Exoplanets
The Brown Dwarf Desert
The missing middle — why almost no 13–80 Jupiter-mass companions orbit close to Sun-like stars
The brown dwarf desert is the observed scarcity of brown-dwarf companions — bodies of 13 to 80 Jupiter masses, spanning the deuterium- to hydrogen-burning range — on close orbits, inside roughly 3–5 AU, around Sun-like stars. Fewer than about 1 percent of solar-type stars host such a companion so near, a deficit of order ten relative to both the giant planets below it and the stellar companions above it, even though a close brown dwarf would produce a reflex Doppler wobble of order 1 km/s and be trivially easy to detect. First mapped by precision radial-velocity surveys in the late 1990s and named by Marcy, Butler and colleagues around 2000, the desert is a fossil clue: it marks the boundary between how planets are built by core accretion in disks and how stellar companions are born by cloud fragmentation — and how migration erases whatever tries to form in between.
- Mass range of the desert~13–80 M_Jup (0.012–0.08 M_sun)
- Deuterium-burning limit~13 M_Jup (core ~5×10⁵ K)
- Hydrogen-burning limit~0.075–0.08 M_sun (~75–80 M_Jup)
- Close-orbit occurrence<1% of Sun-like stars (a < 3–5 AU)
- Depth of the deficit~10× below the flanking populations
- Reflex wobble (40 M_Jup at 1 AU)K ≈ 1 km/s (easily detectable)
- Named / mapped byRV surveys, ~1998–2003 (Marcy & Butler et al.)
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why the brown dwarf desert matters
Line up every companion ever found orbiting a Sun-like star and sort it by mass and orbit. At the low-mass end sit the planets: Earths, super-Earths, Neptunes and gas giants up to about thirteen Jupiter masses, made abundantly by disks. At the high-mass end sit the stellar companions: red dwarfs and Sun-like secondaries, made abundantly by the fragmentation of collapsing gas clouds. Between the two, in the substellar band from roughly 13 to 80 Jupiter masses and within a few AU of the star, the census very nearly falls to zero. That empty band is the brown dwarf desert, and it matters because emptiness is data. The two mechanisms that fill the rest of the diagram both fail in the same place, and the shape of that failure encodes how giant companions form and where they end up.
- It separates two formation channels. Planets grow bottom-up by core accretion; stars form top-down by gravitational fragmentation. The desert is where both run out of steam.
- It is a migration record. A close brown dwarf couples strongly to a gas disk and to tides, so many that formed close were dragged into their star. The gap is partly a graveyard.
- It calibrates the planet/star boundary. The deuterium-burning line at 13 M_Jup and the hydrogen-burning line near 0.08 M_sun bracket the desert, tying it to fundamental nuclear physics.
- It sharpens survey statistics. Because brown dwarfs are so easy to detect by radial velocity, their absence is a clean, near-complete measurement rather than a selection artifact.
- It shapes companion-mass functions. Any theory of binary and planetary formation must reproduce the dip, making the desert a demanding benchmark.
How the desert is measured, step by step
The desert is defined by a companion's mass and orbital separation, both of which fall out of the star's motion:
- Detect the reflex wobble. A companion and its star orbit a common center of mass, so the star's line-of-sight velocity oscillates. A spectrograph measures this Doppler shift to a meter per second.
- Fit the orbit. From the velocity curve you recover the period P, eccentricity e, and semi-amplitude K. Kepler's third law then gives the semi-major axis a from P and the stellar mass M.
- Solve for minimum mass. The amplitude fixes the mass function, yielding m·sin i — the companion mass times the sine of the unknown inclination. For a brown dwarf near a Sun-like star, K is hundreds of meters to a kilometer per second, so the detection is unambiguous.
- Break the sin i ambiguity. Transits (if the orbit is edge-on), astrometry from Gaia, or Hipparcos–Gaia proper-motion anomalies pin down the inclination, converting m·sin i into a true mass and confirming the object lies in the 13–80 M_Jup band.
- Bin the population. Repeating this over thousands of stars builds the companion mass function versus separation. Inside a few AU, the 13–80 M_Jup bins come up nearly empty — the desert.
The reflex-velocity equation
The whole measurement rests on the radial-velocity semi-amplitude of the host star:
K = (2πG / P)1/3 · (m sin i) / (M + m)2/3 · 1 / √(1 − e²)
- K — reflex velocity semi-amplitude of the star, in m/s.
- G — gravitational constant, 6.674×10⁻¹¹ m³ kg⁻¹ s⁻².
- P — orbital period, in seconds.
- m — companion mass; M — stellar mass, in kg.
- i — orbital inclination (90° = edge-on, transiting); only m·sin i is observed from RV alone.
- e — orbital eccentricity (0 = circular).
For a 40 M_Jup brown dwarf at 1 AU (P ≈ 1 yr) around a 1 M_sun star on a circular edge-on orbit, K ≈ 1.1 km/s. Compare Jupiter's pull on the Sun, K ≈ 12.5 m/s, or Earth's, K ≈ 0.09 m/s — the brown dwarf signal is nearly a hundred times larger than Jupiter's, and still an order of magnitude above a typical hot Jupiter's ~100 m/s wobble. It is impossible to miss, which is precisely why the emptiness of the desert cannot be blamed on instruments being too blind to see them.
A worked example: how empty is empty?
Suppose a radial-velocity program monitors 1,000 Sun-like stars with a precision of 3 m/s over a decade, easily good enough to catch any companion above a few Neptune masses out to several AU. Tally the companions inside 3 AU by mass band and the desert appears as a number, not a hand-wave:
| Companion class | Mass range | Formation channel | Frequency (close orbits) |
|---|---|---|---|
| Giant planets | 0.3–13 M_Jup | Core / pebble accretion in disk | ~10–20% |
| Brown dwarfs (the desert) | 13–80 M_Jup | Neither channel efficient | <1% |
| Stellar companions | >80 M_Jup (>0.08 M_sun) | Cloud / core fragmentation | ~10–20% |
The middle row is the desert: a factor of ten or more below its neighbors, in the mass band where detection is easiest. The dip is deepest for the tightest orbits and around 30–40 M_Jup, sometimes described as the driest point of the desert. Widen the aperture, however, and the picture changes. Direct-imaging surveys that probe tens to thousands of AU recover brown dwarfs at frequencies of a few percent, comparable to the low-mass tail of stellar binaries. The desert is therefore a close-in feature of the companion-mass function, not a universal ban on making brown dwarfs.
History and a note on names
Brown dwarfs were hypothesized in the 1960s (Kumar, Hayashi and Nakano) as objects too light to sustain hydrogen fusion, and the first unambiguous one, Gliese 229B, was imaged in 1995 — the same year 51 Pegasi b became the first hot Jupiter around a Sun-like star. As Doppler surveys on the Lick, Keck, ELODIE and CORALIE spectrographs accumulated companions through the late 1990s, an odd pattern hardened: planets piled up below 13 M_Jup, binaries piled up above 80 M_Jup, and the substellar middle at short period stayed nearly empty. Geoffrey Marcy, Paul Butler and collaborators, along with the Geneva team, named this the brown dwarf desert around 1998–2000. Later transit surveys (WASP, HAT, TESS) and the Gaia astrometric mission confirmed and refined it, adding true masses where inclinations were unknown and pushing the census toward completeness.
Common misconceptions
- Brown dwarfs are rare in general. No — they are common as free-floaters and as wide companions. Only the close-orbit population around Sun-like stars is depleted.
- We just can't detect them. The opposite: a close brown dwarf is one of the loudest radial-velocity signals possible (~1 km/s). Their absence is real, not a blind spot.
- The desert is empty because 13–80 M_Jup objects don't exist. They exist — they simply avoid tight orbits around solar-type stars. Fragmentation makes them far out; disks rarely make them at all.
- It is a hard, sharp gap. It is a statistical dip that varies with separation and stellar mass; it is driest for the closest orbits and around the mid-substellar masses.
- Brown dwarfs are "failed stars" and nothing more. They straddle the boundary: below 13 M_Jup they resemble giant planets, above 80 M_Jup they become red dwarfs — the desert lies exactly in between.
- The Sun could hide a close brown dwarf. It cannot — such a companion would have been detected decades ago from the Sun's motion; the desert is why we would be very surprised to find one.
Frequently asked questions
What is the brown dwarf desert?
The brown dwarf desert is the striking rarity of brown-dwarf companions — objects between about 13 and 80 Jupiter masses — orbiting Sun-like stars at close separations, typically inside 3–5 AU or with orbital periods under a few years. Fewer than roughly 1 percent of solar-type stars carry such a companion so close, a deficit of about a factor of ten compared with both the giant planets below 13 Jupiter masses and the stellar companions above 80 Jupiter masses. It was first mapped by radial-velocity planet searches in the late 1990s and 2000s.
Why is 13 to 80 Jupiter masses the brown dwarf range?
The lower bound near 13 Jupiter masses is the deuterium-burning limit: above it, the core reaches temperatures around 5×10⁵ K where deuterium fuses (p + D → ³He), the classic dividing line the IAU uses between planets and brown dwarfs. The upper bound near 75–80 Jupiter masses (about 0.075–0.08 solar masses) is the hydrogen-burning limit, where the core sustains ordinary hydrogen fusion and the object becomes a true star, a red dwarf. Between the two, an object fuses deuterium briefly but never ignites hydrogen — a substellar brown dwarf.
Why does the brown dwarf desert exist?
The favored explanation is that the two populations bordering the desert form differently and are then reshaped by migration. Giant planets grow bottom-up in a protoplanetary disk by core accretion, while stellar companions form top-down by the fragmentation of a collapsing molecular cloud. Neither channel efficiently makes 13–80 Jupiter-mass bodies at small separations. Disk-driven inspiral compounds it: a brown dwarf embedded in a disk exchanges angular momentum strongly and tends to spiral all the way into the star, clearing the close-in region.
How was the brown dwarf desert discovered?
It emerged from precision radial-velocity surveys of nearby Sun-like stars. The reflex wobble of a star scales with companion mass, so a brown dwarf at a few AU induces a velocity signal of hundreds of meters per second — far larger than the meters-per-second signal of a planet and trivially detectable. Yet surveys such as those on the Lick, Keck, ELODIE and CORALIE spectrographs kept finding abundant planets and abundant stellar binaries but almost nothing in the 13–80 Jupiter-mass gap at short period. Marcy and Butler and collaborators named the deficit the brown dwarf desert around 2000.
Is the brown dwarf desert dry at all separations?
No — the desert is a close-in phenomenon. Direct-imaging surveys find that brown dwarfs become considerably more common as wide companions, tens to thousands of AU from their host stars, where they plausibly form like the low-mass tail of binary-star fragmentation. The dryness is concentrated at separations inside a few AU, which is exactly the regime radial-velocity and transit surveys probe best, so the desert is really a gap in the close-orbit companion mass function rather than a total absence of brown dwarfs.
What is a radial-velocity reflex signal and how big is a brown dwarf's?
As a companion orbits, the host star traces a small counter-orbit, and its line-of-sight velocity oscillates — the reflex or Doppler wobble. The semi-amplitude is K = (2πG/P)^(1/3) · m sin i / (M+m)^(2/3) · 1/√(1−e²). For a 40-Jupiter-mass brown dwarf at 1 AU around a Sun-like star, K is roughly 1 km/s, versus about 12 m/s for Jupiter around the Sun and 0.09 m/s for the Earth. Brown dwarfs are therefore among the easiest companions to detect, which makes their near-absence at close orbits all the more telling.
How does the brown dwarf desert relate to hot Jupiters?
Both are shaped by migration, but the outcomes differ. Hot Jupiters — gas giants a few tenths of an AU from their stars — survived their inward journey and parked near the star. Close brown dwarfs are far rarer, because their larger mass makes disk migration and tidal decay more violent, so many spiral into the star or fill their Roche lobes and are consumed. The desert is thus partly a graveyard: the missing close brown dwarfs may have been swallowed, leaving planets on one side and stellar binaries on the other.