1. Ponderomotive Squeezing Quantum Measurement Group
Thomas Corbitt, Keisuke Goda, Nergis Mavalvala, Euginy Mikhalov, David Ottaway
Yanbei Chen(Caltech), Stan Whitcomb(Caltech)

2. Introduction
Crystal based squeezing is the only current possibility for generating squeezed light, and it’s not known whether it will be able to reach frequencies ~ 100 Hz.
We propose using a tabletop interferometer to generate squeezed light, using radiation pressure as the squeezing mechanism. We show that the expected noise sources in our configuration do not prohibit squeezing.

3. Design PD - PD To GW detector. Spectrum analyzer 1 W ~10 kW
Laser
1 gram test mass
0.25 kg test masses
1 W
~10 kW
Detuned arm cavities (10 cm)

4. Frequency independent squeeze angle
Ideally, the squeezer should produce the frequency dependent squeeze angle that’s required by the GW detector.
It is not known how to do this directly, instead -
First produce frequency independent squeezing
Filter this squeezing to produce the desired frequency dependence.

5. Quantum Noise a Amplitude  b1 = a1 Phase  b2 = -k a1 + a2 + h b
Radiation Pressure
Shot Noise

6. Producing a Flat Squeeze Angle
Frequency dependent squeeze angle is due to the frequency response (f-2) of a free mass to a force.
Modify the dynamics of the test mass by attaching it to a spring with a high resonant frequency – below the resonant frequency of the spring, the response is frequency independent.

7. Thermal Noise in Springs
Why not use a mechanical spring?
The thermal noise introduced by the high frequency (mechanical) spring will wash out the effects of squeezing.
An optical spring with a high resonant frequency will not change the thermal force spectrum of the mechanical pendulum.
A low resonant frequency mechanical pendulum may be used to minimize thermal noise, while the optical spring produces the flat response up to high frequencies.

8. Optical Springs
If the detuning increases (cavity becomes longer) then the power in the cavity decreases, and the mirror is pushed back to the detuned point. If the detuning decreases, then the
power in the cavity increases and the mirror is pushed back.
Detuned Point L
Detuning = ωL/c

9. The Arm Cavities High finesse, high power (possibly very difficult)
0.5 W input builds up to ~10kW.
Light (1 gram) test mass allows for large radiation pressure effects and stiff spring.
~10 cm long
Performed in vacuum with seismic isolation
Suspended test masses
Detuned to create optical spring

10. The Arm Cavities
The shared end mass case has worse laser noise couplings, but there may be some difficulty in controlling the independent mass case. The static radiation pressure force on the shared mass cancels each other, while it pushes the independent test masses about 1mm(!).

11. Laser Noise
Mismatches in the arm cavities (losses, detuning and finesse), and an imperfect beamsplitter at the 1% level are assumed.
Relative intensity noise at the level of 10-8 / rt Hz, and frequency noise of 10-4 Hz/rt Hz are assumed. These are what’s currently achievable.

12. Laser Noise

13. Laser Noise Optimization(squeezing)
Michelson on dark fringe
Michelson detuning

14. The Test Mass ~1 cm diameter, 3mm thick Require low thermal noise.
Use fused silica as the material and bond fused silica fibers to use as the suspension.
Coating thermal noise could be an issue.
Very high optical quality – losses ~ 5ppm per bounce.

15. Noise Sources

16. Why is this interesting?
Low frequency alternative to crystal squeezing.
Test quantum limited radiation pressure effects – gain confidence that the modeling is correct.
Test noise cancellations via Michelson detuning.
Squeezing may be produced while having a sensitivity far worse than the SQL due to the optical spring.

17. Current State
Finalizing design.
Building will start soon.