Completely reading the abstract states that even the control test, in which the device was modified to NOT produce thrust, still produced thrust.
That indicates, to me, a problem with the measurement.
When your control produces positive results when you were expecting negative results, you're testing it wrong. Without a stable, predictable, easily measurable control, these results are meaningless.
Nasa already has an "electric propulsion system." Look up info on their ion propulsion research. Sure, it requires a gas to be ionized, but it still provides long lasting continuous thrust that can be used for small satellites. This system, although novel if it works, will not revolutionize the space industry. We already have the capability of exploring every planet and asteroid in our solar system. Beyond that it doesn't matter how much fuel you have, you won't reach it in a very long time.
EDIT: Also, for the record, every "paper" referenced in this article is simply a conference paper, meaning it hasn't completely gone through the rigors of peer review. None of them are actually "published."