As always when I read quantum optics papers, I'm struck by the distance between experimental reports and textbook elementary theory. I suppose the quantized electromagnetic field is being used, with a PoincarГ© invariant vector |0вЊЄ, and squeezed states are something of a variation of coherent states that are theoretically described by the action of an exponentiated quadratic creation and annihilation operators, which one can see here. [For such a theory fluctuations of the EM field are clearly expected in the vacuum state, but they do not represent "energy" (which is not a measurable local variable; the vacuum vector is the unique eigenvector corresponding to the minimum eigenvalue of the Hamiltonian -- a squeezed state has higher energy than the vacuum state), nor do fluctuations represent photons coming in and out of existence (photons, at least by the Wigner definition, are also not local objects).]
For the experimentalist, however, the squeezed states are produced in this case by a crystal: "12-fs near-infrared (NIR) pump pulse (red/yellow envelope) and mid-infrared (MIR) vacuum fluctuations (green band; with amplitude О"Evac) co-propagate into a generation crystal (GX) with second-order nonlinearity". It's only by having something like a specific type of crystal available that a squeezed state can be prepared.
Measurement of the state is as esoteric (in the sense that it is distant from an operator-acting-on-Hilbert-space description), for which "A 6-fs probe pulse (blue envelope) is superimposed to sample the electric field amplitude as a function of delay time tD in an electro-optic detector crystal (DX)", after which the electric field that is generated by the whole process is measured over a period of about 100 femtoseconds (from the Science article, using a "radio-frequency lock-in amplifier").
Essentially no manipulation of the electromagnetic field is possible without the use of matter to introduce a significant degree of nonlinearity, so an alternative description would seem to be that the various light sources are used to measure the nonlinear properties of the various macroscopic crystals rather than that the light and various macroscopic crystals are used to measure the vacuum.
This slightly cantankerous critique of the language and mathematics of the paper linked to does not extend to criticism of the experimental merit. I look forward to other comments. I can't legally attach the Nature paper, but I guess it's OK to attach Figures 1 and 2.Attachment #1: Capture.JPG