Here's a wild thought for you.
When NaCl is dissolved in water the Na+ and Cl- ions are separated by
hydration. They each have a polarized layer of water about them that
separate them, insulates them, and reduces the force between them,
and thus causes the solvation. The energy for this separation comes
from the heat of the water.
Looking at the Kanzius videos, I can see that all the bubbling action
is right at the surface of the test tube. This suggests to me the
possibility that the electrostatic gradient at the solution surface
might be pulling out fully solvated ions, alternate ions on alternate
phases. These now gas phase hydrated ions quickly gain kinetic
energy from the EM radiation, typically opposed direction kinetic
energy, and thus collide and strip off their enveloping water
molecules, and thus recombine, giving off the energy of the reaction
in the form of light and heat. In addition, some of the water is
ionized and recombines as well. The flame is thus essentially an RF
arc, as Bill Beaty suggested earlier. However, if this is all
correct, the mechanism of the arc formation is a bit unusual I think.
It the above is correct, then it will not be possible to collect gas
from the flame, which I think Bill Beaty already suggested. It is
also true that, if the process is taken to completion, the weight of
the residue may be considerably less than the weight of the salt at
the start, due to loss of the nano-salt to air entrainment.
Also of interest is the fact the water in the test tube should
actually cool. The energy provided by solvation is fully expended in
the arc heat. A key ingredient to violating the second law then is
getting the salt back into solution by operating a closed cell.
Getting practical energy then is just a matter of using a Sterling
engine with the hot end heated by the arc and the cold end cooled by
the brine.
A practical design might be to use a flat metallic heat collector for
the Sterling also as a capacitor plate, hovering above the brine at
sufficient distance that the arc forms and the "flame" is sustained.
The other "plate" for the RF field is then provided by the brine
surface itself, with a conductive plate also provided within the
brine so as to minimize the brine resistance, i.e. at a shallow depth
and of the size of the opposed plate. The cool brine can then be
pumped to the cooling end of the sterling if that is necessary.
Best regards,
Horace Heffner
http://www.mtaonline.net/~hheffner/