
Last September, a NASA spacecraft designed to assess humanity’s planetary defense capability approached an asteroid not much larger than the International Space Station. Moments before the Double Asteroid Redirection Test (DART) probe rammed its target into a new orbit, it beamed back photos of what looked like a gravel pit strewn with boulders.
Not so long ago, some planetary scientists would have been surprised to learn that the terrain of a small asteroid resembled a rock quarry. But after a quarter-century’s worth of dynamical modeling, telescope surveys, and asteroid rendezvous in space, DART scientists were confident that the target of their deflection mission—a 150-meter-wide asteroid called Dimorphos, which orbits a larger asteroid named Didymos—is not a monolithic mountain of rock. It is a conglomeration of boulders, stones, and sand that is bound only by gravity. Or, to use the technical term, it is a rubble pile.
DART is the latest of several probes that have allowed planetary scientists to study rubble-pile asteroids up close. It is now clear that many of the solar system bodies measuring less than a few hundred kilometers across are loosely bound aggregates of ancient debris accumulated during hundreds of millions of years’ worth of asteroid collisions. The rubble-pile population also includes nearly every object that has been pulled from the asteroid belt into orbits that bring them close to Earth’s. If researchers want to scrutinize the evolution of rocky objects in the solar system and to safeguard our planet against catastrophic meteor strikes, they will have to make sense of the physics of rubble-pile asteroids.
“If something is going to come for us, it’s going to be a rubble pile,” says Erica Jawin, a geologist at the Smithsonian National Air and Space Museum in Washington, DC. “It’s really important to understand the dynamics and nuances of rubble piles in case we have to deflect one.”
Compiling evidence
Objects in the asteroid belt are defined by their collisional pasts. The more than half-million known asteroids that reside there are likely the remnants of a handful of massive rocky bodies that were forced into orbits between Mars and Jupiter early in the solar system’s history. Over time, collisions have broken up those parent bodies and sent some of the shrapnel into orbits rendered unstable by the gravitational influences of Mars, Saturn, and especially Jupiter. Many of those asteroids have gotten ejected from the belt and forced into eccentric orbits that take them through the inner solar system—and potentially on collision courses with Earth.
For decades the conventional wisdom was that most asteroids, whether belt residents or near-Earth objects, are monolithic chunks of rocky debris that broke away from larger bodies during individual collisions. That view began to change about 50 years ago, when Clark Chapman, Donald Davis, and their colleagues at the Planetary Science Institute in Tucson, Arizona, calculated how asteroids would fare after colliding with projectiles of different sizes.
One of their key findings was that an impactor even a tenth the size of an asteroid would have enough energy to shatter—not just partially fragment—its target. The debris from such collisions, however, would not have the energy to escape into the void. Extrapolating from that scenario, Chapman and Davis hypothesized that the shards would gradually fall back together, beholden to weak but nonnegligible gravitational wells created by the mass of the collective parts. Chapman says he recalls a 1977 conversation with Davis during which they walked through how a once-monolithic asteroid could get shattered on multiple occasions, with a significant subset of the remains recombining to form an increasingly fragmented body each time. Chapman called such an aggregate object a rubble pile.

This new idea, that the solar system is teeming with consolidated rubble, did not garner much of a reaction, Chapman says. After all, no telescope could provide high-resolution views of small asteroids located millions of kilometers away. But then, in 1993, astronomers discovered a comet—or at least what was left of one—spiraling toward Jupiter. Comet Shoemaker–Levy 9 had broken into pieces that were drawn out like a string of pearls. The icy object had presumably fallen victim to tidal forces from the giant planet during a close approach the year before.
Yet the physics did not make sense, says Derek Richardson, an astronomer at the University of Maryland. Despite the outsize gravitational influence of Jupiter, a comet with a monolithic nucleus should not have torn apart so easily. “The body had to be virtually strengthless,” he says. The Shoemaker–Levy 9 spectacle inspired Richardson and others to investigate how the collision-dominated evolution of small objects in the solar system affects their structure.
Richardson teamed with Patrick Michel, an astronomer at the Côte d’Azur Observatory in Nice, France, to simulate the collisional dynamics numerically. Using Michel’s hydrodynamics code, the researchers modeled collisions of a 300-kilometer-wide basalt sphere with various projectiles. Once each hypothetical asteroid had shattered, Richardson’s planet-formation model calculated the gravitational interactions among the fragments. In a 2001 paper written with Willy Benz and Paolo Tanga, Michel and Richardson reported that a single collision would yield, within a day or two, scores of rubble piles.

Although planetary scientists had yet to see a small asteroid up close, observational evidence was arriving that supported the new modeling work. Through wide-field surveys, researchers were amassing asteroid brightness measurements to estimate various bodies’ sizes and rotation speeds. Asteroids are susceptible to spin-up because of the subtle torques that result when their irregularly shaped surfaces re-emit absorbed sunlight.
Plots of the relationship between size and spin rate revealed a clear feature: Very few asteroids have a rotational period of less than about 2.2 hours. The implication, says Kevin Walsh, a principal scientist at the Southwest Research Institute, is that most asteroids do not have sufficient internal strength to rotate rapidly; they would fling themselves apart. The 2.2-hour rotation rate is “right about where things would start to fly off the surface due to centrifugal force,” he says. “That problem isn’t relevant for a monolithic chunk, but an aggregate has a spin limit.” That lack of internal strength was also consistent with the ease with which Jupiter’s tidal forces were able to dismantle Shoemaker–Levy 9.
Ball pits in space
In 2005 the Hayabusa spacecraft launched by the Japan Aerospace Exploration Agency (JAXA) arrived at a near-Earth asteroid named Itokawa. Measuring a half kilometer across at its longest axis, kidney-shaped Itokawa was the smallest asteroid to get a dedicated visit from a spacecraft and the second overall (following Eros, the landing site of NASA’s NEAR probe in 2001).

Images from Hayabusa revealed a surface that did not resemble that of any planet or moon in the solar system. Boulders and large stones dominated most of the landscape, with no fine-grained material visible at all. Itokawa looked like “a pile of rocks from someone’s garden or a mountain scree field,” as Walsh put it in a 2018 review article. The sand and pebbles appeared to have collected in several basins that stood out for their relative smoothness compared with the rest of the asteroid’s craggy terrain. Itokawa sure looked like a rubble pile.
In the years following Hayabusa’s rendezvous, JAXA’s Hayabusa2 visited the near-Earth asteroid Ryugu, and NASA’s OSIRIS-REx camped out for three years in the vicinity of another inner solar system interloper, Bennu. Though differing in several ways from Itokawa, both bodies fit the description of loosely bound aggregates that Chapman and Davis had first discussed in 1977. “Pretty much all near-Earth asteroids look like rubble piles,” Richardson says.
One lesson that has emerged from the recent missions is that rubble piles are remarkably dynamic considering they are composed of rocks that, apart from experiencing thermal effects and collision-induced fracturing, have not changed much over the past 4 billion years. Mass is constantly moving, and it can be pulled in different directions over time due to processes that become relevant only when a hodgepodge of granular material is soaring through space in a microgravity environment.

A rubble pile’s rotational speed is particularly important. If an asteroid spins faster than the period threshold of 2.2 hours, then centrifugal forces can overcome gravity’s precarious grip and yank rocks off the surface beyond their escape velocity. Researchers suspect that DART’s target, Dimorphos, formed when its now partner, Didymos, exceeded its spin limit and flung away some of its mass, with the ejected material eventually clumping together to form a moonlet.
Even when it is not unraveling an entire asteroid, rotation remains a major mover of mass. For a spinning body, the rotational potential energy is generally greatest at the poles, which, in the relative absence of gravity, encourages the movement of material toward lower latitudes. Bennu, which looks like a spinning top, is a textbook product of that migration. Despite hosting an ever-growing ridge of relocated material, Bennu’s equator remains a geopotential valley: A ball would roll up the ridge. “The geopotential doesn’t necessarily follow local topography,” says Jawin, who analyzed the migration of material on the asteroid.
All that material is not migrating over solid bedrock. With gravity serving as the sole binding agent, rubble-pile asteroids are exceedingly porous. In 2019 Hayabusa2 researchers reported that the volume of the kilometer-wide Ryugu is mostly empty space. The asteroid’s density is about 1.2 g/cm3, even though the carbonaceous minerals that the asteroid’s rocks are composed of have densities of at least twice that. Similarly, Itokawa has a porosity of about 44%. There is more space within those asteroids than inside a can of mixed nuts.
The porosity could explain the absence of sand and pebbles on Ryugu and Bennu. Fine-grained material probably percolates through the spaces between larger rocks. The result is an extraterrestrial realization of the Brazil nut effect: Over time, finer-grained material sinks into the asteroid’s interior, and boulders and big stones find their way to the surface.

Scientists with the OSIRIS-REx mission almost learned the hard way how fragile a porous, loosely bound pile of rubble can be. On 20 October 2020, the probe descended to Bennu to sample the surface. Upon touching down in one of the few locales not completely riddled with boulders, the spacecraft fired a puff of nitrogen to kick up material for collection and then quickly lifted away from the asteroid. It spent a mere six seconds on the surface. Still, during its brief visit, OSIRIS-REx sank half a meter into the gravel, like a toddler in a ball pit. The surface behaved like a viscous fluid, with “near-zero interparticle cohesion,” researchers reported in Science last year.
“Bennu didn’t push back on the spacecraft at all,” Jawin says. Had its thrusters continued nudging it downward, the lander could have continued to burrow under the surface. An astronaut standing on Bennu’s surface “might go all the way through the asteroid,” Jawin says. “Or you might stop when you hit a boulder part of the way down.”
Redirecting rubble piles
For all their fragility, rubble-pile asteroids are surprisingly durable. Their abundant pore space ensures that the energy from an impacting object will not propagate through the asteroid nearly as well as it would through a massive chunk of rock. An incoming projectile “energizes only the volume directly beneath where it hit,” Richardson says. “You excavate a big crater, but you sort of cushion the far side.”
Rubble-pile asteroids also enjoy impact protection from their surface boulders. When OSIRIS-REx researchers first analyzed Bennu’s surface, they were perplexed by the dearth of small craters. The reason, they eventually determined, is that boulders cover so much surface area that an incoming projectile is highly likely to strike one. At least some of the energy from the impact then goes into cracking the boulder rather than forming a crater.

Those factors introduced some uncertainty in the lead-up to DART, which marked the world’s first attempt to redirect a near-Earth asteroid. Prior to the impact, nobody knew what Dimorphos looked like, let alone where on the surface the spacecraft would strike. Researchers on the mission’s impact modeling team nonetheless had to predict how much momentum the spacecraft would transfer to Dimorphos: The greater the momentum transfer, the more Dimorphos would stray from its pre-impact orbit.
The simulations they ran indicated a large range of potential results, says Angela Stickle, a Johns Hopkins University Applied Physics Laboratory planetary scientist who leads the mission’s impact working group. A lot of the variance was due to uncertainty about Dimorphos’s mass and density, which in turn determine the porosity. The site of the impact was another key variable. If the bus-size spacecraft struck a boulder of similar size, then the momentum transfer would be limited, perhaps to the momentum carried by DART and nothing more. In the most optimistic scenarios, the probe would strike a relatively smooth area and excavate a large crater, sending chunks of Dimorphos flying off in the direction from which DART came. If that were to happen, the collective energy of the ejecta would enhance the impact, leading to an overall momentum transfer exceeding by several times that provided by the spacecraft alone.
Initial results from DART offer good news for planetary defense. The craft’s impact altered the previously 12-hour orbit of Dimorphos around Didymos by about 33 minutes, a team led by Cristina Thomas, a planetary astronomer at Northern Arizona University, reported earlier this month. That figure falls near the upper end of the team’s pre-impact prediction of 7–40 minutes and implies that the momentum from the impact was several times as high as that from the spacecraft itself.

The deflection effectiveness is all the more impressive considering that DART’s twin solar arrays each struck a boulder upon impact, potentially drawing energy away from the excavation of a crater. Apparently the recoil delivered by the estimated 10 million kilograms of rocky material dislodged in the collision more than overcame Dimorphos’s ability to soften the blow. “The moment right before impact, I remember thinking, ‘That’s going to be a lot of ejecta,’ ” Thomas says of seeing the first images of Dimorphos during DART’s final descent.
In three years, the European Space Agency’s Hera spacecraft is scheduled to arrive at the binary asteroid system to survey the DART-excavated crater. The probe also has a radar instrument that will help researchers assess Dimorphos’s structure beneath the surface rubble. For all the recent progress in understanding rubble piles, it is still possible that these asteroids are concealing large, monolithic cores beneath their rock-strewn surfaces. Hera’s measurements should help determine whether Dimorphos and its brethren are truly rubble piles through and through.