Re: why not just use the electricty to accelerate a gas ?
Good question: the US Navy has the answer. They're developing a refinement of the Farnsworth-Hirsch Fusor - an electromagnetic grid/inertially-confined fusion reactor that can run on tritium, deuterium, or boron-11. The basic configuration theoretically can be four meters across and develop a gigawatt.
Following me yet? A four-meter gigawatt reactor with very compact fuel can use Lorentz electric rockets to expel a number of ionizable working fluids (lithium perhaps the most practical) at very, very high specific impulse.
Perhaps the specific impulse from using the gigawatt of electricity from the USN's compact fusion reactor to run one or more Lorentz engines is not as high as a thermonuclear detonation, but the scheme would be much, MUCH cheaper, less politically fraught (since the article is basically describing something that could be altered to be a fully-automatic, high rate of fire micro H-bomb cannon here on Earth or up in the satellite belt), and very scalable.