Transcript University of Birmingham LISA Pathfinder Phasemeter PDR

Homodyne readout of an interferometer
with Signal Recycling
Stefan Hild
for the GEO600 team
October 2007
LSC-Virgo meeting Hannover
Motivation for DC-readout (1)
Increased coupling of laser power noise.
Usually an output mode cleaner (OMC) is required.
Very sensitive to imbalances of the interferometer arms.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 2
Motivation for DC-readout (2)
Reduced shot noise (no contributing terms from 2 times the heterodyne
frequency)
Reduction of oscillator phase noise and oscillator amplitude noise
Stronger low pass filtering of local oscillator (due to PR cavity pole)
Simplify the GW detector




Simpler calibration (GW-signal in a single data-stream, even for detuned SR)
Simpler circuits for photodiodes and readout electronics
Possibility to use photodiodes with larger area => reduced coupling of pointing
Reduced number of beating light fields at the output photodiode => simpler
couplings of technical noise
Requires less effort for injecting squeezed light (=> useful precursor for
GEO-HF)
LO and GW pass the same optical system (identical delay, filtering, spatial
profile) => This advantage is especially important for detectors with arm
cavities.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 3
DC-readout without OMC
nfft/fs = 2.00 : navs = 60 : enbw = 0.75 : nsecs = 120
reference
S5 550 Hz
Heterodyne
-7
Red.
modulation
10%
MIMI
midx
Red.
modulation
10%
MIMI
midx
+ TEM00 carrier
+ carrier from dfo
-8
10
(Watt / Hz1/2)
[W/sqrt(Hz)}
Darkport
Opticalpower
Power @ darkport
10
-9
10
-10
10
2
10
Frequency
[Hz]
Frequency
(Hz)
10
3
Turning down the radio frequency modulation (stable operation is
possible with 10 times lower sidebands)
Dark port is dominated by carrier light (TEM00) from a 50 pm dark
fringe offset
Disadvantage: Still some shot noise contribution from RF-sidebands.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 4
Simulated shot noise: Homodyne vs
Heterodyne detection
Strain [1/sqrt(Hz)]
Homodyne, tuned SR
Heterodyne, tuned SR
10
-21
10
-22
10
2
10
3
Frequency [Hz]
DC-readout with tuned Signal-Recycling:
Stefan Hild
- shape stays constant
- overall level is reduced
LSC-Virgo Meeting 10/2007
Slide 5
Simulated shot noise: Homodyne vs
Heterodyne detection
Strain [1/sqrt(Hz)]
Homodyne, 550Hz
Heterodyne, 550Hz
10
-21
10
-22
10
2
10
3
Frequency [Hz]
DC-readout with detuned SR: - better peak sensitvity
- shape is rotated => better at low freqs, worse at high freqs.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 6
Simulated shot noise: Homodyne vs
Heterodyne detection
Strain [1/sqrt(Hz)]
Homodyne, 1kHz
Heterodyne, 1kHz
10
-21
10
-22
10
2
10
3
Frequency [Hz]
DC-readout with detuned SR: - better peak sensitivity
- shape is rotated => better at low freqs, worse at high freqs.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 7
Strain [1/sqrt(Hz)]
Simulated shot noise: Homodyne vs
Heterodyne detection
10
-21
10
-22
Homodyne, tuned SR
Heterodyne, tuned SR
Homodyne, 550Hz
Heterodyne, 550Hz
Homodyne, 1kHz
Heterodyne, 1kHz
10
2
10
3
Frequency [Hz]
1st Question: Can we confirm the rotation of the shape in our measurements?
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 8
‚Rotation‘ of the optical gain
Optical Gain [a.u.]
Rotated shape of optical response
confirmed by measurement:
10
5
DC-readout
Heterodyne
Phase [deg]
200
10
2
10
3
Frequency [Hz]
100
0
-100
-200
10
2
10
3
Frequency [Hz]
Rotated shape of optical response can be
understood by looking at the phases of
the contributing light fields. => change of
the optical demodulation phase.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 9
Strain [1/sqrt(Hz)]
Simulated shot noise: Homodyne vs
Heterodyne detection
10
-21
10
-22
Homodyne, tuned SR
Heterodyne, tuned SR
Homodyne, 550Hz
Heterodyne, 550Hz
Homodyne, 1kHz
Heterodyne, 1kHz
10
2
10
3
Frequency [Hz]
2nd Question: Can we confirm the change of the relative shape of tuned and
detuned SR with DC-readout ?
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 10
Simulated shot noise: Homodyne vs
Heterodyne detection
Strain [1/sqrt(Hz)]
Homodyne, tuned SR
Homodyne, 1kHz
10
-21
10
-22
10
2
10
3
Frequency [Hz]
2nd Question: Can we confirm the change of the relative shape of tuned and
detuned SR with DC-readout ?
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 11
Strain sensitivity [1/sqrt(Hz)]
Comparison of measured and simulated
optical transfer function for DC-readout
10
-21
10
-22
Simulated shot noise, DC 1000Hz
1/Optical Gain, DC 1000Hz, measured
Simulated shot noise, DC 0Hz
1/Optical Gain, DC 0Hz, measured
10
2
Frequency [Hz]
10
3
The simulated optical transfer function for tuned and detuned SR wit DCreadout is reproduced by our measurements.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 12
Best sensitivity so far with DC-readout
and a SR detuning of 550 Hz
Strain [1/sqrt(Hz)]
10
10
10
10
-19
Simulated shotnoise, DC-readout 550Hz
Simulated shotnoise, heterodyne 550Hz
h(f), heterodyne 550 Hz
h(f), DC-readout 550 Hz
-20
-21
-22
10
2
3
10
Frequency [Hz]
Increased
technical
noise
Stefan Hild
Peak sensitivity roughly
same as with heterodyne
(2e-19m/sqrt(Hz))
LSC-Virgo Meeting 10/2007
Shot noise
 Increased at high
freqs in DC-readout
(with detuned SR)
Slide 13
Noise budget for DC-readout (detuned SR)
• is partly
limiting at low
frequencies
10
• overall seems
to be less of a
problem than
initially expected
ASD [V/sqrt(Hz)]
Laser power
noise (LPN):
10
10
10
10
3rd Question: Do
we understand
the laser power
noise coupling?
Stefan Hild
10
-2
Noise projection for DC-readout with detuned SR
Mid vis
MID AA FB Rot
MID AA FB Tilt
Signal recycling longitudinal noise
Laser amplitude noise (BS ar)
PR error
Dark noise
Modelled shot noise
Sum of the noise
-3
-4
-5
-6
-7
2
10
10
3
Frequency [Hz]
LSC-Virgo Meeting 10/2007
Slide 14
Understanding the LPN in DC-readout
Mag [a.u.]
measured data, SR 550 Hz
Simulation, SR 550 Hz
10
0
10
2
10
3
Frequency [Hz]
Good agreement between measurement and simulation !!
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 15
Summary
Demonstrated DC-readout with tuned and detuned SignalRecycling (without OMC)
Going to DC-readout changes the optical demodulation
phase (rotated shape of optical response)
Measurements and simulations agree pretty well:
 Optical response
 Laser intensity noise coupling
Achieved a displacement sensitivity of 2e-19m/sqrt(Hz)
(currently worse sensitivity than in heterodyne readout)
Laser power noise is not as bad as rumors suggest (due to
filtering of PR cavity pole)
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 16
Where to go in future ??
Best shot noise at low
and high frequencies.
This combination of SR
tuning and DC-readout
would allow an ‚easy‘
implementation of
squeezed light (no filter
cavity necessary to get
full benefit)
See talk by S.Chelkowski @ QND-meeting
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 17
Additional
slides
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 18
Output mode for positive and negative
dfo: observation vs simulation
negative dfo
Stefan Hild
dark fringe
LSC-Virgo Meeting 10/2007
positive dfo
Slide 19
Output mode for positive and negative
dark fringe offset (dfo)
positive dfo
negative dfo
Wave front radii of returning beams
@ beam splitter:
horizontal: north > east
vertical: north < east
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 20
Realisation of tuned signal recycling
For tunings < 250 Hz we cannot achieve a reasonable control signal.
Signal amplitude [a.u.]

0
SR Mod = 9016395 Hz,
tuning = 1000 Hz
SR Mod = 9017065 Hz,
tuning = 330 Hz
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
9
10
Signal amplitude [a.u.]
Position of MSR [nm]
0
-4
SR Mod = 9017065 Hz,
tuning = 330 Hz
SR Mod = 9017195 Hz,
tuning = 200 Hz
SR Mod = 9017395 Hz,
tuning = 0 Hz
-3
-2
-1
0
1
2
3
4
Position of MSR [nm]


Developed a new technique: We ‘kick‘ MSR in a controlled way to jump to
tuned SR, where a reasonable control signal can be obtained again.
MSR is caught at the tuned operating point again.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 21
2 different possibilities for going to
tuned signal recycling
1.
2.
Keep the modulation frequency and jump to center zerocrossing.
Change the modulation frequency (corresponding to 0 Hz tuning)
=> only a single zerocrossing exists.
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 22
Laser intensity noise coupling for tuned
and detuned SR
Mag [a.u.]
10
measured data, SR 550 Hz
Simulation, SR 550 Hz
Simulation, tuned SR
Simulation, tuned SR,
adjusted mod freq
Simulation, tuned SR,
adjusted mod freq,
lower mod index
1
10
0
10
2
10
3
Frequency [Hz]
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 23
Optical Gain []
Tuned DC with various dark fringe
offsets
Tuned DC,
Tuned DC,
Tuned DC,
Tuned DC,
10
data set
data set
data set
data set
3
2b
4
5
data set 2b:
small –dfo
data set 3:
large –dfo
5
2
10
10
3
Frequency [Hz]
150
data set 4:
small +dfo
Phase [deg]
100
data set 5:
large +dfo
50
0
-50
-100
2
10
10
3
Frequency [Hz]
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 24
Comparison of heterodyne 550 Hz,
tuned heterodyne and tuned DC
10
DER_DATA_H (nominal S5= 550 Hz, heterodyne)
2007-05-16 03:20:30 TUNED SR, heterodyne readout (P-quad), (Sept 2006)
TUNED SR, DC-readout (4513)
DC 0Hz (simulated shot noise)
Hetero 0Hz (simulated shot noise)
Hetero 550Hz (simulated shot noise)
-20
STRAIN [1/sqrt(Hz)]
10
-19
10
10
-21
While in the
two heterodyne
cases the
sensitivity is
close to
simulated shot
noise at 2 kHz,
this is not the
case for tuned
DC.
-22
10
2
10
3
Frequency [Hz]
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 25
Combination of tuned SR and
squeezing– an option for GEO HF?
• Squeezed light is available for injection
“Coherent Control of Vacuum Squeezing in the Gravitational-Wave
Detection Band“, Vahlbruch et al, PRL 97, 011101 (2006)
• Tuned Signal-Recycling operation was demonstrated
„Demonstration and comparison of tuned and detuned Signal-Recycling in a large
scale gravitational wave detector“ , S Hild et al, CQG. 24 No 6, 1513-1523.
 No need for long filter cavity !
Stefan Hild
LSC-Virgo Meeting 10/2007
Slide 26