Transcript C. Mow-Lowry (ANU)

Quantum Noise Measurements at
the ANU
Sheon Chua, Michael Stefszky,
Conor Mow-Lowry, Sheila Dwyer,
Ben Buchler, Ping Koy Lam, Daniel
Shaddock, and David McClelland
Centre for Gravitational Physics
Australian National University
Homodyne Detection
• Homodyne detectors work by comparing a weak signal
beam with a strong local oscillator
• The two beams are interfered on a beamsplitter and
detected on two photodiodes
• The subtraction of the diodes can give either the amplitude
or the phase projection of the noise on the signal beam
• The subtraction gives
enormous common mode
rejection
• Uncorrelated technical
noise masks the signal.
Homodyne Detection
Scatter
• Small angle scatter which propagates in the (0,0) mode
interferometrically couples in phase fluctuations from
mirror motion and air currents
• Depending on the location of the principal scattering
sources, this can create uncorrelated intensity noise.
Scatter
• By sweeping the phase of a parasitic interferometer with a
PZT, the phase noise can be moved out of band.
• This technique can be used to diagnose the presence of
scattered light, and to shift it out of the measurement band.1
1
de Vine et. al., Phys. Rev. Lett., Accepted for publication (2010)
Scatter
• A PZT was used to modulate the path length at two separate
points of the apparatus at a variety of modulation
frequencies and amplitudes.
• In an effort to increase the effect, a scatter source was
introduced.
• In all cases, there was no evidence that a parasitic
interferometer was present, neither in reduction of low
frequency noise nor in the broadening of the modulation
peak.
Dust
• Dust moving through the beam after the beamsplitter
causes non-stationary uncorrelated intensity fluctuations 1
• For the figure below, each diode had an equivalent of 6
Volts incident, with measured subtraction to 1 part in 1000
• The largest dust excursions result in worse than 1 part in
100 subtraction
1
Chua et al., J. Phys.: Conf. Ser. 122 012023 (2008)
Pointing
• Experiments by McKenzie et al.1 demonstrated
coupling of pointing to homodyne readout
• Confirmed in our apparatus by driving PZTs
• Pointing noise generates uncorrelated noise on the
two diodes due to detector inhomogeneities.
• Even after sealing the chamber, the homodyne
readout was very susceptible to anthropogenic
noise.
1
McKenzie et al. Applied Optics 46 3389 (2007)
Pointing
• After the homodyne
chamber was sealed,
noise slowly improved
with time
• Monday morning
anthropogenic noise
caused further large
disturbances, exciting
the spectrum (not
shown)
• No modecleaner
installed, using AEI
detector.
Modecleaner
• One of the key improvements was placing a small,
moderate finesse (~300) modecleaner inside our
chamber.
• The modecleaner converts uncorrelated pointing
noise and mode shape disturbances into common
intensity noise
• This truly common noise is rejected by more than
60 dB, finally rendering the homodyne output
resistant to anthropogenic noise
Electronic Noise
• We investigated two couplings of electronic noise:
– Additive dark noise, and
– Non-linear electronic noise
• One potential mechanism for non-linear noise is
uncorrelated ‘gain noise’ which couples due to the
large dynamic range required to see shot noise.
Non-linear electronic Noise
Low-pass filtered
DC voltage with
huge (~80 dB)
common mode
rejection showed
voltage dependent
noise
Current Subtraction
• It is possible to avoid gain noise by directly
subtracting the diode photocurrent.
• Both homodyne diodes are placed on the same
circuit-board and subtracted before the
transimpedance amplifier1:
1
Designed by the squeezing team at AEI Hannover
Shot Noise (I)
Shot Noise (II)
Conclusions
• Isolation from the general lab environment was
required to prevent dust and air current
disturbances
• Scatter and stray light did not cause an issue
despite stock optics and imperfect cleanliness
• Beam jitter was a strong source of noise mitigated
by the introduction of a modecleaner inside a
common chamber
• Non-linear electronic noise was limiting
performance in prior experiments, but is no longer
an issue when using a current-subtraction detector.
Squeezing
Proof of concept
experiments have shown
sensitivity improvements
(ANU, MIT, AEI)
GEO is also investigating
the introduction of
squeezed states currently
Squeezed Hanford 4km Project
Squeezing to be injected
into Hanford 4km detector
asymmetric port Faraday
Isolator
Investigation into:
– The Impact of the
squeezer on LIGO
operation
– injection losses
– The effect of scattered
light from LIGO on
the OPO
– The effect on LIGO
sensitivity (!)
Coherent control of vacuum squeezing in the gravitational-wave detection band
Vahlbruch et al. Phys. Rev. Lett. 97, 011101 (2006)
The LIGO Injection Test OPO
PZT Actuator
Squeezing Out
Pump light In
Crystal
Oven/
Temperature
Sensor
Squeezing June 2009
Improvements
 New OPO constructed (Mk II) including new crystal
 Further optimised locking loops
 New homodyne detector installed (courtesy H. Vahlbruch, AEI)
 Chamber used to isolate homodyne detector and modecleaner
 Mitigation of scattered/ stray light with dichroics, dumps, and
cleaning of optics
Improvements
Homodyne isolation
chamber:
Squeezing January 2010
Future Directions
 Installation of new, high quality optics, including new crystals
 ANU OPO delivered to MIT, awaiting installation and testing
 Investigation of long term squeezing stability
 Delivery of complete squeezing table from MIT to Hanford
 Injection of squeezing into an operational gravitational wave
detector.