2 au depending on the compilation. These separations are unusually wide relative to known substellar binaries: typical substellar binary separations are usually below $9\%$3 au, and brown dwarf–brown dwarf binaries in the field peak near $9\%$4 au (Rodriguez et al., 2024, Parker et al., 1 May 2025).

The multiplicity statistics are a central part of the anomaly. Among $9\%$5 PMO candidates, the lower-limit multiplicity fraction is at least $9\%$6. In the Trapezium comparison assembled by Pearson and McCaughrean, the wide binary fraction is about $9\%$7–$9\%$8 for massive stars, declines to about $9\%$9 for brown dwarfs in the Trapezium, and then rises again to at least $11' \times 7.5'$0 for PMOs. This breaks the otherwise familiar monotonic decline of multiplicity with decreasing primary mass (Pearson et al., 2023).

The systems are also not predominantly extreme equal-mass pairs. The mean mass ratio is $11' \times 7.5'$1, and only $11' \times 7.5'$2 of systems have $11' \times 7.5'$3, which is weaker than the strong equal-mass preference often reported for brown-dwarf binaries. Representative catalogued systems span a broad range, from JuMBO 1 at $11' \times 7.5'$4 au to JuMBO 24 at $11' \times 7.5'$5 au, and include higher-order architectures such as JuMBO 25 and JuMBO 42 with tertiary components listed (Pearson et al., 2023).

The survey also argued against simple contamination explanations. Expected foreground T-dwarf contamination across the entire mosaic is $11' \times 7.5'$6 objects, and the expected number of random chance alignments within $11' \times 7.5'$7 arcsec is $11' \times 7.5'$8. Both components of the candidate binaries exhibit PMO-like SEDs with H$11' \times 7.5'$9O and CH$1.2 \times 0.8$0 signatures, which was taken to disfavor ordinary reddened background stars as a blanket explanation for the original sample (Pearson et al., 2023).

3. Taxonomy, membership, and the planetary-mass controversy

A persistent misconception is that JuMBOs are already established as free-floating planetary-mass binaries. In fact, their actual nature remains unresolved. The initial interpretation treated them as planetary-mass binaries found by JWST in the Orion Nebula Cluster, but later work emphasized that this interpretation is contested and that spectroscopy is needed to decide whether the sources are genuinely very low-mass Orion members or instead reddened background stars (Rodriguez et al., 22 Apr 2025).

That taxonomic uncertainty matters because many dynamical arguments depend sensitively on the mass scale. One analytical study of extremely soft binaries in dense clusters concluded that if the Trapezium candidates are truly binary planets, their observed approximately flat semi-major-axis distribution is difficult to reconcile with encounter-driven evolution; by contrast, if they are actually binary brown dwarfs, a continuous in-situ formation scenario with an initial minimal semi-major axis around $1.2 \times 0.8$1 AU and a formation rate of $1.2 \times 0.8$2 Myr$1.2 \times 0.8$3 can plausibly match the observed number of single objects, binary number, binary fraction, and semi-major-axis distribution (Wang et al., 2024).

A later critique sharpened the skepticism toward the planetary-binary interpretation in a different way. It argued that wide JuMBOs are extremely difficult to form and preserve in the Trapezium under the stripping-by-stellar-flyby scenario, and concluded that either a different formation channel is required or the observed JuMBOs require thorough confirmation (Zwart et al., 15 Jul 2025).

This suggests that “JuMBO” is currently best understood as an observational and theoretical label for a candidate population rather than a settled taxonomic class. The observational sample is real in the sense of resolved JWST sources, but whether those sources are planetary-mass binaries, more massive substellar binaries, or something else remains under active dispute.

4. Cluster dynamics and the problem of survival

Even if the candidates are genuine binaries in Orion, their survival in a dense star-forming region is nontrivial. Direct $1.2 \times 0.8$4-body simulations that place planet–planet binaries into dense, substructured, subvirial star-forming regions find substantial destruction on Myr timescales. In simulations with $1.2 \times 0.8$5, fractal dimension $1.2 \times 0.8$6, virial ratio $1.2 \times 0.8$7, and JuMBO-like separations, the binary fraction drops from $1.2 \times 0.8$8 to $1.2 \times 0.8$9 in $390$0 Myr for a dense case with $390$1, and from $390$2 to $390$3 in $390$4 Myr for a lower-density case with $390$5. The study summarized this as $390$6–$390$7 destruction on timescales of a few Myr, with wider binaries preferentially disrupted (Parker et al., 1 May 2025).

A related analytical treatment of extremely soft binaries derived an ionization cross-section appropriate to the extreme mass-ratio regime and found that the surviving large-$390$8 population should asymptote toward $390$9 or $34.9$0. In that framework, the observed roughly flat Trapezium distribution is difficult to reproduce for binary planets unless there is continuous formation with a favorable input distribution; the same study therefore found the brown-dwarf interpretation dynamically easier to sustain (Wang et al., 2024).

An alternative dynamical picture emphasizes widening rather than mere attrition. Analytical work combined with three-body simulations argued that critical stellar flybys can transform a primordial population of tight JuMBOs into a mixture of tight, wide, and ionized systems. Applied to the Trapezium, the preferred parameters were an initial semimajor axis $34.9$1–20 au and a density-weighted residence time $34.9$2 Myr pc$34.9$3, yielding wide products out to $34.9$4–400 au and a semimajor-axis distribution roughly $34.9$5 with median $34.9$6 au (Huang et al., 2024).

The dynamical literature therefore agrees on fragility but not on the dominant evolutionary channel. One class of models treats the wide pairs as primordial systems being eroded; another treats them as widened descendants of tighter primordial binaries; a third concludes that the observed wide binary-planet population is so soft that continual replenishment or reinterpretation is required.

5. Proposed formation channels

The formation problem begins with the masses themselves. Pearson and McCaughrean explicitly noted that the reported JuMBO masses extend down to $34.9$7, below the minimum masses usually associated with 3D fragmentation, 2D shocks, and opacity-limited fragmentation arguments. At the same time, a simple planet-formation-plus-ejection explanation has difficulty accounting for the survival of wide binary planets after expulsion from a circumstellar system (Pearson et al., 2023).

Several specific channels have therefore been proposed. One dynamical route invokes close stellar encounters with planetary systems containing two giant planets. Direct few-body calculations showed that if the planets are nearly aligned at closest approach, a passing star can eject both while leaving them bound to each other. In that scenario, the JuMBO semimajor axis typically peaks near $34.9$8, and the eccentricity distribution is high and superthermal, with typical values around $34.9$9–$540$0. The formation efficiency can rise to a few percent in dense clusters for wide planetary systems, but the dependence on geometry and environment is strong (Wang et al., 2023).

That ejection picture remains controversial once subsequent cluster evolution is included. One critique argued that wide JuMBOs formed in this way are too soft to survive in the Trapezium, quoting $540$1 au$540$2 for a $540$3 au binary of two $540$4 objects and an ionization timescale $540$5 kyr at $540$6 and $540$7. In that reading, formation in isolated scattering experiments does not translate into a steady observed population in a dense young cluster (Zwart et al., 15 Jul 2025).

A distinct in-situ alternative is photoerosion of fragmenting prestellar cores by Lyman continuum radiation from nearby massive stars. In that model, a core begins fragmenting into a binary or higher-order system, but photoerosion strips the envelope and suppresses later accretion, reducing the final mass into the JuMBO range while leaving the primordial separation largely intact. For $540$8 and $540$9, the final masses lie comfortably in the observed JuMBO range, and the wide separations of $\le 13\,M_{\rm Jup}$0–$\le 13\,M_{\rm Jup}$1 au are interpreted as inherited from the separations of more massive binaries that would have formed from the same cores without photoerosion (Diamond et al., 2024).

More speculative proposals address the angular-momentum problem directly. One paper argued that wide JuMBOs may arise in vortical collapse if angular-momentum removal is aided by cosmic ray viscosity, with the collapse condition

$\le 13\,M_{\rm Jup}$2

In that picture, viscosity is more effective in smaller, less massive collapsing regions, potentially explaining why the wide-binary fraction could rise toward planetary masses; for a JuMBO-forming region with $\le 13\,M_{\rm Jup}$3, $\le 13\,M_{\rm Jup}$4, $\le 13\,M_{\rm Jup}$5, and $\le 13\,M_{\rm Jup}$6, the paper estimated $\le 13\,M_{\rm Jup}$7 (Katz, 2023).

A further line of work shifts attention from bound binaries to the production of single free-floating planets. Three-dimensional simulations of disc fragmentation in very young stellar binaries found that Jupiter-mass fragments formed at tens to $\le 13\,M_{\rm Jup}$8 au are usually ejected by chaotic migration and gravitational slingshot encounters with the growing secondary star. The ejection fraction reaches $\le 13\,M_{\rm Jup}$9 for $28$00, and the ejecta have $28$01. This does not readily explain stable wide JuMBOs, but it does provide a route to the broader free-floating Jupiter-mass population with which JuMBOs are observationally associated (Ćalović et al., 20 Nov 2025).

6. JuMBO 24 as the benchmark system

JuMBO 24 occupies a special place because it is the most massive and closest-separation case in the sample, with a total mass of about $28$02 and a projected separation of about $28$03 au, and because it is the only JuMBO previously detected in the radio continuum. That makes it the best current target for testing whether a radio source is physically associated with a JWST JuMBO candidate rather than being an unrelated background object (Rodriguez et al., 22 Apr 2025).

Archival VLA observations established the first radio counterpart to a JuMBO at 6.1 and 10.0 GHz. The source is coincident with JuMBO 24 to within about $28$04 arcsec and has measured flux densities of $28$05 at 6.1 GHz in 2012, $28$06 at 10.0 GHz in 2018, and $28$07 at 6.1 GHz in 2022. Across five individual 2012 epochs, the fluxes were all consistent within uncertainties with a steady value of $28$08, and no circular polarization was detected. The radio source was marginally resolved, with deconvolved size $28$09, aligned broadly with the JWST infrared elongation, which was interpreted as evidence that the emission may come from both components. The weighted-mean spectral index was $28$10, excluding a Rayleigh-Jeans optically thick thermal blackbody spectrum with $28$11 and the very steep negative indices typical of pulsars with $28$12, while the radio luminosity was $28$13 (Rodriguez et al., 2024).

New radio observations strengthened the kinematic constraint. Six VLA A-configuration epochs at 10 GHz in late 2024 and early 2025 detected JuMBO 24 in all epochs at a flux consistently around $28$14Jy, while an HSA observation at 5 GHz on 2024 January 22 yielded no long-baseline detection, giving a $28$15 upper limit of $28$16Jy on the recovered flux density. Combining earlier radio positions with the new VLA positions gave

$28$17

with no significant proper motion at the $28$18 level of about $28$19. Using

$28$20

and $28$21 pc, this corresponds to an upper limit of approximately $28$22 in the plane of the sky (Rodriguez et al., 22 Apr 2025).

The interpretation remains cautious. The low transverse speed is more compatible with a stationary, star-like origin than with a strongly ejected object, and it is also consistent with scenarios in which JuMBOs form in place, such as photoerosion of prestellar cores. The radio emission mechanism remains uncertain, but the combination of steady $28$23Jy flux, lack of strong variability, non-detection on long HSA baselines, and absence of detectable circular polarization at the level of about $28$24 does not favor a non-thermal origin (Rodriguez et al., 22 Apr 2025).

In that sense, JuMBO 24 does not resolve the JuMBO problem, but it narrows it. The system strengthens the case that at least one candidate is a compact, stationary Orion source with radio emission plausibly associated with both components, while leaving open the larger question of whether the entire class consists of planetary-mass binaries, more massive substellar binaries, or a heterogeneous mixture.