Actually, i don't quite get how the experiment tested E=mc². I am not at work this week so i cannot consult the Nature article, which i will be certainly doing next week.
Here is what i got out of the link:
When a neutron is captured by an atom, the total mass of the atom with one extra neutron is
[tex]m_{total} = m_{atom} + m_{neutron} - \frac{E_{binding}}{c^2}[/tex]
From experiment (with the magnetic traps, measuring the revolutions about the B field lines) they acquired the total mass (silicon ion with neutron) and the mass of a Silicon ion.
The [tex]E_{binding}[/tex] consists out of emitted gamma rays and a recoil energy of the nucleus. This is all straighforeward. They measured the gamma ray energy.
So from experiment we have both the gamma energy, the ionic mass and the mass of the ion WITH the neutron. In order to verify if E=mc², we need to have the [tex]E_{binding}[/tex], right ? Well, the one thing they are missing is the nuclear recoil. How did they deal with that ? Probably it was much smaller than their error margin ?
regards
marlon