The DU Lounge

The high temperature reached by the explosion (100s of millions K usually cited) will volatilize anything, including refractories, and an explosion near the surface will eject a great deal of mass, including rock, dust, vapor, and plasma, all serving as reaction mass for a very short-lived rocket engine. I don't know what calculations have shown the optimum depth to be, but it may be a substantial fraction of the diameter. You don't want to waste the energy of the explosion on high-velocity shrapnel so much as create as much hot, compressed, low-molecular-weight gas as possible for that temperature -- the most effective reaction mass has the lowest momentum per particle, with individual atoms (in practice, molecules) being optimum. The hot exhaust may continue to escape to the surface for some time after the individual explosion, suggesting even a tunnel between the bomb and surface, which only needs to last long enough to enable most of the gas to escape.

Comets, made of mostly volatile materials, should be particularly effectively deflected, but also in greater danger of falling apart.

PS: An explosion directed out of the ecliptic plane should move a comet or asteroid into a more out-of-plane orbit, with fewer opportunites for close approaches. This may represent the most beneficial change for the least amount of energy.

PPS: IIRC, this is how the inkjet printer was discovered. An engineer accidentally touched a hot soldering iron to a syringe containing liquid, and the increase in pressure caused it to shoot out a few droplets of liquid.