> But, Burns asks, what if the ring’s mass is much greater when it slides in one direction than the other? ... In fact, a simplistic implementation of Burns’s concept would be to replace the ring with a circular particle accelerator, in which ions are swiftly accelerated to relativistic speed during one stroke, and decelerated during the other.

OK... that's great, but how do you plan to quickly accelerate particles without expending lots of energy (and without receiving a 'reaction')?