Your voltage standoff goes with distance, and inverse
to doping where your blocking depletion region will be.
You want a good low ohmic backside for current but
a lightly doped layer above, of high quality. An implant
that was capable of hammering the backside and broad
(tWafer-xDeplExtent) would (a) probably be impossible
to get - check out energies required for, say, a 20um
depth) and (b) really damage the intervening silicon.
Epi growth is well known art, clean and controllable
and about as defect free as the starting material and
equipment set.