At least 10% (3 out of 28) of the largest known impact craters on Earth and a
similar fraction of all impact structures on Venus are doublets (i.e. have a
companion crater nearby), formed by the nearly simultaneous impact of objects
of comparable size. Mars also has doublet craters, though the fraction found
there is smaller (2-3%). These craters are too large and too far separated to
have been formed by the tidal disruption of an asteroid prior to impact or
from asteroid fragments dispersed by aerodynamic forces during entry. We
propose that some fast rotating rubble-pile asteroids (e.g. 4769 Castalia),
after experiencing a close approach with a planet, undergo tidal breakup and
split into multiple co-orbiting fragments. In some cases these fragments
evolve into stable binary systems, which re-encounter and impact the planet
during a later pass, creating two distinct craters.
To test this idea, we modeled close encounters between between fast-rotating
contact-binary asteroids, our first-order approximation for rubble-pile
asteroids, and a chosen planet. Our results show that Earth's tidal forces
frequently create binary asteroids, but that the separation distance between
the binary's components is almost always too small to produce a doublet crater
at impact. However, once the components are orbiting one another, small
perturbations from repeated distant Earth encounters, along with mutual tidal
forces between the components, frequently increase the separation distance
between the components in a random-walk fashion. To model these effects, we
combined our numerical model of planetary encounters with a Monte-Carlo code
that computes the frequency and characteristics of repeated Earth encounters
as well as mutual tidal effects occurring between Earth encounters. Our
results show that ~15% of all Earth-crossing asteroids evolve into co-orbiting
binary asteroids with well-separated components. Asteroids on solely
Mars-crossing orbits produce a smaller fraction of binaries (< 5%).
Folding these results into another model treating impact encounters between
binary asteroids and a chosen planet, we found we could duplicate the observed
fraction of doublet craters found on Earth, Venus, and Mars. Our results
suggest that any search for asteroid satellites should place emphasis on
km-sized Earth-crossing asteroids with short-rotation periods.