There are techniques with which deorbiting 1 million 10 gram fragments might end up being quite cheap. E.g. the use of a single laser-equipped base that selectively hits the debris to cause localized evaporation and thrust, disrupting their trajectory.
If a tiny bit of undirected evaporation is enough to push the fragment to deorbit, it's already so low that it would have deorbited on its own very soon anyway.
In that case evaporating it completely would be a better option, as a gas it's harmless
