Rogue planets may originate fromplanetary systems in which they are formed and later ejected, or they can also form on their own, outside a planetary system. TheMilky Way alone may have billions to trillions of rogue planets, a range the upcomingNancy Grace Roman Space Telescope is expected to refine.[5][6] The odds of a rogue planet entering the solar system, much less posing a direct threat to life on Earth are slim to none with the odds being about one in one trillion within the next 1,000 years.[7]
Some planetary-mass objects may have formed in a similar way to stars, and theInternational Astronomical Union has proposed that such objects be calledsub-brown dwarfs.[8] A possible example isCha 110913−773444, which may either have been ejected and become a rogue planet or formed on its own to become a sub-brown dwarf.[9]
The two first discovery papers use the names isolated planetary-mass objects (iPMO)[10] and free-floating planets (FFP).[11] Most astronomical papers use one of these terms.[12][13][14] The term rogue planet is more often used for microlensing studies, which also often uses the term FFP.[15][16] Apress release intended for the public might use an alternative name. The discovery of at least 70 FFPs in 2021, for example, used the terms rogue planet,[17] starless planet,[18] wandering planet[19] and free-floating planet[20] in different press releases.
Isolated planetary-mass objects (iPMO) were first discovered in 2000 by theUK team Lucas & Roche withUKIRT in theOrion Nebula.[11] In the same year theSpanish team Zapatero Osorio et al. discovered iPMOs withKeck spectroscopy in theσ Orionis cluster.[10] The spectroscopy of the objects in the Orion Nebula was published in 2001.[21] BothEuropean teams are now recognized for their quasi-simultaneous discoveries.[22] In 1999 theJapanese team Oasaet al. discovered objects inChamaeleon I[23] that were spectroscopically confirmed years later in 2004 by theUS team Luhman et al.[24]
Astrophysicist Takahiro Sumi of Osaka University in Japan and colleagues, who form theMicrolensing Observations in Astrophysics and theOptical Gravitational Lensing Experiment collaborations, published their study ofmicrolensing in 2011. They observed 50 million stars in the Milky Way by using the 1.8-metre (5 ft 11 in) MOA-II telescope at New Zealand'sMount John Observatory and the 1.3-metre (4 ft 3 in)University of Warsaw telescope at Chile'sLas Campanas Observatory. They found 474 incidents of microlensing, ten of which were brief enough to be planets of around Jupiter's size with no associated star in the immediate vicinity. The researchers estimated from their observations that there are nearly two Jupiter-mass rogue planets for every star in the Milky Way.[25][26][27] One study suggested a much larger number, up to 100,000 times more rogue planets than stars in the Milky Way, though this study encompassed hypothetical objects much smaller than Jupiter.[28] A 2017 study by Przemek Mróz of Warsaw University Observatory and colleagues, with six times larger statistics than the 2011 study, indicates an upper limit on Jupiter-mass free-floating or wide-orbit planets of 0.25 planets per main-sequence star in the Milky Way.[29]
The cold planetary-mass object WISE J0830+2837 (marked orange object) observed with theSpitzer Space Telescope. It has a temperature of 300-350K (27-77°C; 80-170°F).
Microlensing planets can only be studied by the microlensing event, which makes the characterization of the planet difficult. Astronomers therefore turn to isolated planetary-mass objects (iPMO) that were found via thedirect imaging method. To determine a mass of abrown dwarf or iPMO one needs for example the luminosity and the age of an object.[32] Determining the age of a low-mass object has proven to be difficult. It is no surprise that the vast majority of iPMOs are found inside young nearbystar-forming regions of which astronomers know their age. These objects are younger than 200 Myrs, are massive (>5MJ)[4] and belong to theL- andT-dwarfs.[33][34] There is however a small growing sample of cold and oldY-dwarfs that have estimated masses of 8-20MJ.[35] Nearby rogue planet candidates of spectral type Y includeWISE 0855−0714 at a distance of7.27±0.13 light-years.[36] If this sample of Y-dwarfs can be characterized with more accurate measurements or if a way to better characterize their ages can be found, the number of old and cold iPMOs will likely increase significantly.
The first iPMOs were discovered in the early 2000s via direct imaging inside young star-forming regions.[37][10][21] These iPMOs found via direct imaging formed probably like stars (sometimes called sub-brown dwarf). There might be iPMOs that form like a planet, which are then ejected. These objects will however bekinematically different from their natal star-forming region, should not be surrounded by acircumstellar disk and have highmetallicity.[22] None of the iPMOs found inside young star-forming regions show a high velocity compared to their star-forming region. For old iPMOs the coldWISE J0830+2837[38] shows a Vtan of about 100 km/s, which is high, but still consistent with formation in our galaxy. ForWISE 1534–1043[39] one alternative scenario explains this object as an ejected exoplanet due to its high Vtan of about 200 km/s, but its color suggests it is an old metal-poor brown dwarf. Most astronomers studying massive iPMOs believe that they represent the low-mass end of the star-formation process.[22]
Astronomers have used theHerschel Space Observatory and theVery Large Telescope to observe a very young free-floating planetary-mass object,OTS 44, and demonstrate that the processes characterizing the canonical star-like mode of formation apply to isolated objects down to a few Jupiter masses. Herschelfar-infrared observations have shown that OTS 44 is surrounded by a disk of at least 10 Earth masses and thus could eventually form a large satellite system.[40] Spectroscopic observations of OTS 44 with the SINFONI spectrograph at the Very Large Telescope have revealed that the disk is actively accreting matter, similar to the disks of young stars.[40]
In theOrion Nebula a population of 40 wide binaries and 2 triple systems were discovered. The discovery was surprising for two reasons: the trend ofbinaries of brown dwarfs predicted a decrease of distance between low mass objects with decreasing mass. It was also predicted that the binary fraction decreases with mass. These binaries were namedJupiter-mass Binary Objects (JuMBOs); they make up at least 9% of the iPMOs and have a separation smaller than 340AU.[44] It is unclear how these JuMBOs formed, but an extensive study argued that they formed in situ, like stars.[45] If they formed like stars, then there must be an unknown "extra ingredient" to allow them to form. If they formed like planets and were later ejected, then it has to be explained why these binaries did not break apart during the ejection process. Future measurements with JWST might resolve if these objects formed as ejected planets or as stars.[44]Kevin Luhman reanalysed theNIRCam data and found that most JuMBOs did not appear in his sample of substellar objects. Moreover, the color was consistent with reddened background sources or low signal-to-noise sources. He considers onlyJuMBO 29 as a good candidate for a binary planetary-mass system.[46]
There are likely hundreds[47][44] of known candidate iPMOs, over a hundred[48][49][50] objects with spectra and a small but growing number of candidates discovered via microlensing. Some large surveys include:
As of December 2021, the largest-ever group of rogue planets was discovered, numbering at least 70 and up to 170 depending on the assumed age. They are found in theOB association betweenUpper Scorpius andOphiuchus with masses between 4 and 13MJ and age around 3 to 10 million years, and were most likely formed by eithergravitational collapse of gas clouds, or formation in aprotoplanetary disk followed by ejection due todynamical instabilities.[47][17][51][19] Follow-up observations with spectroscopy from theSubaru Telescope andGran Telescopio Canarias showed that the contamination of this sample is quite low (≤6%). The 16 young objects had a mass between 3 and 14MJ, confirming that they are indeed planetary-mass objects.[50]
In October 2023, an even larger group of 540 planetary-mass object candidates was discovered in the Trapezium Cluster and inner Orion Nebula with JWST. The objects have a mass between 13 and 0.6MJ. A surprising number of these objects formed wide binaries, which was not predicted.[44]
There are in general two scenarios that can lead to the formation of an isolated planetary-mass object (iPMO). It can form like a planet around a star and is then ejected, or it forms like a low-mass star or brown dwarf in isolation. This can influence its composition and motion.[22]
Recent research indicates that rogue planets may form both through direct gravitational collapse within stellar nurseries and through ejection from their natal planetary systems, later interacting with established systems and influencing their orbital architectures and overall demographics. Many of these objects likely originated within planetary systems before being dynamically expelled, while others may have formed in isolation. Besides altering system stability during close encounters or possible capture events, rogue planets can also deliver volatiles that enhance prebiotic chemistry and create conditions conducive to increased biological diversity. These combined formation, dynamical, biochemical, and ecological effects play a significant role in shaping the distribution and evolution of exoplanetary systems.[52]
Objects with a mass of at least oneJupiter mass were thought to be able to form via collapse and fragmentation ofmolecular clouds from models in 2001.[53] Pre-JWST observations have shown that objects below 3-5MJ are unlikely to form on their own.[4] Observations in 2023 in theTrapezium Cluster with JWST have shown that objects as massive as 0.6MJ might form on their own, not requiring a steep cut-off mass.[44] A particular type ofglobule, calledglobulettes, are thought to be birthplaces for brown dwarfs and planetary-mass objects. Globulettes are found in theRosette Nebula andIC 1805.[54] Sometimes young iPMOs are still surrounded by a disk that could formexomoons. Due to the tight orbit of this type of exomoon around their host planet, they have a high chance of 10-15% to betransiting.[55]
Some very young star-forming regions, typically younger than 5 million years, sometimes contain isolated planetary-mass objects with infrared excess and signs ofaccretion. Most well known is the iPMOOTS 44 discovered to have adisk and being located inChamaeleon I. Chamaeleon I and II have other candidate iPMOs with disks.[56][57][33] Other star-forming regions with iPMOs with disks or accretion are Lupus I,[57]Rho Ophiuchi Cloud Complex,[58] Sigma Orionis cluster,[59] Orion Nebula,[60]Taurus,[58][61]NGC 1333[62] andIC 348.[63] A large survey of disks around brown dwarfs and iPMOs withALMA found that these disks are not massive enough to formearth-mass planets. There is still the possibility that the disks already have formed planets.[58] Studies ofred dwarfs have shown that some have gas-rich disks at a relative old age. These disks were dubbedPeter Pan Disks and this trend could continue into the planetary-mass regime. One Peter Pan disk is the 45 Myr old brown dwarf2MASS J02265658-5327032 with a mass of about 13.7MJ, which is close to the planetary-mass regime.[64] Recent studies of the nearby planetary-mass object2MASS J11151597+1937266 found that this nearby iPMO is surrounded by a disk. It shows signs of accretion from the disk and also infrared excess.[65] In May 2025 researchers using JWST found that the disk aroundCha 1107−7626 containshydrocarbons. Cha 1107−7626 (6-10MJ) is one of the lowest-mass objects with a dusty disk.[66] Additional JWST spectroscopy did show that silicates and hydrocarbons are a common feature in disks of planetary-mass objects. The disks showed strong evidence of grain growth and crystallization, similar to what is seen in disks around brown dwarfs and stars. This showed that these disks are capable to formrocky companions.[67]
Ejected planets are predicted to be mostly low-mass (<30M🜨 Figure 1 Ma et al.)[68] and their mean mass depends on the mass of their host star. Simulations by Ma et al.[68] did show that 17.5% of 1M☉ stars eject a total of 16.8M🜨 per star with a typical (median) mass of 0.8M🜨 for an individual free-floating planet (FFP). For lower massred dwarfs with a mass of 0.3M☉ 12% of stars eject a total of 5.1M🜨 per star with a typical mass of 0.3M🜨 for an individual FFP.
Hong et al.[69] predicted that exomoons can be scattered by planet-planet interactions and become ejected exomoons. Higher mass (0.3-1MJ) ejected FFP are predicted to be possible, but they are also predicted to be rare.[68] Ejection of a planet can occur via planet-planet scatter or due a stellar flyby. Another possibility is the ejection of a fragment of a disk that then forms into a planetary-mass object.[70] Another suggested scenario is the ejection of planets in a tiltedcircumbinary orbit. Interactions with the central binary and the planets with each other can lead to the ejection of the lower-mass planet in the system.[71][72] Although the effectiveness of this mechanism depends on the encounter geometry, which is not well constrained yet both observationally and theoretically
Formation via encounters between young circumstellar disks
Encounters between young circumstellar disks, which are marginally gravitationally stable, can produce elongated tidal bridges that collapse locally to form iPMOs.[73] These iPMOs host expansive disks similar to observations,[60] which the ejected planet hyperthesis can hardly explain. They also have a high multiplicity fraction in their formation, as suggested by iPMOs in the Trapezium cluster.[44] Although the effectiveness of this mechanism depends on the encounter geometry, which is not well constrained yet both observationally and theoretically.[74]
If a stellar or brown dwarf embryo experiences a halted accretion, it could remain low-mass enough to become a planetary-mass object. Such a halted accretion could occur if the embryo is ejected or if its circumstellar disk experiencesphotoevaporation nearO-stars. Objects that formed via the ejected embryo scenario would have smaller or no disk and the fraction of binaries decreases for such objects. It could also be that free-floating planetary-mass objects form from a combination of scenarios.[70]
Most isolated planetary-mass objects will float in interstellar space forever.
Some iPMOs will have a close encounter with aplanetary system. This rare encounter can have three outcomes: The iPMO will remain unbound, it could be weakly bound to the star, or it could "kick out" the exoplanet, replacing it. Simulations have shown that the vast majority of these encounters result in a capture event with the iPMO being weakly bound with a lowgravitational binding energy and an elongated highlyeccentric orbit. These orbits are not stable and 90% of these objects gain energy due to planet-planet encounters and are ejected back into interstellar space. Only 1% of all stars will experience this temporary capture.[75]
Interstellar planets generate little heat and are not heated by a star.[76] However, in 1998,David J. Stevenson theorized that some planet-sized objects adrift in interstellar space might sustain a thick atmosphere that would not freeze out. He proposed that these atmospheres would be preserved by the pressure-induced far-infrared radiation opacity of a thickhydrogen-containing atmosphere.[77]
During planetary-system formation, several small protoplanetary bodies may be ejected from the system.[78] An ejected body would receive less of the stellar-generatedultraviolet light that can strip away the lighter elements of its atmosphere. Even an Earth-sized body would have enough gravity to prevent the escape of the hydrogen and helium in its atmosphere.[77] In an Earth-sized object thegeothermal energy from residual core radioisotope decay could maintain a surface temperature above themelting point of water,[77] allowing liquid-water oceans to exist. These planets are likely to remain geologically active for long periods. If they have geodynamo-created protectivemagnetospheres and sea floor volcanism,hydrothermal vents could provide energy for life.[77] These bodies would be difficult to detect because of their weak thermal microwave radiation emissions, although reflected solar radiation andfar-infrared thermal emissions may be detectable from an object that is less than 1,000astronomical units from Earth.[79] Around five percent of Earth-sized ejected planets with Moon-sizednatural satellites would retain their satellites after ejection. A large satellite would be a source of significant geologicaltidal heating.[80]
The table below lists rogue planets, confirmed or suspected, that have been discovered. It is yet unknown whether these planets were ejected from orbiting a star or else formed on their own assub-brown dwarfs. Whether exceptionally low-mass rogue planets (such asOGLE-2012-BLG-1323 andKMT-2019-BLG-2073) are even capable of being formed on their own is currently unknown.
Age uncertain, but old due to solar vicinity object;[94] candidate even for an old age of 12 Gyrs (age of the universe is 13.8 Gyrs). Closest known probable rogue planet
These objects were discovered viamicrolensing. Rogue planets discovered via microlensing can only be studied by the lensing event. Some of them could also be exoplanets in a wide orbit around an unseen star.[110]
CandidateALMA detection; although the object's brightness and proximity is consistent with it being the same object that eclipsed the starV1400 Centauri in 2007, follow-up observations by ALMA are needed to confirm whether it is moving, let alone in the right direction.[123]
A Pail of Air (1951) — a science fiction short story byFritz Leiber whereEarth is pulled out of the Solar System by ablack hole. Although the Earth is explicitly stated to orbit the black hole, the net effect is the same as ejecting it out of the Solar System as a rogue planet.[124][125]
Metroid Prime 2: Echoes (2004) - video game set in a rogue planet, whose ecosystem is sustained by its natural energy, called by its inhabitants as "Light of Aether".
Remina (2004–2005) – horror manga byJunji Ito featuring a sentient rogue planet capable of eating planets and stars
Melancholia (2011) – science fiction film byLars von Trier about a rogue planet on a collision course with Earth
Gemini Home Entertainment (2019–present) – horror anthology web series by Remy Abode, the main antagonist of which is a sentient rogue planet named "the Iris" that is masterminding an invasion of the Solar System, particularly Earth andNeptune[127]
^abcdefPearson, Samuel G.; McCaughrean, Mark J. (2 October 2023). "Jupiter Mass Binary Objects in the Trapezium Cluster". p. 24.arXiv:2310.01231 [astro-ph.EP].
^abcdefghijkSanghi, Aniket; Liu, Michael C.; Best, William M.; Dupuy, Trent J.; Siverd, Robert J.; Zhang, Zhoujian; Hurt, Spencer A.; Magnier, Eugene A.; Aller, Kimberly M.; Deacon, Niall R. (6 September 2023). "The Hawaii Infrared Parallax Program. VI. The Fundamental Properties of 1000+ Ultracool Dwarfs and Planetary-mass Objects Using Optical to Mid-IR SEDs and Comparison to BT-Settl and ATMO 2020 Model Atmospheres".arXiv:2309.03082 [astro-ph.SR].
