For decades, comet scientists have debated whether comet nuclei are primordial
rubble piles, formed at their current sizes through gentle accretion in the Solar Nebula,
or if they are collisional rubble piles formed in the aftermath of violent collisions
between larger parent bodies. The Rosetta mission to comet 67P/Churyumov-Gerasimenko and
the Stardust sample-return mission to comet 81P/Wild 2, combined with observations by
Cassini or from ground of irregular giant planet satellites captured from the primordial
disk, have provided a variety of physical, mineralogical and chemical information that
allow us to revisit the problem of comet formation with greater confidence than
previously.
The emerging picture is that thermal processing due to short–lived radionuclides, combined
with collisional processing during accretion in the primordial disk, is expected to create
a population of medium–sized bodies that are comparably dense, compacted, strong, heavily
depleted in supervolatiles and that have experienced extensive aqueous alteration due to
the presence of liquid water. Irregular satellites Phoebe and Himalia are potential
representatives of this population. Collisional rubble piles inherit these properties from
their parents. Contrarily, comet nuclei have low density, high porosity, weak strength,
are rich in supervolatiles, and do not display convincing evidence of in situ aqueous
alteration. Therefore, comet nuclei do not resemble collisional rubble piles, but display
all properties expected for primordial rubble piles.
We outline a comet formation scenario that starts in the Solar Nebula and ends in the
primordial disk, that reproduces these observed properties, and additionally explains the
presence of extensive layering on 67P/Churyumov–Gerasimenko (as well as on 9P/Tempel 1
observed by Deep Impact), its bi–lobed shape, the extremely slow growth of comet nuclei as
evidenced by recent radiometric dating, and the low collision probability that allows
primordial nuclei to survive the age of the Solar System.
We argue that TNOs formed due to streaming instabilities at sizes below ~ 400 km and that
~ 350 of these grew slowly in a low–mass primordial disk to the size of Triton, Pluto, and
Eris, causing little viscous stirring during growth. We thus propose a dynamically cold
primordial disk, that prevented medium–sized TNOs from breaking into collisional rubble
piles, and allowed for the survival of primordial rubble–pile comets. We argue that comets
formed by hierarchical agglomeration out of material that remained after TNO formation,
and that this slow growth was a necessity in order to avoid thermal processing by
short–lived radionuclides (that would have led to aqueous alteration and loss of
supervolatiles), and that allowed comet nuclei to incorporate ~ 3 Myr old material from
the inner Solar System.