Novel atomic clocks: can quantum and many-body effects help? I am trying to learn if there are any proposals concerning the application of quantum and many-body effects to atomic clocks.
From what I understand, optical lattices have been used for timekeeping only due to a superior signal to noise ratio (SNR). The interactions between atoms are considered to be harmful as they introduce an extra frequency shift. I have a couple of questions related to this:
1) What is the reason for using a lattice instead of a continuum ultracold gas
in a trap?
2) Have people considered using quantum effects, like coherent superpositions of states (e.g., squeezed states) in order to improve these clocks?
3) Are there any many-body effects that actually help to increase the accuracy (besides increasing the SNR)?
 A: You might want to look up the chip scale atomic clock made by symmetricom.  It "resonates" with the hyperfine transition using CPT (coherent population trapping) weird QM Atom-light field interaction.  SNR is many times worse than the standard Rb atomic clock.      
A: I think you have about the right idea. To get the most accurate atomic clock possible, you want all the atomic transitions to be at exactly the same frequency, and if the atoms have substantial kinetic energy, this spreads out the possible transition frequencies. From a band structure perspective, turning on an extremely deep optical lattice makes the atoms all sit in a band that is extremely flat. Also, tight confinement of the atoms makes it almost certain that two will not occupy the same site, because the energy cost of the double occupation will be very high. This is in a sense a many-body contribution to the accuracy, although I'm not sure if it is what you had in mind. The end result is simply that each atom looks like it is in its own little harmonic oscillator trap, and then they do all sorts of work to make all of those oscillators have nearly the same frequency or at least know the deviations (paywalled article).  I'm not aware of any other proposals for improvements to lattice clock accuracy using many-body physics, although that certainly doesn't mean they aren't out there.
People are also looking at using squeezed states for lattice clocks, but it is at the proof-of-concept stage right now as far as I know. Here is an APS writeup, for example.
Edit: A slightly different issue is that Jun Ye is now using his lattice clock setup to study some many-body physics, as is mentioned for example on his website and this paper (paywalled). But these experiments are not in the same regime as the clock experiments themselves, and when these many-body effects do appear in their clock experiments they would be a source of noise that must be understood, not a benefit.
