Transcript EXPLORER (CERN) - LAPP

The status of VIRGO
E. Tournefier
LAPP(Annecy)-IN2P3-CNRS
Journées SF2A, Strasbourg
27 juin – 1er juillet 2005
•
The VIRGO experiment
•
The commissioning of VIRGO
•
Towards a global network
•
Conclusions
1
How to detect gravitational waves?
•
Effect of a gravitational wave on free masses:
 A Michelson interferometer is suitable:
- suspend mirror with pendulum => ‘free falling masses’
- Gravitational wave => phase shift
 4 hL

Suspended
mirror
L  
Suspended
mirror
hL
2
L  
h = L/L
L = length difference between the 2 arms
L = arm length
Beam splitter
LASER ()

Light
Detection
2
hL
2
The shot noise and the VIRGO optical design
Limitation of a Michelson interferometer due to photon shot noise:
the minimum measurable relative displacement is ~  1 2
h
4 L
P
=> Can reach h ~ 3.10-23 with L=100km and P=1kW
How to achieve that?
1/ Fabry-Perot cavities to increase the effective length:
2
( F = finesse )
L'   F  L

=> L’ = 100km for L=3km and F=50
2/ Recycling mirror to increase the effective power:
P’ = R P
(R = recycling gain)
=> P’ = 1kW with P=20W and R=50
3
Gravitational wave signal
Noise sources in interferometers
Thermal
Noise
Acoustic
Noise
Seismic
Noise
Index fluctuation
Shot
Noise
Laser Noises
Detection
Noise
4
Noise sources: seismic noise
Seismic noise spectrum for f few Hz:
Seismic noise measured on the Virgo site
a m
~
xs  2
f
Hz
a ~ 10-6 - 10-7
m/Hz
  shot noise !
 Need a very large attenuation!
Solution:
suspend the mirrors to a chain of pendulums
Transfert function
10-12
100 Hz
With 6 pendulums: attenuation of the seismic noise
by more than 10 orders of magnitude above 4 Hz!
5
Suspensions and control of the interferometer
All mirrors are suspended to a cascade of 6 pendulums:
 Large attenuation in the detection band ( > 10 Hz)
 Large residual motion at low frequencies: < ~1mm
 Need active controls to:
- maintain the interferometer’s alignment
- maintain the required interference conditions
1/ Local control of the suspensions:
 Residual motion ~2 m/sec
 Obtain interference fringes
2/ To keep the interferometer at interference conditions:
– Need to control the cavity length to 10-12 m
 Use the interferometer signals (photodiodes)
6
VIRGO design sensitivity
Main sources of noise limiting the VIRGO design sensitivity
Seismic noise
Thermal noise
Shot noise
Shot noise
1
•
7
Gravitationnal waves sources and VIRGO design sensitivity
(sources: see talk by N. Leroy)
Distance to the Virgo cluster = 10Mpc
8
VIRGO
French-italian collaboration (CNRS – INFN)
Site : Cascina close to Pisa
5 french labs: Annecy (LAPP), Lyon (LMA), Nice (OCA), Paris (ESPCI), Orsay (LAL)
6 italian labs: Firenze, Frascati, Napoli, Perugia, Pisa, Roma (all INFN)
9
The commissioning of VIRGO
Started in summer 2003
•
The steps of the VIRGO commissioning:
- control of the north FP cavity: Oct 2003
- control of the west FP cavity: Dec 2003
- recombined (Michelson) ITF: Feb 2004
- recycled (full VIRGO) ITF: Oct 2004
• Technical runs (3 to 5 days) at each step
C1(Nov 2003),…, C5(Dec 2004)
Lock stability
Sensitivity/noise studies
Data taking on ‘long’ period
West arm
•
Fabry-Perot cavities
input mode cleaner
l=150m
beam
splitter
laser
recycling
mirror
l=6m
output mode cleaner
Gravitational wave signal
North arm
L=3km
10
Recombined interferometer
•
Recombined interferometer:
keep the two Fabry-Perot cavities on resonance + the Michelson on the dark fringe
Example of lock acquisition
Power ‘stored’ inside the FP cavities
Power at the interferometer output
Lock on the dark fringe
11
The lock of the full VIRGO
•
Lock of the recycled interferometer (full VIRGO):
– Need to control 4 degrees of freedom (3 cavities + Michelson)
– The lock is acquired in several steps:
• Start without recycling
+
• Slowly increase the recycling gain
 2 technical runs:
- C5 (3 days, Dec 2004)
- C6 (2 weeks) planned for this summer
POWER IN THE RECYCLING CAVITY
Lock acquisition sequence
With power recycling
Laser
Recycling gain ~ 30
-
Without power recycling
12
Noise studies
Sensitivity measured during C4 run and identified sources of noise
Attention a l’unite!
Noise hunting
1/ Identify the sources of noise which limit the sensitivity
2/ Perform the necessary improvements / implement new controls
=> see talk by R. Gouaty
13
Typical unforseen difficulties
•
Injection bench:
– A small fraction (bigger than expected) of the light reflected by the
interferometer is retro-diffused by the input mode cleaner mirror
 spurious interferences
Temporary solutions:
- rotate the mode cleaner mirror
- reduce the incident light (/10)
 We are now working with only Pin = 0.7 Watts
Final solution: install a Faraday isolator
 A new input bench will be installed in september
14
Sensitivity summary
Single arm, P=7 W
Recombined, P=7 W
Recycled, P=0.7 W
P = 10W
h ~3. 10-21/Hz
The VIRGO sensitivity will significantly improve with:
- the implementation of the automatic alignment of the mirrors (low frequency)
15
- the full power (high frequency)
Data analysis
•
Calibration and reconstruction of the signal: Watts -> meters
- Apply a known displacement to the mirrors
•
A lot of tests on simulated data including interferometer noise
•
Test of the data analysis on real data
from the technical runs:
– Test the full chain of data analysis
– Learn how to put vetoes
– Inject events in the real data:
software and hardware injections
-> measure efficiencies, false alarm rate,…
•
- Injected events
Start collaboration with LIGO:
Coincident analysis will help the detection of gravitational waves
=>decrease false alarm rate (rare events in non gaussian noise)
Combined data analysis is necessary to extract the source parameters
16
Towards a worldwide network
•
•
Look for events in coincidence
Combined analysis is needed to extract information on the source
LIGO
GEO
VIRGO
TAMA
AIGO
17
Status of LIGO
•
Two sites:
– Hanford (Washington):
4km and 2 km interferometer
– Livingston (Louisiana):
4km interferometer
•
•
•
•
Same optical configuration as Virgo
Less sophisticated suspensions
The commissioning started in 1999
The three interferometers are
operational
•
Long science runs have started:
– S1 (Aug 2002)
– S2 (March-April 2003)
– S3 (Nov-Dec 2003)
– S4 (Fev-March 2005)
– 6 month run this year
•
The LIGO sensitivity is now very close
to the design sensitivity !
LIGO Hanford
Observatory
(LHO)
reaching
the Virgo cluster
!
H1 : 4 km arms
H2 : 2 km arms
LIGO Livingston Observatory
(LLO)
L1 : 4 km arms
18
Conclusion
•
VIRGO commissioning is progressing
– The recycled (full VIRGO) interferometer is working
– Sensitivity will make big progress with
• Automatic alignment of the mirrors
• New input bench
– First scientific run in 2006?
•
LIGO is very close to its design sensitivity
– Long science runs will start this year
•
The detection with the first generation of detectors is not guaranteed
– A global network is needed
– A second generation of detectors is being prepared to reach h~few 10-24 /Hz
=> Upgraded VIRGO and LIGO ~ 2010-2013
19
Future: how to improve the sensitivity?
The first generation of detectors might not be able to see gravitational waves
 Need to push the sensitivity further down:
•
Seismic noise:
– The VIRGO suspensions already meet the requirements for next
generation interferometers
•
The main limit: thermal noise
– Monolitic suspensions (silica)
– Better mirrors (material,
geometry, coating)
•
Shot noise
– More powerful lasers
– Signal recycling technique
•
And the technical noises
– Better sensors
– Better electronics
– Better control systems
Shot noise
20
Future: How to go to lower frequencies
•
Frequency range limited on the earth due to seismic noise
=> go to the space: the LISA project
LISA:
•Spatial interferometer (NASA-ESA)
• 3 satellites, size = 5.106 km
• start: 201?
• Much lower frequencies: 10-4 – 1 Hz
• It is complementary to terrestrial
detectors
21
GEO (UK, Germany)
•
•
600m long arms
An interferometer for the development of new techniques:
– Signal recycling
– Monolitic suspensions (-> reduce thermal noise)
600m arm (no FP)
Power recycling
Laser
Signal recycling
22
TAMA (Japon)
•
•
•
•
•
Located at Tokyo
Same optical configuration as VIRGO
Started the commissioning in 1997
Reached first a sensitivity of h ~ 3.10-21 Hz –1/2
But limited by the small arm length
design
10-21
23
Gravitational wave detectors: resonant bars
•
The gravitational wave excites the resonant mode of the bar
 Good sensitivity for frequency = mode of the bar
•
•
First detectors in 1960
Many improvements since then:
– Cryogenic
– New transducers for the
detection of bar oscillation
•
Several detectors in operation
=> perform coincidence data analysis
24
Bars events
EXPLORER (CERN)
NAUTILUS (Italie)
•
Coincident analysis between Explorer
and Nautilus
– 2001 data
– Small excess of events when the
detectors are optimally oriented
with respect to the galactic plane
– Excess not confirmed by recent
data taking
25
The mirrors
•
Fused silica mirrors
•
Coated in a class 1 clean room at SMA-Lyon
(unique in the world).
– Low scattering and absorption:
< few ppm
– Good uniformity on large dimension:
< 10-3  400 mm
• Large mirrors (FP cavities):
–  35 cm, 10 cm thick
– 20 kg
26
The injection and detection systems
•
•
•
Laser: powerful and stable
- 20W
- Power stability: 10-8
- Frequency stability: Hz
The input and output mode cleaners:
- optical filter => improve signal to
noise ratio
Signal detection:
- InGaAs photodiodes, high efficiency
27
The commissioning of the CITF
•
Commissioning of the central interferometer: 09/2001 -> 07/2002
– CITF = Recycled Michelson interferometer (no Fabry-Perot cavities)
- a lot of common points with VIRGO
unit = meters!
•
•
The evolution: configuration and
sensitivity: 4 runs of 3 days each
- E0/E1: Michelson
- E2: Recycled Michelson
- E3: + automatic angular alignment
- E4: + final injection system
Results:
– Viability of the controls
– Sensitivity curve understood
– And gain experience for the
VIRGO commissioning
- Improvements triggered by the CITF experience
28
Data analysis
•
Supernovae
– The signal shape is not well known
 several techniques are developed to detect bursts
 Problem of non gaussian detector noise
 Detection in coincidence with other detectors is needed
•
Binary coalescences
– Well known signal
 Use a matched filtering technique
 The parameters of the sources can be extracted
•
Pulsars
– Need to integrate on long periods
– But the signal is distorted by Doppler effect due to the earth’s rotation
 Huge parameter space
 Limited by computational resources
29
Fabry-Perot cavities
input mode cleaner
l=150m
beam
splitter
Control of the
laser frequency
laser
recycling
mirror
l=6m
L=3km
output mode cleaner
30
Control of the cavity