Perhaps you could do the same with Berylium Oxide?
It also has a hardness of 9, but lower density at 3.016, (while Al2O3 is 3.4 to 4) for space applications. My Hawley's Chem Dictionary mentions that it is undamaged by nuclear radiation (low Barns cross section, I believe), and has high heat stress resistance. The m.p. is 2570 C versus 2030 C for the alumina. Really awesome stuff, though I suppose carbon nanotubes out do it greatly in tensile strength and m.p.. Still, it could provide rigidity and hardness for heat shields, maybe as a light weight matrix to fuse with the carbon fibers.
Best of luck on your enterprise.