Free-Floating Planetary-Mass Objects (FFPMOs)

• Free-floating planetary-mass objects (FFPMOs) are isolated bodies with masses below the deuterium-burning threshold (<13 MJup), situated between planets and brown dwarfs.
• They are identified via microlensing and deep infrared imaging in star-forming regions, covering a mass range from Mars-like to gas giant scales.
• Studying FFPMOs constrains formation models by informing mass functions, ejection mechanisms, and disk properties, key to understanding star and planet formation.

Free-floating planetary-mass objects (FFPMOs) are isolated, non-stellar bodies with masses below the deuterium-burning threshold (≲13 MJup) and not gravitationally bound to any host star. This population spans a diverse spectrum, from objects near the planetary–brown dwarf boundary (~10–13 MJup) down to masses comparable to Mars and Pluto. FFPMOs are now robustly detected across the Galaxy via microlensing and deep infrared imaging in young star-forming regions, with upcoming space missions poised to deliver a comprehensive census. Their study constrains initial conditions for star and planet formation, population synthesis models, and the dynamical histories of planetary systems.

1. Definition, Physical Properties, and Classification

A free-floating planetary-mass object is defined as an isolated (non-bound) object with $M<13\,M_{\mathrm{Jup}}$ (deuterium-burning limit) (Miret-Roig, 2023, Scholz et al., 2022). This incorporates diverse origins—direct collapse (“star-like”), ejection from planetary systems (“planet-like”), or formation in tidal debris—but is operationally agnostic to formation mechanism.

• Physical properties: FFPMOs occupy the mass and temperature continuum extending from young gas-giant exoplanets ($\sim$1–15 MJup, $T_\mathrm{eff}\sim$1300–2400 K, L- and T-type spectra at a few Myr) down to sub-Earth-mass objects detectable only through microlensing signatures.
• Population subtypes: Some recent taxonomies distinguish between “Wide”, “Kuiper”, and “Oort” FFPs—bound at extremely large separations (100–10⁴ au), but essentially indistinguishable from true FFPMOs in transient lensing events (Gould et al., 2024, Gould, 2016).

FFPMOs should be contrasted with brown dwarfs (13–75 MJup, no sustained H burning) and bound planetary-mass companions.

2. Detection Methods and Survey Capabilities

Micro-lensing is the primary extragalactic tool for all mass regimes, providing sensitivity down to $M\sim 0.1\,M_\oplus$ for well-designed high-cadence surveys. Space-based surveys (Roman, JWST/NIRCam/NIRISS, Euclid) and ground-based campaigns (KMTNet, OGLE) have redefined the FFPMO landscape (Johnson et al., 2020, Coleman et al., 2024, Scholz et al., 2022).

• Microlensing event timescales:

$$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$

with $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$ (Mars-mass) up to $30$ d ($M\sim100\,M_\oplus$) (Johnson et al., 2020).

• Detection criteria: FFPMO microlensing events are isolated, short-duration pulses with timescales $<1$ day (low-mass) to a few days (Neptune–Jupiter mass) (Gould, 2016). The Nancy Grace Roman Space Telescope will achieve sensitivity from Mars mass ($\sim$0.1 $\sim$0) up to gas-giant masses, with detection criteria requiring at least 6 points $\sim$1 above baseline and $\sim$2 relative to a flat model (Johnson et al., 2020, Gould et al., 2024).
• Finite-source effects: For $\sim$3, the lens Einstein angle $\sim$4 becomes smaller than the source angular radius, so the light curve is lengthened and peak magnification reduced, but still recoverable at 15 min cadence and 0.01 mag precision (Johnson et al., 2020, Gould et al., 2024).

Infrared direct imaging and spectroscopy enable detection and spectral classification in young clusters down to $\sim$55 $\sim$6 (Langeveld et al., 2024, Damian et al., 7 Jul 2025), with spectral energy distributions confirming effective temperatures, surface gravities, disks, and accretion signatures (e.g., Paschen-$\sim$7 emission, silicate features).

3. Mass Spectrum, Abundance, and Population Synthesis

Current microlensing-based mass functions for FFPMOs show a nontrivial, non-power-law structure reflecting multiple ejection and formation channels (Coleman et al., 2024, Guo et al., 5 Nov 2025). Key results:

• Mass function parameterizations:
  • Log-uniform: $\sim$8 planet/star/dex (sometimes assumed but disfavored at low masses).
  • Broken power-law (Cassan-inspired):

  $\sim$9

  (Johnson et al., 2020) - Multimodal population synthesis:

  $T_\mathrm{eff}\sim$0

  with $T_\mathrm{eff}\sim$1, $T_\mathrm{eff}\sim$2, $T_\mathrm{eff}\sim$3, $T_\mathrm{eff}\sim$4, $T_\mathrm{eff}\sim$5, $T_\mathrm{eff}\sim$6 (Coleman et al., 2024).

• Total abundance: Current predictions yield $T_\mathrm{eff}\sim$71.07–1.20 ejected planets per star in $T_\mathrm{eff}\sim$8 (Guo et al., 5 Nov 2025, Coleman et al., 2024), with a total mass budget $T_\mathrm{eff}\sim$9 per star (majority at Neptune mass or higher).

• Characteristic mass distribution: There is a pronounced peak at $M\sim 0.1\,M_\oplus$0, a trough near $M\sim 0.1\,M_\oplus$1, and a steep decline at $M\sim 0.1\,M_\oplus$2. The low-mass end ($M\sim 0.1\,M_\oplus$3) rises only shallowly, rejecting simplistic single power-law extrapolations for sub-Earth masses.

• Relation to binding energy: The fraction of true free-floaters versus "detached" (widely bound) objects is under active investigation (Hadden et al., 11 Jul 2025). Dynamical models indicate that up to half of Neptunian-mass FFPMOs detected by lensing may actually be on extremely wide, bound orbits.

4. Formation Channels and Evolutionary Scenarios

The observable FFPMO population is supplied by at least three distinct formation mechanisms, with relative weights depending on mass and environment (Miret-Roig, 2023, Guo et al., 5 Nov 2025, Fu et al., 2024):

• Planet–planet scattering and ejection: Instability and mutual scatterings in multi-giant-planet systems naturally populate interstellar space with Neptune-mass planets at wide orbits and a flat-to-declining spectrum toward lower masses. Close-in low-mass planets are generally retained, while wide-orbit Neptunes are ejected most efficiently. Simulations yield $M\sim 0.1\,M_\oplus$4/star in $M\sim 0.1\,M_\oplus$5 (Guo et al., 5 Nov 2025).

• Circumbinary scattering: In circumbinary disks, repeated close binary–planet interactions at the inner edge drive ejection with a peak at the pebble isolation mass ($M\sim 0.1\,M_\oplus$6), producing characteristic peaks and troughs in the FFP mass function (Coleman et al., 2024).

• Turbulent fragmentation (star-like mechanism): Direct fragmentation of molecular cloud cores, and possibly disk instability, produces FFPMOs extending the initial mass function (IMF) tail below the brown dwarf limit. Yet this mechanism alone generally underpredicts FFPMO abundance in young regions (Miret-Roig, 2023, Langeveld et al., 2024).

• Bridge/filament fragmentation in disk–disk collisions: In dense clusters, hydrodynamical simulations demonstrate that near-coplanar disk–disk encounters can trigger fragmentation of tidal bridges, producing planetary-mass objects and binaries independent from other channels (Fu et al., 2024).

• Empirical evidence for disks and accretion: The presence of massive, flared disks, active accretion (e.g., H$M\sim 0.1\,M_\oplus$7, Pa$M\sim 0.1\,M_\oplus$8), and scaling of accretion rate with central mass down to a few $M\sim 0.1\,M_\oplus$9 demonstrates continuity with the pre-main-sequence stellar pathway (Joergens et al., 2014, Seo et al., 27 Jan 2025, Damian et al., 7 Jul 2025).

5. Disk Properties, Multiplicity, and Atmospheric Signatures

Circumplanetary disks are ubiquitous among young FFPMOs with $$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$05–15 $$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$1, as revealed by infrared excess, silicate emission/absorption, and hydrocarbon lines (Seo et al., 27 Jan 2025, Damian et al., 7 Jul 2025, Langeveld et al., 2024).

• Disk fractions:

  • In clusters such as IC 348 and NGC 1333, disk fractions among FFPMOs are 40–50% at ages 1–5 Myr, comparable to brown dwarfs and low-mass stars. Disk lifetimes decline sharply after ∼5–10 Myr (Seo et al., 27 Jan 2025).
  • Disk masses (e.g., OTS 44): $$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$2 with $$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$3 (Joergens et al., 2014).
• Spectral properties: 1–13 $$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$4m spectra confirm strong silicate features, evidence for grain growth and crystallization, and emission from hydrocarbon molecules (CH$$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$5, C$$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$6H$$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$7) in disks around $$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$8–10 $$t_E = \frac{\theta_E}{\mu_{\rm rel}}, \quad \theta_E = \sqrt{\kappa M \pi_{\rm rel}}$$9 objects (Damian et al., 7 Jul 2025). Photospheric absorption by amorphous silicates is directly detected (Damian et al., 7 Jul 2025).
• Multiplicity:
  • Binary fraction among $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$0 FFPMOs is $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$1 at $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$2 au (1/55 in surveyed populations) (Bouy et al., 17 Jun 2025).
  • In Taurus (quiescent, low density), $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$3; in Upper Sco (higher density), $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$4, suggesting environmental suppression of wide binaries (Bouy et al., 17 Jun 2025).
  • Bridge/filament fragmentation simulations naturally yield binaries with $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$5–15 au, matching observed multiplicities (Fu et al., 2024).

6. Observational Challenges and Future Prospects

• Host discrimination: Short microlensing events cannot, in isolation, distinguish truly unbound FFPMOs from planets on wide ($t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$6 Einstein radii) orbits. AO imaging several years post-event can resolve hosts at Kuiper/Oort-scale separations ($t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$7 au) with ELTs (Gould, 2016, Gould et al., 2024).
• Detection efficiency for lowest-mass FFPMOs is severely limited by cadence; e.g., in Roman, detection of Pluto–Mars-mass objects requires at least six data points above baseline, motivating a cadence increase from $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$8 to $t_E \sim 1.4\,\mathrm{hr}\,(M/0.1\,M_\oplus)^{1/2}$9 to recover full sensitivity (Gould et al., 2024).
• Direct mass measurements using simultaneous space-based microlens parallax yield individual masses/distances for super-Earth and even sub-Earth FFPs. Forthcoming campaigns (Roman+Euclid, L2-spacecraft+ground) will precisely determine the mass function down to Mars mass and below (Penny et al., 2019, Gould et al., 2020).
• Cluster environments: Deep JWST spectroscopy in clusters like NGC 1333 finds planetary-mass object fractions of $30$010–15%, exceeding lognormal-IMF predictions, with a dearth of T-type (sub-4 $30$1) objects marking a floor to star-like fragmentation (Langeveld et al., 2024).
• Implications: Accurate census of FFPMOs and their formation modes is essential for constraining planet formation theory, dynamical evolution of planetary systems, and the initial mass function at the planetary-mass tail.

7. Summary Table: Key FFPMO Survey Results

Survey/Mission	Mass Sensitivity	Disk Fraction (1–5 Myr)	Multiplicity (≥7 au)	Comments
Roman Microlensing	$30$2–$30$3	—	—	$30$4250 FFPMOs/5 yr (Johnson et al., 2020)
JWST (e.g., NGC 1333)	$30$5–$30$6	$30$7–$30$8	%%%%66$\sim$67%%%%2%	$M\sim100\,M_\oplus$110–15% of members are FFPMOs (Langeveld et al., 2024)
Taurus/Upper Sco	$M\sim100\,M_\oplus$2–$M\sim100\,M_\oplus$3	—	$M\sim100\,M_\oplus$4	Binaries more common in Taurus (Bouy et al., 17 Jun 2025)
IC 348 (Spitzer)	$M\sim100\,M_\oplus$5–$M\sim100\,M_\oplus$6	$M\sim100\,M_\oplus$7	—	Disk duration comparable to brown dwarfs (Seo et al., 27 Jan 2025)
Population Synthesis	$M\sim100\,M_\oplus$8–$M\sim100\,M_\oplus$9	—	—	$<1$0–$<1$1 ejected FFPs per star (Guo et al., 5 Nov 2025, Coleman et al., 2024)

Detailed, statistically robust measurement of the FFPMO mass function—including sub-Earth regime—awaits next-generation microlensing and imaging surveys, with the combination of mass, age, and multiplicity distributions poised to decisively link Galactic planet/BD formation to stellar population synthesis and planetary system evolution.