Compact double-pass AOM setup

Double-pass your AOM!
It’s actually pretty cool. When you retro-reflect the diffracted light back into the AOM and diffract it in the same order again (like +1 and +1, or -1 and -1), a few nice things happen:

1. Tuning range:
   There will be an effectively larger tuning range of the radio frequency drive. Because a df change of the drive frequency translates to a 2.df frequency change on the light.

2. Pointing:
   If the AOM is placed in the focal plane of a lens followed by the retro-reflector, the pointing of the twice-diffracted light stays fixed when the drive frequency is changed.

3. Thermal sensitivity:
   Because of feature number 2, the typical temperature sensitivity in AOMs that leads to beam pointing instability can be largely mitigated.

I recently wanted to divide my laser beam into 5 branches and set up a double-pass AOM for each branch. They are then used for Raman sideband cooling in a Ytterbium tweezer experiment. Because I wanted to push the setup for compactness, I ended up using a polarizing beamsplitter cube glued on a mirror mount together with the AOM on a mirror mount. Combined, they provide 4 degrees of freedom to align the beam position into the AOM crystal and hit the Bragg angle of the crystal. Here’s the schematic of this arrangement:

And a top view of the experimental setup:

It gives me satisfaction to see all 4 ports of each cube being used!

Bonus (!):
If the retro-reflector is exactly f away from the lens, i.e. a cat’s eye arrangement, a collimated beam entering the AOM leaves collimated after twice diffraction. Now, if the retro-reflector is moved back and forth, the beam’s divergence can be tuned. This means that if we are focusing the double-diffracted light using a single lens, the focal point can be moved along the optical axis. In a recent work from Manuel Endres team, authors propose using a grating as the retro-reflector which enables using the acoustic diffraction frequency to move the beam’s focus position along the optical axis.