FU Orionis Outburst

A star that flared up and never came back down

In the autumn of 1936, an obscure variable star in the constellation Orion brightened by almost six magnitudes — roughly a hundredfold — over the course of about a year. That alone would have been unremarkable; the sky is full of variable stars that surge and fade. What made FU Orionis extraordinary is that it never faded back. Ninety years later it is still sitting within about a magnitude of its 1936 maximum, lit up like a porch light that someone switched on in the 1930s and forgot to turn off. The eruption was not a flash. It was a permanent — or at least decades-long — change of state.

We now understand FU Orionis and its relatives as the most dramatic visible manifestation of episodic accretion: the idea that young stars do not gather their mass at a steady trickle but in violent, intermittent gulps. In an FU Orionis outburst (the eruptions are called FUor events, and the class is named for the archetype), the rate at which gas from the surrounding disk falls onto the star jumps from a quiescent value near 10⁻⁷ solar masses per year to a burst value of order 10⁻⁴ M☉/yr — a factor of about a thousand. The inner disk is heated to thousands of kelvin, it blazes, and the whole system brightens by about 5 magnitudes — a factor of 100 — in the optical. The brightening lasts decades to a century. During the burst the star can swallow the equivalent of a Jupiter mass or more of disk gas.

How the outburst works

A young stellar object is surrounded by a rotating accretion disk left over from its collapsing natal cloud. Gas in the disk must shed angular momentum to spiral inward and reach the star, and it does so through viscous and magnetic torques. In the quiescent state these torques are weak in the inner disk, so material delivered from the outer disk cannot get through fast enough. Gas backs up. A reservoir builds in an annulus a few tenths of an AU from the star — the disk's "dead zone," where the gas is too cold and too poorly ionized to couple to magnetic fields and transport itself inward.

The reservoir grows hotter and denser until it crosses an instability threshold. At that point the disk flips, locally, from a low-viscosity, low-accretion state to a high-viscosity, high-accretion state. Once the switch flips, the dammed-up gas pours inward all at once. The accretion rate through the inner disk leaps by three orders of magnitude, the inner disk heats to several thousand kelvin, and the system lights up. The outburst persists until the reservoir is drained and the inner disk cools back below the threshold — which, because the reservoir holds a Jupiter-to-many-Jupiter mass of gas, takes decades.

Crucially, almost all the light during an FUor outburst comes from the disk, not the stellar photosphere. The hot inner disk outshines the young star by a large factor. This is why FUor spectra are weird: they look like an F or G supergiant in the optical, but like a cooler K or M giant in the near-infrared, because at each wavelength we are looking at a different, progressively cooler annulus of the disk. The "spectral type" depends on which color you observe — a fingerprint no single stellar photosphere can produce.

Worked example: the energetics of a single burst

Let us put numbers to a generic FUor event around a young star of mass M ≈ 0.5 M☉ and radius R ≈ 2 R☉ (young stars are puffy). The accretion luminosity released when gas falls from the disk onto the stellar surface is

L_acc = G M Ṁ / R

Plug in the burst accretion rate Ṁ = 10⁻⁴ M☉/yr = 6.3 × 10¹⁸ kg/s, M = 0.5 M☉ = 1.0 × 10³⁰ kg, and R = 2 R☉ = 1.4 × 10⁹ m:

L_acc = (6.67×10⁻¹¹ × 1.0×10³⁰ × 6.3×10¹⁸) / (1.4×10⁹)
      ≈ 3.0×10²⁷ W
      ≈ 770 L☉

So a half-solar-mass protostar accreting at the burst rate radiates several hundred solar luminosities — overwhelmingly from the disk. In quiescence at 10⁻⁷ M☉/yr the same formula gives only about 0.8 L☉. The luminosity ratio between burst and quiescence is the ratio of accretion rates, ~1000×, but because the protostar already has an intrinsic photospheric luminosity of order 1 L☉ that does not switch off, the observed optical brightening is the more modest ~100× (5 magnitudes) quoted for the class.

How much mass is delivered? If the burst sustains 10⁻⁴ M☉/yr for, say, 50 years before tapering, the star gains

ΔM ≈ 10⁻⁴ M☉/yr × 50 yr = 5×10⁻³ M☉ ≈ 5 Jupiter masses

A single FUor outburst therefore moves several Jupiter masses of gas from disk to star. Repeat such a burst a few dozen times over the embedded lifetime and you have assembled a substantial fraction of a solar mass — the heart of the episodic-accretion solution to the luminosity problem.

The problem FUors solve

For decades, observers of embedded protostars faced a quiet contradiction known as the protostellar luminosity problem. If a star accretes steadily, its accretion luminosity should equal G M Ṁ / R for whatever Ṁ is required to build the star within the roughly 0.5-million-year embedded phase. That demands a mean Ṁ of order a few × 10⁻⁶ M☉/yr and a corresponding luminosity of tens of solar luminosities. But surveys — most decisively the Spitzer "Cores to Disks" census — found that embedded protostars are systematically fainter than that, with a median luminosity of only a few L☉. They are not accreting fast enough, on average, to be building themselves on schedule.

Episodic accretion resolves the paradox. A protostar spends the overwhelming majority of its time in faint quiescence at ~10⁻⁷ M☉/yr — which is why surveys catch most of them looking dim — and assembles the bulk of its mass during a handful of brief, luminous FUor-like bursts at ~10⁻⁴ M☉/yr that any single snapshot survey is statistically unlikely to catch. The time-averaged accretion rate works out right; the instantaneous rate is almost always far below or, rarely, far above the average. FU Orionis outbursts are the visible, nearby proof that those bursts happen.

Variants and regimes: FUors, EXors, and the continuum between

FU Orionis outbursts sit at the extreme end of a continuum of accretion-driven young-star eruptions. The two named classes anchor the ends:

• FUors (named for FU Orionis): 4–6 magnitude brightenings, durations of decades to a century, disk-dominated absorption-line spectra with the wavelength-dependent spectral type, strong winds. Rare; each star likely produces only a few dozen over its embedded life.
• EXors (named for EX Lupi): 2–4 magnitude brightenings, durations of months to a couple of years, repeating every few years, T Tauri-like emission-line spectra in which the underlying star is still visible. Milder, more frequent accretion events.
• Hybrids and intermediates: objects such as V1647 Orionis and HBC 722 show amplitudes, durations, or spectra between the two classes, suggesting FUors and EXors are not distinct phenomena but different amplitudes of the same disk-instability process.

There is also a regime question about the trigger. Three mechanisms compete, and different FUors may be driven by different ones: a pure thermal (hydrogen-ionization) instability in the inner disk, analogous to the disk instability that powers dwarf-nova eruptions; a two-stage gravitational-instability-plus-MRI model in which spiral arms or clumps in a self-gravitating outer disk migrate inward and dump gas until they ignite the magnetorotational instability in the hot inner disk; and external triggering by a close stellar companion on an eccentric orbit or by an infalling clump of envelope material. Modern multi-band light curves and ALMA disk imaging are being used to discriminate among them.

Observational status: from FU Ori to HBC 722

The class was built from three classical members. FU Orionis (in Orion, distance ~400 pc) erupted in 1936–1937 and remains the defining case. V1057 Cygni erupted in 1969 — the first FUor caught reasonably early — and has since faded more rapidly than FU Ori, demonstrating that not all FUors share the same decline timescale. V1515 Cygni rose more gradually over years rather than in a single fast jump, showing that the rise itself can vary from months to years.

The modern era is defined by events caught from the pre-outburst state and tracked densely. V1647 Orionis illuminated McNeil's Nebula when it brightened in 2003, faded, and re-brightened in 2008. HBC 722 (also catalogued V2493 Cygni) erupted in 2010 in the Pelican Nebula with archival pre-outburst photometry already in hand, providing one of the cleanest before-and-after datasets for testing the accretion-instability picture. Today roughly two dozen confirmed or candidate FUors are known. That is far fewer than the eruption rate implied by the luminosity problem would predict at any instant — but the discrepancy is expected, because each object spends only a percent or so of its time in outburst, and most FUors are deeply embedded and visible only in the infrared.

FU Orionis outbursts versus other young-star and accretion variables

Phenomenon	Amplitude	Duration	Burst Ṁ (M☉/yr)	Spectrum / driver
FUor (FU Orionis)	~5 mag (×100)	decades–century	~10⁻⁴	disk absorption lines; major disk instability
EXor (EX Lupi)	2–4 mag	months–few yr, recurring	~10⁻⁶	T Tauri emission; milder accretion burst
Classical T Tauri (steady)	<1 mag flicker	continuous	~10⁻⁸	magnetospheric accretion onto star
Stellar flare (YSO)	~1–3 mag	minutes–hours	—	magnetic reconnection on photosphere
Dwarf nova	2–6 mag	days–weeks, recurring	—	same thermal disk instability, white-dwarf scale
Classical nova	~8–15 mag	weeks–months	—	thermonuclear runaway on a white dwarf surface
Type Ia supernova	~20 mag	weeks–months	—	thermonuclear disruption of a white dwarf

The instructive comparison is with dwarf novae: both are driven by essentially the same thermal disk instability, just at very different physical scales — an AU-scale protostellar disk versus a solar-radius white-dwarf disk — which is why FUor outbursts last decades while dwarf-nova outbursts last days. The comparison with novae and supernovae underscores that an FUor is not a thermonuclear or explosive event: no material is destroyed, nothing detonates. It is gravity-powered accretion, simply switched into a high gear.

Common pitfalls and misconceptions

• "FU Orionis exploded." It did not. There is no explosion, no shell ejection, no thermonuclear runaway. The brightening is accretion luminosity from gas falling onto the star, released gravitationally. The star and disk survive intact.
• "The star's surface got hotter and brighter." Mostly the disk brightened, not the photosphere. The inner disk outshines the star, which is why the optical spectrum looks like a supergiant while the infrared looks like a cool giant — a single photosphere cannot do that.
• "5 magnitudes means 5 times brighter." Magnitudes are logarithmic. A 5-magnitude rise is a factor of 100 in flux (each magnitude is ~2.512×, and 2.512⁵ = 100), which is exactly why the headline number is "about a hundredfold."
• "FUors fade like ordinary novae, in months." The defining feature is the opposite: the decline is extraordinarily slow, decades to a century, because the inner-disk reservoir that feeds the burst holds several Jupiter masses of gas and drains slowly.
• "We've seen plenty, so they're common." Only ~two dozen are known. They look common in the sense that most young stars probably undergo many over their lifetimes, but at any instant only a tiny fraction of protostars are in outburst — which is precisely the statistical fact that the luminosity problem exposed.
• "It's just a brighter T Tauri star." The accretion rate is ~1000× higher than a classical T Tauri star's, the light is disk-dominated rather than star-dominated, and the spectrum shows absorption lines and a strong wind rather than the emission lines of magnetospheric accretion. It is a qualitatively different state.

Quantitative analysis: why the instability switches and how long it lasts

The thermal-instability picture rests on the shape of the disk's local thermal-equilibrium curve — the relation between surface density Σ and effective temperature (equivalently, accretion rate) at a given radius. For a region near hydrogen's ionization temperature (~10⁴ K at the disk surface, fed by an interior near a few thousand K), the opacity rises steeply with temperature. This makes the equilibrium curve S-shaped: at fixed radius there are two stable branches — a cool, neutral, low-Ṁ branch and a hot, ionized, high-Ṁ branch — separated by an unstable middle.

In quiescence the dead-zone annulus sits on the lower branch, accumulating mass from the outer disk faster than it can pass it inward. Σ rises until the lower branch ends; the annulus has nowhere to go but to jump to the hot branch, where the ionized gas couples to magnetic fields, the magnetorotational instability switches on, viscosity soars, and Ṁ leaps to ~10⁻⁴ M☉/yr. The annulus then drains, Σ falls, the hot branch ends, and it drops back to the cool branch — completing a limit cycle. The schematic mass budget is:

quiescent fill:   Ṁ_outer ≈ 10⁻⁷ M☉/yr feeds the reservoir
reservoir built:  ΔM_res ≈ few × 10⁻³ M☉ (a few Jupiter masses)
burst drain:      Ṁ_burst ≈ 10⁻⁴ M☉/yr
burst duration:   t_burst ≈ ΔM_res / Ṁ_burst
               ≈ (5×10⁻³ M☉) / (10⁻⁴ M☉/yr)
               ≈ 50 years

The duration is therefore not arbitrary: it is set by the reservoir mass divided by the burst rate, and both numbers conspire to give the decades-to-century timescale observed. The recurrence time between bursts is the reservoir mass divided by the slow refill rate:

t_recur ≈ ΔM_res / Ṁ_outer
       ≈ (5×10⁻³ M☉) / (10⁻⁷ M☉/yr)
       ≈ 5×10⁴ years

So a typical protostar would undergo an FUor-like burst roughly every ~10⁴–10⁵ years and stay in outburst for ~10–100 years each time — meaning it is in the high state only about 0.1% of the time. Over a ~0.5-Myr embedded phase that is order ten bursts, each delivering several Jupiter masses, which is enough to assemble a substantial fraction of the final stellar mass through these episodes. The exact numbers depend on disk viscosity prescriptions and on whether gravitational instability in the outer disk, rather than pure thermal instability, sets the refill rate; this is an active research frontier, with ALMA disk masses and decades-long photometric monitoring providing the constraints.