4. An experiment
How to produce squeezed light
There is a variety of methods to produce
squeezed light.
Probably the most popular and so
far most successful one is the so called
parametric downconversion.
Let us look more closely at this specific form of the generation process of nonclassical states of the
light field, which was also used to generate the data shown in the previous chapter. We
overlap a
coherent state
of very small amplitude (the signal input) within a nonlinear
crystal with a light wave of twice its frequency (the pump wave). Due to the
polarizability of the crystal, the pump wave will induce a periodic change in the medium
of propagation of the signal wave. (See also the article about
nonlinear optics.)
The effect on the signal wave is very much dependent on its relative phase with respect
to the pump wave. Experimentally this phase is controlled by a mirror mounted on a
piezoelectric tranducer in the beam path of the signal wave. If the signal wave is in
phase with the pump wave, it gets amplified, if it is out of phase, it gets deamplified.
This behaviour is best visualized in
phase space.
The light field's state is depicted by
its uncertainty area in phase space, which is just the contour (a horizontal cut) of its
Wigner function. The action of the pump wave in the nonlinear medium (mathematically
denoted as the squeezing operator) is a distortion of the phase space along
hyperbolic lines. In the picture below the purple circles indicate the coherent input
states out of which the squeezed output states (red) emerge. The phase angle of the
input states in these coordinates is given by the relative phase angle between pump and
signal wave. It can clearly be seen how all kinds of squeezed states can be generated by
altering the phase or the ampitude of the input state.
The first experiments of this kind were performed by Wu and Kimble. The first squeezing
experiments in general by Slusher et al. (see also the book of Walls/Milburn). (refs
6,7,8)
How were the measurements performed
The crucial technique that has been employed for the measurements shown above, is the
technique of
homodyne detection.
Here a weak signal beam (the squeezed or coherent light wave) is
overlapped with a strong coherent local oscillator of the same optical frequency at a
50/50
beam splitter.
The fields emerging from the beam splitter are the sum and the
difference of the two input waves. Subtracting the
photocurrents
of two detectors at the
beam splitter output ports yields a current directly proportional to the signal wave (ref.
10). By this method the noise at the high frequency of the signal waves natural oscillation
is mixed down to an electronically accessible RF range. The following picture depicts this
detection scheme:
Part of the actual setup in the lab
The author in a desperate attempt to produce thermal light
