Luminous Fast Blue Optical Transients (LFBOTs)

• Luminous Fast Blue Optical Transients (LFBOTs) are rare, rapidly evolving, blue transients characterized by fast rise/decline times, high optical luminosity, and multi-wavelength detections.
• They exhibit strong X-ray, radio, and millimeter emissions that indicate fast outflows interacting with dense circumstellar material, as evidenced by temperatures >30,000 K and high expansion velocities.
• The physical origins remain uncertain, with competing models including compact-object accretion, mergers, massive star core-collapse, and tidal disruption events.

Luminous Fast Blue Optical Transients (LFBOTs) are a rare subclass of fast blue optical transients defined observationally by rapid optical evolution, blue colors at peak brightness, high optical luminosity, a visible host galaxy, and an X-ray or radio detection. One recent six-object analysis adopted the practical criteria $t_{1/2}\leq 12$ d, $g-r\leq -0.5$ mag, and $M_g<-20$ (Sevilla et al., 26 Jan 2026). Across the class, the defining phenomenology includes rise and fade on timescales of days, hot blue continua with early temperatures of order $\sim 30{,}000$ K, luminous X-ray and radio emission, and evidence for dense circumstellar material (CSM), yet the physical origin remains uncertain and multiple progenitor channels remain under active consideration (Chrimes et al., 16 Dec 2025).

1. Observational definition and class properties

LFBOTs occupy the luminous end of the broader FBOT population. Reported observational hallmarks include rise times of days, decay within $\lesssim 15$–20 d, peak absolute magnitudes from about $-20$ to $-22$, early blue colors, low ejecta mass inferred from rapid evolution, and luminous X-ray, radio, and millimeter emission (Chrimes et al., 16 Dec 2025). In the optically selected six-object sample, most events decline by 3–4 mag in 10 d, are non-nuclear, and show blue, approximately blackbody continua near maximum light (Sevilla et al., 26 Jan 2026).

The observational diversity within the class is already substantial. AT2018cow is the prototypical event; AT2022tsd displayed minutes-timescale optical flares tens of days after the initial transient; AT2023fhn established that LFBOTs can occur at large projected offsets from their hosts; CSS161010 showed broad, entirely blueshifted hydrogen emission reaching velocities up to $\sim 0.1c$; and AT2024wpp is the most luminous LFBOT discovered to date (Ouyed, 25 Jun 2025, Chrimes et al., 2023, Gutiérrez et al., 2024, LeBaron et al., 31 Aug 2025). A plausible implication is that “LFBOT” is a phenomenological label defined by a compact region of timescale–luminosity–color space, rather than by a single already-established physical mechanism.

AT2024wpp illustrates the upper end of the currently observed parameter space. Its ultraviolet-to-near-infrared campaign found $L_{\rm pk}\approx (2$–$4)\times 10^{45}$ erg s$g-r\leq -0.5$0, a rise to peak within $g-r\leq -0.5$1 d, and $g-r\leq -0.5$2 erg radiated in the first $g-r\leq -0.5$3 d, with a persistently blue thermal continuum and blackbody temperature $g-r\leq -0.5$4 K at peak that remained $g-r\leq -0.5$5 K thereafter (LeBaron et al., 31 Aug 2025). These values place LFBOTs well outside ordinary radioactive supernova parameter space.

2. Broadband phenomenology

Optical and ultraviolet emission are characterized by high temperatures, rapid evolution, and prolonged spectral simplicity. AT2024wpp remained dominated by a blue thermal continuum for the first $g-r\leq -0.5$6–35 d, with $g-r\leq -0.5$7–$g-r\leq -0.5$8 cm at peak and inferred expansion velocities from $g-r\leq -0.5$9 of $M_g<-20$0–$M_g<-20$1 before declining at later epochs (LeBaron et al., 31 Aug 2025). In a six-object LFBOT sample, the optical selection was complemented by radio or X-ray detection precisely because broadband behavior is part of the empirical class definition (Sevilla et al., 26 Jan 2026).

X-ray behavior is luminous, variable, and heterogeneous. In the six-object sample, X-ray luminosities span $M_g<-20$2–$M_g<-20$3 erg s$M_g<-20$4 at $M_g<-20$5 d (Sevilla et al., 26 Jan 2026). AT2024wpp reached $M_g<-20$6 erg s$M_g<-20$7, remained roughly constant for the first 7 d, then decayed rapidly, and later re-brightened near $M_g<-20$8 d while developing a transient Compton hump peaking near $M_g<-20$9 keV (Nayana et al., 31 Aug 2025). The same study interpreted the X-ray properties as favoring an embedded high-energy source shining through asymmetric expanding ejecta.

Radio and millimeter data consistently indicate fast outflows in dense media. In the six-object sample, most 10 GHz light curves peak at $\sim 30{,}000$0–100 d with peak radio luminosities of $\sim 30{,}000$1–$\sim 30{,}000$2 erg s$\sim 30{,}000$3, and synchrotron modeling gives $\sim 30{,}000$4–$\sim 30{,}000$5 shocks in dense media with $\sim 30{,}000$6–$\sim 30{,}000$7 cm$\sim 30{,}000$8 (Sevilla et al., 26 Jan 2026). For AT2024wpp, the radio emission peaked at $\sim 30{,}000$9 erg s$\lesssim 15$0 Hz$\lesssim 15$1 at $\lesssim 15$2 d, while equipartition modeling favored $\lesssim 15$3 for $\lesssim 15$4 km s$\lesssim 15$5, a dense shell at $\lesssim 15$6 cm, and a CSM profile $\lesssim 15$7 (Nayana et al., 31 Aug 2025). That work further noted that all radio-bright LFBOTs in its comparison set exhibit similarly steep $\lesssim 15$8 profiles.

Some LFBOTs show additional short-timescale or long-wavelength components. The quark-nova interpretation of AT2022tsd attributes tens-of-minutes optical flares to optically thin fragments of ejecta releasing trapped radiation, with flare luminosities comparable to the LFBOT peak value (Ouyed, 25 Jun 2025). Independent TESS and ZTF monitoring of 12 of the 14 known LFBOTs found no confirmed late-time optical flares outside AT2022tsd, ruling out comparable flares in AT2024qfm on 40–65 d timescales and ruling out similarly luminous flares in AT2022tsd itself at 380–430 d (Jayaraman et al., 15 Jun 2026). Near-infrared excesses have also been reported: a dust-echo model predicts $\lesssim 15$9–$-20$0 erg s$-20$1 lasting weeks, compatible with the previously unexplained NIR excess in AT2018cow (Metzger et al., 2022).

3. Spectral appearance and the suppression of line features

The early optical and ultraviolet spectra of LFBOTs are commonly blue and nearly featureless. AT2024wpp maintained a persistently blue, featureless thermal continuum for several weeks, and faint H and He features with two velocity components centered at $-20$2 km s$-20$3 and $-20$4 km s$-20$5 emerged only after $-20$6 d (LeBaron et al., 31 Aug 2025). More generally, broad H and He lines often emerge after $-20$7 weeks in the class-level summary of observational properties (Chrimes et al., 16 Dec 2025).

A dedicated radiative-transfer study mapped the appearance of H, He I, and He II lines in luminous transients to source luminosity, characteristic radius, and expansion velocity. It found that featureless spectra occur for high luminosities $-20$8 erg s$-20$9, compact ejecta radii $-22$0 cm, and sufficiently large velocities $-22$1, with the photospheric temperature scaling as

$-22$2

Intermediate regimes produce He II-dominated spectra, whereas larger radii or lower luminosities allow conspicuous H and He I emission (Aspegren et al., 2 Jan 2026). The same work argued that persistent optical and UV featurelessness likely requires non-homologous, compact outflows rather than a single homologously expanding supernova-like ejecta.

AT2024wpp supplied an empirical counterpart to this framework. Its ultraviolet-to-near-infrared study interpreted the long-lived featureless continuum as a consequence of continuous energy injection from a central source of high-energy emission that keeps the ejecta highly ionized, suppressing line formation until the continuum fades and the apparent photosphere recedes (LeBaron et al., 31 Aug 2025). This suggests that line suppression is not merely an atmospheric detail but a global constraint on outflow geometry, ionization state, and engine longevity.

CSS161010 demonstrated that not all late-time spectral evolution within the LFBOT luminosity–timescale domain is alike. After an initially blue, featureless phase, it developed very broad, entirely blueshifted H$-22$3 and H$-22$4 emission with line maxima offset by $-22$5 to $-22$6 km s$-22$7 and bluest edges up to $-22$8, without emission at the rest wavelength (Gutiérrez et al., 2024). That study favored a partial tidal disruption event by an intermediate-mass black hole (IMBH) as the origin of CSS161010, underscoring that spectroscopic diversity is central to the debate over whether LFBOTs are a single physical class.

4. Host galaxies, local environments, and spatial distribution

A uniform host-galaxy analysis of 11 LFBOTs found that all hosts are actively star-forming and frequently show recent bursts of star formation. The median host properties were $-22$9, present-day SFR $\sim 0.1c$0, and gas-phase oxygen abundance $\sim 0.1c$1 (Nugent et al., 24 Mar 2026). Relative to comparison samples, LFBOT hosts are more star-forming than core-collapse supernova hosts but less star-forming than SLSN-I hosts; they are more metal-poor than SN Ibc and II hosts but more metal-rich than SLSN-I and LGRB hosts (Nugent et al., 24 Mar 2026).

The spatial distribution of events within hosts is a major empirical constraint. The same study found median physical offsets of $\sim 0.1c$2 kpc and median host-normalized offsets of $\sim 0.1c$3, and reported that more than 30% of LFBOTs occur in their hosts’ faintest pixel or outside their host galaxy’s light (Nugent et al., 24 Mar 2026). This differs from the behavior of ordinary CCSNe and LGRBs, which more strongly track bright star-forming regions.

AT2023fhn made this issue especially prominent. Hubble Space Telescope imaging placed it at host-normalized offsets of $\sim 0.1c$4 and $\sim 0.1c$5 from the two nearest galaxies at the same redshift, with fractional host light $\sim 0.1c$6 (Chrimes et al., 2023). Its transient site showed very faint local surface brightness and no clear underlying H II region or resolved compact cluster (Chrimes et al., 2023). A later multiwavelength analysis of the same object argued that neither its host galaxy nor its circumstellar environment are unusual compared with previous LFBOTs, and that its diffuse yet young local environment together with a similar CSM favors a massive-star progenitor with strong winds (Chrimes et al., 2024). Taken together, these results indicate that LFBOT environments are star-forming on host-galaxy scales but do not necessarily coincide with the brightest local star-forming regions.

The host study therefore favored a compact-object and Wolf-Rayet star merger progenitor scenario, rather than nuclear IMBH tidal disruptions or prompt stellar-mass compact-object tidal disruptions, because LFBOTs are generally non-nuclear, not limited to dwarf galaxies, and do not trace active star-forming regions within their hosts (Nugent et al., 24 Mar 2026). This conclusion is model-dependent, but it is directly anchored in the current host demographics.

5. Progenitor models and the central-engine debate

Several physically distinct models have been proposed, and no single framework is presently established for the full class. One line of interpretation is compact-object accretion. For AT2024wpp, the X-ray and radio analyses favored a binary progenitor involving super-Eddington accretion onto a compact object launching mildly relativistic disk-wind outflows, while the ultraviolet-to-near-infrared study argued that the outflow geometry, persistent ionization, and multi-component velocities are naturally realized by super-Eddington accretion disks around neutron stars or black holes (Nayana et al., 31 Aug 2025, LeBaron et al., 31 Aug 2025).

A broader merger family has been developed in several theoretical papers. Wolf-Rayet/black-hole or Wolf-Rayet/neutron-star mergers were proposed to explain fast rise times, peak luminosities $\sim 0.1c$7 erg s$\sim 0.1c$8, low $\sim 0.1c$9Ni yields, aspherical ejecta spanning velocities from $L_{\rm pk}\approx (2$0 km s$L_{\rm pk}\approx (2$1 to $L_{\rm pk}\approx (2$2–$L_{\rm pk}\approx (2$3, dense H-depleted-but-not-free CSM on scales from $L_{\rm pk}\approx (2$4 cm to $L_{\rm pk}\approx (2$5 cm, and an embedded compact object producing variable non-thermal X-ray or gamma-ray emission (Metzger, 2022). A related delayed-dynamical-instability model attributes LFBOTs to helium-core–black-hole mergers after prolonged stable mass transfer, predicting optical peaks of $L_{\rm pk}\approx (2$6–$L_{\rm pk}\approx (2$7 erg s$L_{\rm pk}\approx (2$8, compact nearby CSM from the inspiral, and extended CSM at $L_{\rm pk}\approx (2$9 cm from earlier binary mass loss (Klencki et al., 10 Oct 2025). A 2026 study further argued that at least a subset of ultra-long GRBs and LFBOTs may share a common origin in compact-object–massive-star mergers, with SN 2011kl broadly consistent with an LFBOT origin but exhibiting a longer plateau and stronger UV suppression than the current LFBOT population (Villar et al., 8 Jul 2026).

Massive-star core-collapse channels remain active alternatives. One study explored very massive star failed supernovae forming black holes with masses $4)\times 10^{45}$0–40 $4)\times 10^{45}$1, finding a formation rate similar to the observed LFBOT rate, a bias toward low metallicity $4)\times 10^{45}$2, hydrogen- and helium-poor progenitors, dense circumstellar media, and long-lived accretion-disk emission capable of matching late-time luminosities (Chrimes et al., 3 Oct 2025). This scenario overlaps with the “super-kilonova” progenitor space and was proposed as a possible channel for prompt $4)\times 10^{45}$3-process enrichment (Chrimes et al., 3 Oct 2025).

More exotic models have also been advanced. The quark-nova model proposes delayed conversion of a neutron star into a highly magnetized hybrid star, ejecting $4)\times 10^{45}$4 at $4)\times 10^{45}$5 with $4)\times 10^{45}$6 erg; the resulting diffusion and transparency timescales of $4)\times 10^{45}$7 d and $4)\times 10^{45}$8 d were argued to match the fast optical evolution of LFBOTs, while ejecta fragmentation explains the minute-scale optical flares of AT2022tsd (Ouyed, 25 Jun 2025). The same work connected the neutron-rich ejecta to heavy-element enrichment and proposed delayed neutron-star conversion as an explanation for hostless or off-galaxy events such as AT2023fhn (Ouyed, 25 Jun 2025).

Tidal disruption models remain important but controversial. CSS161010 was interpreted as a partial TDE of a hydrogen-rich star by an IMBH, based on its spectroscopic uniqueness and dwarf-host context (Gutiérrez et al., 2024). For AT2024wpp, a systematic multiwavelength modeling study found that none of the tested optical, radio, and X-ray models could reasonably explain all aspects of the data, although the authors stated that other physical arguments favor a stellar-mass/IMBH TDE of a low-mass star together with a synchrotron blast wave (Omand et al., 6 Jan 2026). However, the environmental study of 11 hosts explicitly disfavored IMBH TDEs as a class-wide explanation because LFBOTs are not nuclear and are not limited to dwarf galaxies (Nugent et al., 24 Mar 2026). This is the central controversy in current LFBOT research: some individual objects admit TDE-like interpretations, while host and demographic data argue against a universal TDE origin.

6. Population context, related transients, and observational prospects

The volumetric rate of LFBOTs is itself part of the model discrimination problem. A 2040s white paper summarized the observational estimate as about $4)\times 10^{45}$9 of the core-collapse supernova rate, corresponding to only $g-r\leq -0.5$00–2 events per year presently discovered (Chrimes et al., 16 Dec 2025). The very massive-star failed-supernova study likewise compared LFBOTs to an observed volumetric rate of $g-r\leq -0.5$01 of the CCSN rate and found that black holes above $g-r\leq -0.5$02 match this rate robustly (Chrimes et al., 3 Oct 2025). By contrast, the quark-nova model predicted an LFBOT rate of $g-r\leq -0.5$03 of the local core-collapse supernova rate (Ouyed, 25 Jun 2025). These numbers indicate that rate estimates remain model-sensitive and that the true completeness of present surveys is uncertain.

LFBOTs also appear increasingly connected to neighboring transient classes. EP240414a was discovered as a fast X-ray transient with no reported gamma-ray counterpart; its early blue featureless spectrum, multi-peaked optical light curve, and later broad-lined Type Ic supernova led its authors to argue for a massive star creating a jet-forming supernova inside a dense envelope, thereby linking fast X-ray transients, GRBs, and some LFBOT-like optical phenomenology (Dalen et al., 2024). GRB 210704A, at $g-r\leq -0.5$04, displayed an extremely luminous and rapidly evolving optical/IR excess peaking at $g-r\leq -0.5$05 mag around rest-frame 2 d, explicitly described as resembling the emission seen in LFBOTs and Einstein Probe fast X-ray transients; in that case the favored interpretation was an energetic refreshed shock (Pieterse et al., 15 Apr 2026). These events suggest that at least some LFBOT-like optical components can coexist with successful relativistic jets.

Multi-messenger considerations extend the scope further. A neutrino study argued that LFBOTs are likely powered by compact objects launching asymmetric fast outflows, examined neutrino production in choked jets and CSM interaction, and found that the IceCube upper limit on AT2018cow excludes part of the choked-jet parameter space otherwise allowed by electromagnetic data (Guarini et al., 2022). The same work concluded that LFBOTs do not constitute the bulk of the diffuse astrophysical neutrino background, but that nearby events remain promising targets for IceCube and IceCube-Gen2 (Guarini et al., 2022).

Late-time engine activity appears to be uncommon. TESS and ZTF constraints found no confirmed luminous late-time flares in 12 monitored LFBOTs apart from the earlier AT2022tsd episode, with minutes-duration flare duty cycles constrained to $g-r\leq -0.5$06–0.01 during TESS observations (Jayaraman et al., 15 Jun 2026). This suggests that minutes-scale luminous flaring is either intrinsically rare, short-lived, or strongly anisotropic.

Current survey limitations remain severe. The 2040s white paper argued that most LFBOTs are missed by present photometric surveys or are not efficiently selected for detailed follow-up, and recommended deeper, wider, higher-cadence surveys, very short exposures, automated photometric and spectroscopic classification, spectroscopic capability reaching S/N $g-r\leq -0.5$07 at $g-r\leq -0.5$08 mag, routine multi-epoch polarimetry, and rapid concurrent radio, X-ray, and NIR coverage (Chrimes et al., 16 Dec 2025). This suggests that a decisive resolution of the LFBOT problem will require not only more events, but earlier discovery and denser multiwavelength sampling than has usually been available so far.