A planet does not have to shine to be detected. It does not even have to orbit a star.
Free-floating planets, sometimes called rogue planets, travel through the Milky Way without being gravitationally bound to a host star. That makes them extraordinarily difficult to find. There is no parent star for the planet to transit, no stellar wobble to measure, and an old, low-mass planet may give off so little light that direct imaging is impractical.
Gravitational microlensing gets around that problem in an elegant way: instead of looking for light from the planet, astronomers look for what the planet’s gravity does to light coming from a much more distant star.
Why free-floating planets are so difficult to see
Most familiar exoplanet-detection techniques depend in some way on a planet’s relationship with its star. The transit method detects the tiny decrease in starlight when a planet crosses its star. Radial-velocity measurements detect the star moving slightly in response to an orbiting planet’s gravity.
A genuinely free-floating planet provides neither signal.
Young giant planets can sometimes be bright enough in infrared wavelengths to image directly because they still retain heat from their formation. Older or much smaller isolated planets are another matter. They can be extremely faint against the darkness of interstellar space.
Microlensing is different because it depends primarily on mass and alignment rather than emitted light. An otherwise invisible object can therefore announce its presence through gravity.
How gravitational microlensing works
Einstein’s general theory of relativity tells us that mass curves spacetime. Light travelling through that curved spacetime follows a bent path.
Imagine three objects that happen to line up from our point of view:
- a distant background star, called the source;
- a closer object, called the lens;
- and a telescope observing from Earth or space.
If the foreground object passes sufficiently close to the line of sight toward the background star, its gravity bends some of the star’s light toward us. The foreground object effectively becomes a natural magnifying lens.
The telescope usually cannot resolve the multiple tiny images produced by the lensing geometry. Instead, astronomers measure something much easier to detect: the background star temporarily becomes brighter.
NASA’s microlensing overview describes this temporary amplification as the observational signature that allows an intervening object to be detected even when the lens itself cannot be seen.
A free-floating planet creates its own brief flash
If the foreground lens is a lone planet rather than a star, the same physics applies. The planet’s gravity bends and magnifies the background star’s light as the alignment improves. The star brightens, reaches a maximum magnification, and then returns toward its normal brightness as the planet continues moving across the sky.
The planet itself has not flashed or emitted a burst of energy. The changing brightness belongs to the distant source star. The planet has simply acted as a temporary cosmic magnifying glass.
This is particularly useful for finding low-mass isolated objects because microlensing does not require the lens to be luminous.
Why short microlensing events can point to planets
One important quantity measured during a microlensing event is the Einstein timescale, usually written as tE. Roughly speaking, it describes how long it takes the source and lens to move across the characteristic angular scale of the event.
Lower-mass lenses generally have smaller Einstein radii and therefore tend to produce shorter events. A stellar lens may produce a signal lasting many days or weeks, while planetary-mass lenses can create events lasting days, hours, or in extreme cases less than an hour.
But duration is not a weighing scale by itself. The event timescale also depends on the distances of the source and lens and on their relative motion across the sky. A very short event is therefore evidence of a low-mass lens, not an automatic measurement of the planet’s mass.
An extreme example lasted only about 42 minutes
A striking demonstration came from the event OGLE-2016-BLG-1928. Researchers reported an Einstein crossing time of about 0.0288 day, or approximately 41.5 minutes, making it one of the shortest microlensing events identified at the time.
The team also detected a finite-source effect, allowing an angular Einstein radius of about 0.842 microarcsecond to be measured. Depending on the unknown distance to the lens, the object could have had a mass in roughly the Mars-to-Earth range. The investigators found no evidence for a stellar companion out to a projected separation of about 8 astronomical units, making the object a strong terrestrial-mass free-floating-planet candidate rather than proof beyond all possible ambiguity.
The details were published in The Astrophysical Journal Letters study of OGLE-2016-BLG-1928.
How astronomers get more than an event duration
Finding a short brightening event is only the beginning. Astronomers want to determine what produced it, and that requires extracting additional information from the light curve whenever possible.
Finite-source effects
The simplest microlensing model treats the background star as a point of light. During an extremely close alignment, however, the apparent angular size of the star itself becomes important. Different portions of the stellar disk are magnified differently as the lens passes across or very near it.
These distortions in the light curve are called finite-source effects. If astronomers can estimate the angular size of the background star, the finite-source signature can help them determine the angular Einstein radius, written as θE.
That additional measurement significantly improves what can be learned about the lens.
Microlensing parallax
Another valuable measurement is microlensing parallax. Observers at sufficiently different locations can see slightly different versions of the same alignment, just as viewing a nearby object first with one eye and then the other changes its apparent position against a distant background.
If both the angular Einstein radius and microlensing parallax can be determined, astronomers can derive the lens mass much more directly. Without both pieces of information, researchers may instead have to use a Galactic population model and Bayesian analysis to estimate the most probable mass and distance.
A second Earth-mass candidate shows what surveys can find
Analysis of nine years of observations from the Microlensing Observations in Astrophysics, or MOA, survey produced another notable candidate known as MOA-9y-5919.
Its measured Einstein timescale was about 0.057 day, roughly 82 minutes, and its angular Einstein radius was only about 0.90 microarcsecond. A Bayesian analysis placed the lens near an Earth mass, with a broad uncertainty range.
The wider importance of the study was not simply one unusual object. Researchers analysed thousands of microlensing events and modelled the survey’s ability to detect extremely short signals, which is essential when estimating how common low-mass free-floating planets may be. The results are described in the 2023 Astronomical Journal study of terrestrial- and Neptune-mass free-floating-planet candidates.
Does microlensing prove that a planet has no star?
Not always, and this is one of the most important qualifications in free-floating-planet research.
A short planetary microlensing event may show no detectable contribution from a host star. That makes an isolated planet a plausible explanation. But a planet on an extremely wide orbit around a distant host could sometimes produce a similarly isolated-looking signal if the star is far enough away in projected separation.
Researchers therefore often use the term free-floating-planet candidate. They can search the light curve for evidence of a stellar companion and may use later high-resolution observations to look for a possible host, but the absence of a detected star does not automatically prove that no star exists anywhere nearby.
For population studies, scientists may consequently discuss free-floating or very wide-orbit planets when the observations cannot completely distinguish the two possibilities.
Why astronomers watch the crowded centre of the Milky Way
Microlensing requires a chance alignment. Individual events cannot be predicted simply by choosing one nearby planet and waiting for it to cross a convenient star.
The practical solution is statistical: monitor enormous numbers of stars in extremely crowded regions of the sky and look for transient brightenings.
This is why major microlensing projects such as OGLE, MOA and the Korea Microlensing Telescope Network have concentrated heavily on the Galactic bulge, where dense stellar fields provide huge numbers of potential background sources.
High observing cadence matters just as much as the number of stars. An Earth-mass or sub-Earth-mass lens may create a signal that develops and disappears between widely spaced observations. Missing a few hours can mean missing the planet.
The Roman Space Telescope could transform the search
NASA’s Nancy Grace Roman Space Telescope adds a powerful space-based observatory to this effort. Roman launched on August 30, 2026 and is undergoing commissioning ahead of its science programme.
One of Roman’s major programmes is the Galactic Bulge Time-Domain Survey, designed to repeatedly observe crowded regions toward the centre of the Milky Way. Its wide field, infrared sensitivity, stable space-based observations and frequent measurements are particularly valuable for detecting short microlensing events.
NASA expects Roman’s microlensing work to discover large numbers of planets that are difficult for other techniques to reach, including planets on relatively wide orbits and objects apparently wandering without host stars. Roman’s Galactic Bulge Time-Domain Survey is therefore expected to turn what has so far been a sparse collection of intriguing candidates into a much more useful statistical sample.
As of October 2026, Roman is still in its commissioning phase rather than producing its full planetary census. NASA expects the first science images in early 2027.
Why finding rogue planets matters
A census of free-floating planets would answer questions that extend well beyond finding strange worlds in the dark.
Planetary systems can be dynamically violent while they are forming. Interactions among several planets, stars or other companions can throw a planet onto a drastically altered orbit or eject it from the system entirely. The number and masses of isolated planets therefore preserve clues about how frequently these gravitational upheavals occur.
Other planetary-mass objects may have formed in relative isolation through processes more closely resembling the collapse that produces stars and brown dwarfs. Determining the mass distribution of free-floating objects can help astronomers distinguish among different formation channels.
Recent MOA analyses indicate that low-mass free-floating or extremely wide-orbit planets may be numerous, but the precise Galactic population remains uncertain because detecting short events is difficult and survey selection effects must be modelled carefully.
Microlensing turns darkness into an advantage
Free-floating planets pose a peculiar observational problem: the planets astronomers most want to count may be the ones giving them almost no light to work with.
Microlensing changes the question. Astronomers do not ask whether the planet is bright enough to see. They ask whether its gravity can briefly alter the light of something behind it.
For a few hours, or sometimes only minutes, an otherwise invisible world can leave a measurable fingerprint in a distant star’s light curve. The event may never repeat, and determining the planet’s precise mass or proving the complete absence of a host star can remain difficult. Yet that fleeting brightening is enough to reveal an object that might otherwise wander through the Galaxy unseen.