Impact-crater ejecta on Bennu indicate a surface with very low strength

14 15 A planetary surface’s resistance to change is generally described as its “strength” 16 (units of stress). The surface strength of small, rubble-pile asteroids, which consist of 17 fragments of larger bodies that were collisionally disrupted, is poorly constrained due 18 to their wide departure from terrestrial analogs. Here, we report the observation of 19 an ejecta deposit surrounding an impact crater that limits the maximum surface 20 strength of the near-Earth rubble-pile asteroid (101955) Bennu. The presence of this 21 deposit implies that ejecta were mobilized with velocities less than the escape velocity 22 of Bennu, 20 cm/s. Because ejecta velocities increase with surface strength, the ejecta deposit can only be explained if the effective strength of the surface material near the crater is exceedingly low, ≤ 100 Pa. This is three orders of magnitude below values commonly used for asteroid surfaces, but is supported by previous observations of an artificial impact crater on a similar asteroid, Ryugu. Our findings indicate a mobile surface that has likely been renewed multiple times since Bennu’s initial assembly and have far-reaching implications for interpreting observations of Bennu and other rubble piles.

hemisphere. In multispectral images, the color of the crater and the surrounding smooth area 62 is more homogenous than that of the rest of Bennu's surface, which varies at the scale of 63 boulders (meters to tens of meters) 12 . This region shows a distinct b′/v normalized band 64 ratio > 1, which is typical for smoother and younger (as inferred from space-weathering 65 trends) terrains on Bennu (Fig. 1d, S3). The surface is smoother by a factor of 2 than the 66 Bennu average, as determined by measures of roughness such as variations in slope over 67 length scales of 1 to 5 m 13 and tilt variation (Fig. 1c). There are two boulders to the northeast 68 ( Fig. 1a), behind which the terrain distal to radially from the crater is rockier and 2 to 5 m 69 lower (Fig. S1).  to elevation and a volume of 9x10 3 m 3 ±50% 14 . The crater resides on a ~23° regional slope 94 ( Fig. S4 shows the detailed topography). The northern crater wall has a steeper slope than 95 the southern wall, which has more large boulders.  Because the smooth, uniform terrain surrounds and inhabits the crater, we infer that they 110 formed concurrently. A crater that post-dated the terrain would have roughness and color 111 that differed from the surrounding terrain, and a crater that pre-dated the terrain would 112 show evidence of infilling, particularly at the downslope crater wall, which would be shal-113 lower than the upslope wall, instead of steeper as we observe. We therefore conclude that 114 the material that composes the uniform terrain is a product of or triggered by the cratering  Further, the uniform ring of terrain on and just beyond the rim uphill and to the south, east, 118 and west (Fig. 1a) is best explained by material that left the crater in those directions rather 119 than by mass wasting, which would create a more unilinear set of features (c.f.) 16,17 . We 120 therefore infer that this terrain consists of ejecta from the impact that formed Bralgah crater. For ejecta to fall back onto Bennu's surface, the particles must be ejected at speeds lower 125 than Bennu's escape velocity of 20 cm/s 18,19 . Impact-scaling relationships (Table 1), devel-126 oped from terrestrial testing and combined with assumptions about impact velocity and ma-127 terial properties, enable parameters such as ejecta velocities to be estimated from the crater 128 size 20,21 . The stronger and more cohesive the surface material, the higher the ejection veloc-129 ities. Analyses of crater formation are parameterized in either a strength or gravity regime.

130
(Armoring, when the impactor is smaller than the target particle, requires different anal- ries that re-impact Bennu within a crater diameter (Fig. 3).

142
On a slope, an ejecta deposit is asymmetric even for impacts that occur at near-normal inci-   Table 1. Scaling relationships for both gravity and strength regimes used for calculations 153 and simulations of impact cratering and resulting ejecta 21 . R is the final crater radius; Y is a 154 measure of surface strength (Pa); is surface density; U, , and m respectively are impactor 155 velocity, density, and mass; and x is the radial distance from the crater center. , , C1, C4, H1, 156 and H2 are fitted constants.

Parameter
Values and Relationships Crater radius (strength regime) ( ) Crater radius (gravity regime) ( ) Transition strength = (The gravity regime applies when surface strength is less than Yt; for Bennu, Yt <1 Pa for a < 15 m) Ejection velocity At the crater radius (x=R) simplifies to v~=√ Ejection velocity (gravity regime) At the crater radius (x=R) simplifies to v~=√ / Mass ejected faster than v (strength regime) Mass ejected faster than v (gravity regime)  this plot shows the fraction of total ejected mass that has speeds below the plotted value.

171
Only 20% of the ejected mass is ejected faster than 8 cm/s.   The flow field and ejecta blanket are likely related. Downslope ejecta would have re-con-228 tacted Bennu at relatively shallow angles of 25 to 30° to the surface and with velocities of 5 229 to 7 cm/s tangent to the surface (Fig. 3b). This velocity is sufficient to dislodge particles. For 230 example, assuming that the scaling laws remain valid for impacts at very low speeds, a 10-231 cm-diameter ejecta particle returning to a strengthless surface at 5 cm/s would create a 40- be such a collection (Fig. 1a). The lack of boulders larger than a few meters in the uniform 249 terrain suggests that larger boulders have been removed or buried.  There are a few other large candidate craters at high latitude, but they appear older and de-295 graded 17 , and it is possible that any associated ejecta or flow fields are weathered or dis-296 turbed past recognition. We estimate the size of the impactor that made Bralgah Crater by first using the gravity-314 regime parameterizations. Using the range of values in Table 1, the impactor had a radius, a,   Mapping and measurements 473 We mapped the ejecta blanket and flow field on a global OCAMS/PolyCam mosaic of Bennu 474 with a pixel scale of ~5 cm/pixel 11 (Fig. 1a) and on OCAMS/MapCam image 475 ocams20190322t233553s104_map_iofl2pan_78685 (Fig. 1b), which was collected on 22

476
March 2019 and has a pixel scale of 0.29 m/pixel. Elevations, slopes, and tilts are from SPC 477 shape models 13 . Tilt variation (Fig. 1c)  Applying laboratory-based scaling relationships (  Constraining surface strength 543 We constrain the possible strength by investigating several parameterizations and placing a 544 value that is likely an extreme but that encloses many of the possible conditions. To set a 545 maximum value on strength of the surface material ejected during formation of Bralgah 546 Crater, we examine the slowest ejecta speeds, which increase as surface strength increases.

547
The slowest ejecta are launched near the crater rim in the final stages of the cratering pro-548 cess. Using one crater radius as the distance that certainly contains ejecta, the minimum de-549 duced ejecta speed is 3.5 cm/s, the speed required to land within one crater radius 550 downslope. (There appear to be ejecta closer than one radius from the crater rim, so this is 551 a conservative speed.) The solid lines in Fig. S2a are the slowest available speeds using the 552 equations for ejecta velocities in Table 1 for different material properties. If surface strength 553 is >20 Pa, then no ejecta for any of the analyzed materials will be sufficiently slow (red line 554 in Fig. S2a) to land as close as 1 crater radius from the rim. The strength may in fact be much less, but that cannot be discerned from comparing the 568 Bennu observations to the results of terrestrial testing.  Table 1 and published parameters 21 for the different material analogs for Bennu's regolith.

577
WCB is weakly cemented basalt, and "Base" has the constant C3=1 in the strength equation  Figure S1 shows the drop in elevation to the north-620 west.