Successful kinetic impact into an asteroid for planetary defence

The DART spacecraft was launched on 24 November 2021. The spacecraft carried a narrow-angle imager called the Didymos Reconnaissance and Asteroid Camera for Optical navigation (DRACO), which was used for optical navigation, terminal guidance and asteroid characterization. DRACO detected Didymos—the primary asteroid in the binary system—in summed optical navigation images 61 days before impact. On 27 August 2022, 30 days before impact, DRACO began taking optical navigation images of Didymos every 5 hours, which were processed by the ground optical navigation team.

On 26 September 2022 at 19:09:24 utc, 4 h and 5 min before impact, the spacecraft’s autonomous Small-body Maneuvering Autonomous Real Time Navigation (SMART Nav) system took control of spacecraft navigation (Fig. 1a). SMART Nav processed DRACO images onboard to identify Didymos, and once resolved, Dimorphos. To achieve impact with Dimorphos, the spacecraft needed to distinguish between the two asteroids and hit the smaller, dimmer one. Owing to the dynamics of the binary system and the spatial resolution of DRACO, it was known that the spacecraft’s ultimate target—Dimorphos—would be hidden from DRACO’s view during most of the autonomous phase. By design, SMART Nav manoeuvred the spacecraft towards Didymos until Dimorphos became reliably detected. SMART Nav first detected Dimorphos 73 min before impact, and at 50 min before impact, SMART Nav began manoeuvring towards Dimorphos (Fig. 1). As planned, SMART Nav manoeuvring ended at 23:11:52 utc, 2.5 min before impact, to give the spacecraft time to settle to minimize jitter and smear in the final images. The spacecraft impacted Dimorphos at 23:14:24.183 ± 0.004 utc (Methods). From the start of autonomous navigation until impact, the spacecraft continuously streamed images to the ground, which were immediately broadcast to the public. The final full image was acquired 1.818 s before impact and has a pixel scale of 5.5 cm. The final image received on the ground was a partial image acquired 0.855 s before impact with a pixel scale of 2.6 cm.

Little was known about the shape or surface of Dimorphos until DRACO obtained high-resolution images. Ground-based radar observations led to a diameter estimate of 150 ± 30 m. Analysis of telescopic photometric observations yielded a comparable diameter for Dimorphos, 171 ± 11 m. Although DRACO imaged only a portion of Dimorphos, and illumination was limited at 60° solar phase, the images were used to construct an asteroid shape model (Methods). The shape model revealed Dimorphos to be an oblate spheroid with a volume-equivalent diameter of 151 ± 5 m (Table 1, Fig. 2 and Extended Data Fig. 1). The shape of Dimorphos is unusual relative to other near-Earth asteroids visited by spacecraft and differs from other binary asteroid secondaries observed so far that have measured elongations. However, oblate secondaries show little or no measurable light-curve amplitude, which biases the observational sample towards elongated secondaries. A size estimate for Didymos from DRACO images (Table 1 and Methods) combined with previous telescopic observations enables calculation of a more accurate visible (0.55 μm) geometric albedo for the system of 0.15 ± 0.02. This value is on the low side, but within 1σ, of the mean geometric albedo for S-type asteroids.

The spacecraft trajectory and pointing were reconstructed to locate the impact site (Methods and Fig. 2). The spacecraft impacted Dimorphos at 8.84 ± 0.45° S, 264.30 ± 0.47° E, within 25 m of the centre of the figure, which is very near the scenario for maximizing momentum transfer with an impact through the centre of the figure. The 1σ uncertainty in the impact site location is ±68 cm (Methods), which is smaller than the size of the spacecraft bus (Fig. 3). The impact angle was 73 ± 7° from local horizontal (Methods and Extended Data Fig. 2). The impact site was near two large boulders, labelled boulder 1 (6.5-m long and at its highest point standing about 2.2 m above the surrounding terrain; Extended Data Fig. 3) and boulder 2 (6.1-m long and at its highest point standing about 1.6 m above the surrounding terrain; Extended Data Fig. 3) in Fig. 2b,c. The spacecraft approached the asteroid with its solar arrays canted slightly towards the surface (Fig. 3). The leading edge of the +Y solar array contacted the surface of boulder 1, and this solar array directly hit boulder 1 (Fig. 3). Almost immediately thereafter, the −Y solar array grazed boulder 2, with the leading edge of the −Y array contacting the surface near the base of boulder 2 in downrange direction (Fig. 3). Finally, the spacecraft bus hit between boulders 1 and 2 (Fig. 3). Although the solar arrays contacted Dimorphos just before the spacecraft bus, the bulk of the spacecraft’s energy was transferred by the bus, which accounted for about 88% of the spacecraft mass at the time of impact.

DART images of Dimorphos revealed a boulder-strewn surface (Fig. 2a, and Extended Data Figs. 4 and 5) resembling other small near-Earth asteroids, such as the S-type (25143) Itokawa, and carbonaceous asteroids (101955) Bennu and (162173) Ryugu, suggesting a rubble-pile structure for Dimorphos. The boulder-rich nature of the surface is apparent in images as coarse as 2–3-m pixel scale (Extended Data Fig. 4). No unambiguous impact crater candidates are observed, which indicates a young surface, although craters can be difficult to identify on boulder-covered terrains. The appearance of Dimorphos contrasts with impressions from lower-resolution images of the Didymos surface, where regional roughness variations are apparent (Extended Data Fig. 4).

The final complete DRACO image captured an approximately 880-m² area surrounding the impact site at a 5.5-cm pixel scale. The impact region (Fig. 2b) exhibits blocky terrain resembling the rest of the observed hemisphere (Fig. 2a, and Extended Data Figs. 4 and 5). There is evidence for variation within boulders, ‘rocks on rocks’ similar to observations on Bennu, and partially buried boulders (Extended Data Fig. 5). The longest axes of boulders counted in the final complete image are 0.16 m to 6.5 m in length. The impact region has fewer boulders in the 0.2–0.5 m size range than expected if the cumulative distribution followed a single power law, even though the pixel scale of the image is sufficient for their identification (Extended Data Fig. 6). There is no evidence for expansive smooth deposits (grain size smaller than the image pixel scale) such as those seen on Itokawa. The blocky nature of the impact site probably influenced crater formation, ejecta and momentum enhancement, as seen in impact experiments, numerical simulations and the Small Carry-on Impactor experiment on Hayabusa2.

DART did not measure the mass of Dimorphos. Instead, the mass of Dimorphos is estimated using the orbital properties of the binary system, the total volume of the system and an assumption that Didymos and Dimorphos have equal bulk densities (Methods). This assumption cannot be rigorously tested from DART data, but this approach leads to a bulk density of Dimorphos of 2,400 kg m⁻³ (Table 1) with difficult-to-quantify uncertainties.

On the basis of analyses of reflectance spectra, the best meteoritic analogues for Didymos are L and LL chondrites. L and LL chondrites have grain densities of 3,580 ± 10 kg m⁻³ and 3,520 ± 10 kg m⁻³, respectively. If one assumes Dimorphos has the same composition as Didymos and that these meteorite values represent the grain density of Dimorphos, then a bulk density of about 2,400 kg m⁻³ implies a Dimorphos bulk porosity of the order of 30% (with a difficult-to-quantify uncertainty; Methods). This level of bulk porosity is not inconsistent with a rubble-pile structure for Dimorphos, a structure suggested by the boulder-rich character of Dimorphos’s surface. This bulk porosity probably exists as a combination of macroporosity between pieces of rubble and microporosity within individual pieces of rubble. L and LL chondrite samples have porosities of 8.0 ± 0.3% and 9.5 ± 0.6%, respectively, which would imply that macroporosity is substantial on Dimorphos. Estimates of the density and porosity of Dimorphos will improve when the European Space Agency’s Hera mission arrives at the Didymos system in early 2027.

DART’s successful autonomous targeting of a small asteroid with limited prior knowledge is a key accomplishment on the path to advancing kinetic impactor technology to an operational capability. The impact of DART indicates that a precursor reconnaissance mission is not a prerequisite for intercepting a subkilometre asteroid, although the characterization done by a precursor mission would provide valuable information for optimizing, planning and predicting the outcome with greater certainty. Kinetic impactor technology for asteroid deflection requires having sufficient warning time—at least several years but preferably decades—to prevent an asteroid impact with Earth. Nevertheless, this successful step to demonstrate the viability of kinetic impactor technology for planetary defence builds optimism about humanity’s capacity to protect Earth from an asteroid threat.