People have be filling it with different gasses to get different colors.
More specifically, for a given exhaust velocity and grid spacing, the space charge limited thrust density (thrust/area) of an ion engine scales as the square of the mass/charge ratio of the ions. So you really want heavy singly charged ions. This is completely unlike thermal rockets, where you want low molecular weight exhaust gases.
Plasma engines that accelerate a quasi-neutral plasma aren't subject to space charge limits, but even there heavy ions help because they reduce the energy used in ionizing the propellant per unit propellant mass.
With Radon it might even be conceivably possible (not sure how hard it is to get and if any restrictions apply because of its radioactivity), and it would work for a few years, because it has a half-life of 3.825 days (EDIT: this is of course complete nonsense, the "." is a decimal point, so it will only work for a few days). In the quantities needed for a gas tube (and as long as it stays in the tube!), I guess it should also be relatively safe, but I'm not an expert. Apparently it produces red light when used in a tube. Oganesson however has a half-life of 0.7 ms, so, aside from how expensive it would be to synthesize enough of it, it's doesn't stay around long enough for any experiments...
In a year the radon would've undergone about a hundred halvings, so around one 10^30th of the original radon nuclei would be left. Which is to say, almost certainly zero. One mole worth of radon would've decayed down to the last atom after less than 300 days (mostly to lead-210, which would then comparatively slowly decay to stable lead-206 with a half-life of about 22 years).