Or, in this case, the “not”
The technique involved is neat: for some atoms you can find wavelengths where the AC Stark shift is the same for the two levels in the clock transition, so the atom is unperturbed by the presence of the trapping light. So you trap them in an optical lattice, with confinement like a far off-resonance dipole force trap (FORT). This means you can continue to confine the atoms while it is in the superposition where it is oscillating between the two clock states.
The big advantage of this method is that you can trap millions of atoms easily in an optical lattice and that should make such a clock much more robust than a fountain, while achieving at least the same kind of accuracy.
Actually, that’s not the big advantage. Fountains trap millions of atoms (even billions, depending on your collection technique). The advantages are that you’d keep that many atoms (fountains lose signal from the original collection because the cloud spreads out, so the number you toss is an order of magnitude bigger than the number that return), you can interrogate the atoms for a longer period of time (an advantage shared by ion trap clocks/frequency standards) and avoiding cold-collision frequency shifts (atoms in close proximity tend to interact strongly, as they can interact for a relatively long time, and this changes the state of the atom, introducing an error in the signal)
However, “at least the same kind of accuracy” isn’t enough. I’ve noted before that international standards are a political issue. Cesium beam standards are commercially available. Furthermore, dozens of labs have or are building fountains, at some investment of time and money to gain the expertise in doing so (because atomic fountains are not, nor are they likely to become, a commercially available item). The countries doing this will likely be reluctant to switch to a standard that requires even more money and acquired expertise in a new technique for marginal gain in accuracy and precision. Especially in light of how many new options for secondary standards have emerged in just the last decade — an even better candidate may emerge as technology advances.