Asteroid Rubble Pile

A pile, not a rock

The cartoon asteroid — a single dark boulder tumbling through space — is wrong for most of the population. The reality is closer to a slow-motion gravel bank: a loose pile of fragments, ranging from house-sized boulders down to dust, held together by their mutual gravity and nothing else. There are no chemical bonds between the pieces, no welded contacts, often no measurable tensile strength at all. The whole structure is suspended in equilibrium between centrifugal stress and the few microgee of self-gravity at its surface. Walk on one and you would float off; lean on a boulder and it might roll away.

This picture was first articulated in the 1970s and 1980s, but for two decades it remained a hypothesis. Asteroids were too small and too far to see in detail. The case had to be built indirectly, from the disagreement between the bulk density of an asteroid — its total mass divided by its total volume — and the density of the meteoritic material it sheds when it breaks up. By the 1990s that gap had become impossible to ignore, and by the 2000s spacecraft visits turned the hypothesis into direct observation.

The density argument: where did the mass go?

Take an ordinary chondrite — the most common class of stony meteorite, fallen from typical S-type and C-type asteroids — and measure its mineral density in a laboratory. You get 3.0 to 3.7 g/cm³ for ordinary chondrites, 2.6 to 2.9 for carbonaceous chondrites. Now measure the bulk density of an asteroid by independent means: spacecraft Doppler tracking, satellite orbits in binary systems, or stellar-occultation timing combined with radar shapes.

Asteroid	Diameter	Bulk density (g/cm³)	Inferred porosity	Method
Itokawa	~330 m	1.9 ± 0.1	~40 %	Hayabusa Doppler
Bennu	~490 m	1.19 ± 0.01	~55 %	OSIRIS-REx tracking
Ryugu	~870 m	1.19 ± 0.02	~50 %	Hayabusa2 tracking
Eros	~16 km	2.67 ± 0.03	~20 %	NEAR–Shoemaker
Mathilde	~53 km	1.3 ± 0.3	~50 %	NEAR flyby
Dimorphos (post-DART)	~160 m	~2.4	~25 %	DART, LICIACube

The bulk values are half — sometimes a third — of the mineral density. The missing mass cannot be in subatomic voids; it has to be in physical empty space inside the body. That space takes two forms. Microporosity is the pore space inside individual fragments, inherited from the chondrite parent material; it is typically 10–25 percent in unaltered chondrites. The rest is macroporosity: the voids between fragments — the gaps in the rubble pile itself. For Bennu's 55 percent total porosity, after subtracting a few tens of percent of micro-porosity, more than 30 percent of the body is open space between rocks. Bennu is mostly hole.

The 2.2-hour spin barrier

The second line of evidence is statistical. Photometric lightcurve surveys have measured rotation periods for thousands of asteroids. Plot rotation period against diameter and a striking pattern emerges: almost no body larger than about 150 metres rotates faster than once per 2.2 hours. Below 150 metres, the cliff disappears and rotations of seconds or fractions of a second are routine.

The reason is purely Newtonian. For a uniform sphere of density ρ rotating with angular speed ω, the centrifugal acceleration at the equator is ω² R; the surface gravity is (4/3) π G ρ R. Setting them equal:

ω² R = (4/3) π G ρ R
ω    = √((4/3) π G ρ)
T    = 2π / ω = √(3π / (G ρ))

For ρ = 2 g/cm³ (the empirical mean rubble-pile density), T ≈ 2.2 hours. A body that tries to spin faster than this and is held only by gravity will fly apart at the equator. The 2.2-hour cutoff is therefore a direct measurement of the spin at which a cohesionless aggregate of density 2 g/cm³ breaks up. Bodies smaller than ~150 m can spin faster because tensile strength — a property of cohesive solid rock — overtakes self-gravity at small sizes. Bodies above that scale cannot; the spin barrier is the smoking gun of the rubble-pile interior.

A small number of fast rotators do exist among 100–300 m bodies. They are interpreted as either monolithic remnants or as rubble piles with modest cohesion of order ~10 to 100 Pa — a few millibar, vastly weaker than rock — provided by van der Waals forces between fine dust grains. Bennu's measured surface cohesion from particle-ejection events is in this range.

Ground truth: the four sample-return targets

Three Japanese and one American mission have now visited rubble piles at close range, two of them returning samples to Earth.

Itokawa (Hayabusa, 2005). The Japanese Hayabusa spacecraft rendezvoused with the 330-metre near-Earth asteroid Itokawa in September 2005, surveyed it for three months, briefly touched down to collect dust, and returned the sample to Earth in 2010. The body turned out to be peanut-shaped — two distinct lobes joined at a thin neck — and covered in a startlingly bimodal surface: a rough "highland" terrain piled with boulders up to 50 m across, and smoother "seas" of pebbles where finer particles had migrated. There was no impact crater large enough to gouge the surface intact. The conclusion was that Itokawa is two rubble piles in contact, with grain sizes from metres down to micrometres distributed by gravitational sorting.

Bennu (OSIRIS-REx, 2018–2021). NASA's OSIRIS-REx spacecraft arrived at the 490-metre carbonaceous asteroid Bennu in late 2018 and spent two and a half years mapping, sampling, and departing with a 122-gram sample (returned to Earth in 2023). Bennu's bulk density of 1.19 g/cm³ implies the highest porosity ever measured for a small body. Its surface is wall-to-wall boulders, with no smooth ponds. Most surprising was an unexpected discovery: Bennu spontaneously ejected small particles from its surface throughout the mission, in events ranging from individual rocks to thousands of grains at once. The mechanism — thermal cracking, micrometeorite impacts, or both — confirmed that surface cohesion is in the tens of pascals at most. The November 2020 "TAG" sampling event saw the spacecraft's sample-collection head sink nearly half a metre into the surface and shoot back out covered in pebbles; the surface had less bearing strength than a children's ball pit.

Ryugu (Hayabusa2, 2018–2019). Hayabusa2 arrived at the 870-metre carbonaceous asteroid Ryugu in mid-2018, deployed three small landers (MINERVA-II1A/1B and the German-French MASCOT), executed two touchdown sampling operations, and fired a copper impactor into the surface to excavate fresh sub-surface material. The body is also covered in boulders, also low-density (1.19 g/cm³, similar to Bennu), and shows a pronounced "top shape" — equatorial bulge — with a strikingly low-relief surface and a chaotic distribution of boulder sizes. The artificial impact crater was nearly twice as large as a non-porous target would have given, in direct quantitative agreement with rubble-pile cratering models.

Dimorphos (DART, 2022). Dimorphos is the 160-metre moonlet of the 780-metre asteroid Didymos. The DART spacecraft, the first dedicated planetary-defence test, impacted Dimorphos on 26 September 2022 at 6.1 km/s. Telescopes on Earth and the Italian LICIACube companion measured the orbital change: Dimorphos's period around Didymos shortened by 32 minutes, far in excess of the 7-minute change expected from a clean inelastic momentum transfer. Hubble and ground-based images showed a long ejecta tail. The interpretation was unambiguous: 13 metric tonnes of Dimorphos's loose surface material were ejected backwards at high speed, recoiling against the moonlet with an additional impulse roughly 2.5 times that of DART itself — a momentum-enhancement factor (β) of about 3.6. Such large β values are characteristic of porous, cohesionless targets and are theoretically impossible for monolithic rock. Dimorphos, like its bigger siblings, is a rubble pile.

How rubble piles form

Two pathways dominate.

Catastrophic disruption + reassembly. Most rubble piles are second- or higher-generation bodies. Their parents — much larger asteroids tens to hundreds of kilometres across — were shattered in the early solar system by hypervelocity collisions. Numerical simulations (Michel, Benz, Richardson and others, 2001 onwards) show that immediately after disruption, fragments fly outward from the impact site at speeds from metres per second to many tens of metres per second. Most exceed the escape velocity of the parent and are lost; the slowest do not, and re-converge under their mutual gravity over hours to days. The reassembled body has a chaotic mix of fragment sizes, low cohesion, and substantial macroporosity — exactly the rubble-pile signature. The remaining debris becomes the asteroid families we still observe today: the Themis, Eos, and Karin families, for example, are coherent groups of bodies whose orbits trace back to single disruption events.

YORP spin-up. The Yarkovsky–O'Keefe–Radzievskii–Paddack (YORP) effect is the rotational analogue of the Yarkovsky effect: anisotropic re-emission of absorbed sunlight as thermal infrared exerts a tiny torque on an irregularly shaped body. Over millions of years, YORP can substantially change the spin rate of an asteroid smaller than a few kilometres. If a rubble pile is spun up past the 2.2-hour barrier, its weakly bound surface layer flows toward the equator, builds an equatorial bulge — the "top shape" — and may eventually shed material entirely. The shed material either escapes or remains gravitationally bound and forms a small secondary moonlet. About 15 percent of near-Earth asteroids larger than 200 m are binaries, almost all believed to be YORP-spawned, and their primaries are characteristically top-shaped. Both Ryugu and Bennu are textbook examples of YORP-shaped rubble piles.

What is actually inside?

"Rubble pile" is a structural class, not a precise internal architecture. Different sub-types are possible.

• Strict rubble pile. A gravitationally bound aggregate of fragments with broad size distribution and effectively zero cohesion. Bennu and Ryugu fit best.
• Shattered/fractured monolith. A body that was disrupted but not fully reassembled — internally fragmented and cracked, with macroporosity, but with some fragments still in contact or partially welded. May retain modest tensile strength. Eros (porosity ~20%) is in this category.
• Contact binary / bilobate. Two distinct rubble piles in tangential contact, often with a clear neck. Itokawa, the comet 67P/Churyumov–Gerasimenko, and the Kuiper Belt object Arrokoth are all contact-binary morphologies.
• Compacted aggregate. A rubble pile that has been gravitationally compressed at its centre by its own self-gravity. The macroporosity decreases with depth; the outer 100 m is the loosest. Hypothesised for the largest rubble piles (>10 km), no direct observation yet.

Why it matters

The rubble-pile picture has consequences far beyond academic interest.

• Planetary defence. Earth-crossing asteroids in the 100 m to 1 km size range are exactly the rubble-pile regime. DART has shown that kinetic-impactor deflection is highly effective on this population — far more so than the canonical β = 1 assumption that applies to monolithic targets. A 10-tonne impactor against a 100-m rubble pile delivers an effective momentum impulse of roughly β × M v, with β ≈ 3 plausible. The corollary is that nuclear deflection options can be smaller, and lead times shorter, than once thought.
• Sample return. Touching down on a rubble pile is easier than landing on a hard surface in some ways (low gravity, no need for legs) and harder in others (no firm anchor, surface that gives way and shoots material back). Both OSIRIS-REx and Hayabusa2 had to design "touch-and-go" sampling that allowed for the surface to recoil. Future missions to comet nuclei (effectively cold rubble piles with ice) face the same challenge.
• Asteroid mining. A loose pile of boulders is mechanically very different from a continuous ore body. Conventional rotary drilling fails; concepts using nets, collection bags, or in-situ centrifuges have been proposed instead.
• Cratering record interpretation. A rubble-pile target absorbs impactor energy by compaction rather than purely by shock heating. Crater diameters scale differently with impactor energy, and the abundance of small craters can be a poor proxy for incident flux. Bennu has fewer small craters than its surface age would predict — they have been erased not by re-cratering but by ongoing surface shaking and slow boulder migration.
• Solar-system formation. If most of today's asteroids are post-disruption rubble piles, the present mass function and orbital distribution of the asteroid belt are heavily reprocessed from the original planetesimal population. Reconstructing primordial conditions from current asteroid data is therefore an exercise in collisional cascade modelling.

Open questions

• How does cohesion scale with grain size? Surface particle ejection from Bennu suggests cohesion in the 10–100 Pa range. But this is a surface measurement at the smallest grain sizes; what binds the metre-scale boulder fields is essentially unconstrained.
• What is the depth profile of porosity? Self-gravity compresses the interior of a large rubble pile. The depth dependence of macroporosity has been modelled but never measured. Future seismic experiments on asteroids — proposed for missions in the 2030s — could resolve it.
• Do largest asteroids (Vesta, Ceres) hide a rubble-pile substructure? Vesta is differentiated and probably mostly intact, Ceres is partly hydrated and may have an icy mantle. But the next tier down — Pallas, Hygiea, Interamnia — could have substantially porous, fragmented interiors. Their bulk densities are intermediate (2.0–2.6 g/cm³) and difficult to interpret.
• What is the size threshold for rubble-pile structure? The 2.2-hour spin barrier sets a lower limit of around 150 m, but the upper limit is unclear. Are 100-km bodies still rubble piles, or do they become consolidated by self-gravity? The largest asteroid directly visited — 53-km Mathilde — is highly porous, suggesting the upper limit may be higher than once thought.

Common pitfalls

• Conflating bulk density with mineral density. A 1.2 g/cm³ asteroid does not mean its rocks are 1.2 g/cm³; they are 2.6 to 3.5 g/cm³, with the rest of the volume being holes. Always state porosity explicitly when quoting density.
• Treating rubble piles as fluids. They are not. They are granular materials with finite angles of repose, intermittent particle flow, and stick-slip dynamics. The relevant analogue is sand, not water — and even sand on Earth's surface gravity is not a perfect model for sand under microgee, where electrostatic and van der Waals forces become non-negligible.
• Assuming all small asteroids are rubble piles. Below ~150 m, fast rotators dominate; those are monolithic. Above that scale, cohesionless behaviour kicks in and rubble piles dominate. The transition is set by the comparison of tensile strength to self-gravity.
• Ignoring the role of ice. Comet nuclei and trans-Neptunian objects show similar low bulk densities to inner-belt rubble piles, but in their case much of the missing mass is ice, not rock-and-vacuum. The structural class is similar but the cosmochemistry is different.
• Generalising β from DART to every kinetic impactor. Dimorphos's β ≈ 3.6 reflects its specific composition, porosity, and impact geometry. Other targets could give β between 1 and 5; mission planners must allow for the uncertainty.