Transcript Post S5 Improvements - LIGO

The Next
Gravity Wave
Interferometers
Rana Adhikari
Caltech
Gravitational Waves?
 Gravitational Waves = “Ripples in space-time”
 Two transverse polarizations - quadrupolar:
+ and x
Example:
Ring of test masses
responding to wave
propagating along z
Amplitude parameterized by
dimensionless strain h: DL ~ h(t) x L
Need to measure strain of ~ 10-21-10-22
We want a very large ‘L’
2
GW Sources in LIGO Band 50-1000 Hz
“chirp”
 Compact binary inspirals:
 NS-NS waveforms are well described.
inspiral is a standard candle.
 BH-BH waveforms are rapidly improving
 Supernovae / Mergers:
1.4 Msolar NS/NS
“burst”
 Short signals. Waveforms not well known.
 Search in coincidence between two or more interferometers and
possibly with electromagnetic and/or neutrinos signals
 Spinning NS:
“continuous”
 search for signals from observed pulsars
 all-sky search computing challenging
 Cosmic Background:
“stochastic”
 Metric fluctuations amplified by inflation, phase transitions in
early universe, topological defects
 Unresolved foreground sources
3
LIGO Observatories
Hanford, WA (H1 4km, H2 2km)
- Interferometers are aligned to be as
close to parallel to each other as
possible
- Observing signals in coincidence
increases the detection confidence
- Determine source location on the sky,
propagation speed and polarization of
the gravity wave
Livingston, LA (L1 4km)
LIGO
GEO
Virgo
TAMA
AIGO (proposed)
4
Michelson Interferometer
 = 2p (dLy - dLx) / l
dP/d  PBS x sin()cos()
Reflected Port
d/dh  L
Ly
Lx
Laser
Anti-Symmetric
(Dark) Port
PAS  PBS x sin2()
Noise  PAS
5
Interferometer Optical Layout
End Test
Mass
Power Recycled
Michelson
Interferometer
with Fabry-Perot
Arm Cavities
4 km Fabry-Perot
arm cavity
Power Recycling
Mirror
300 W
Laser
6W
Photo
detector
Input Test
Mass
20 kW
50/50 Beam Splitter
Signal  Phase shift between
the arms due to GW
6
Seismic:
Natural and anthropogenic
ground motions, filtered by
active/passive isolation
systems.
Depends strongly on in-vac
seismic isolation.
Science Requirements Doc:
The LIGO-I Sensitivity Goal
Thermal:
Brownian noise in the mirrors
and in the mirrors’ steel
suspension wires.
Depends mostly on internal
rubbing in the suspension
wires.
Shot Noise:
Photon counting statistics -> 10 kW in the cavities
~ 200 mW detected power
- Goes down with increased
laser power and better
fringe contrast
7
5 years of debuggin’ in Louisiana…
8
9
What Is Inside
1.2 m diameter - 3mm stainless 50 km of weld
10-9 torr vacuum and no leaks!
Seismic isolation
Stack of masses
and springs
Coils and magnets to control the mirror
Fused silica mirror
25 cm diameter
10 kg mass
10
Dark Port Optical Table
2 mm diameter
InGaAs photodiode
11
Time Line
1999
2000
2001
2002
2003
2004
2005
2006
3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
Inauguration First Lock Full Lock all IFO
4K strain noise
Science
10-17 10-18
HEPI at LLO
10-20 10-21
S1
S2
at 150 Hz [Hz-1/2]
10-22
S3
Now
S4
S5
Runs
First
Science
Data
2006
12
LIGO Scientific Collaboration
~40 institutions, ~550 scientists
Caltech
LIGO Laboratory
LIGO Hanford Observatory
University of Adelaide ACIGA
Australian National University ACIGA
Balearic Islands University
Caltech LIGO
Caltech Experimental Gravitation CEGG
Caltech Theory CART
University of Cardiff GEO
Carleton College
Cornell University
Embry-Riddle Aeronautical University
University of Florida-Gainesville
Glasgow University GEO
NASA-Goddard Spaceflight Center
Hobart – Williams University
India-IUCAA
IAP Nizhny Novgorod
IUCCA India
Iowa State University
MIT
LIGO Livingston Observatory
Loyola New Orleans
Louisiana State University
Louisiana Tech University
MIT LIGO
Max Planck (Honnover) GEO
Max Planck (Potsdam) GEO
University of Michigan
Moscow State University
NAOJ - TAMA
Northwestern University
University of Oregon
Pennsylvania State University
Southeastern Louisiana University
Southern University
Stanford University
Syracuse University
University of Texas-Brownsville
Washington State University-Pullman
University of Western Australia ACIGA
University of Wisconsin-Milwaukee
13
LIGO Science Run
 The fifth science run started in November 2005
 S5 goal is to collect one year of triple coincidence data at the
design sensitivity
 Optimistic event rates: NS/NS ~3/year, BH/NS ~30/year Nakar,
Gal-Yam, Fox, astro-ph/0511254
 Plan to reach the Crab pulsar spin down limit
 Expect to beat the Big-Bang Nucleosynthesis limit on
gravitational wave density in the LIGO band
 GEO interferometer joined the S5 run in January 2006.
 Virgo interferometer plans to join S5 later this year.
14
NS-NS Inspiral Range
Improvement
Time progression since the start of S5
Design Goal
Commissioning
breaks
Histogram
Stuck ITMY optic
at LLO
15
S5 Duty Factor
One week running average
S5 Goal is 85%
for single
interferometer
and 70% for
triple
coincidence
Commissioning
breaks
Stuck ITMY optic
at LLO
16
S5 Duty Factor
H1
H2
L1
Uptime
72%
79%
60%
Wind, Storms, Earthquakes
4.5%
Nearby Logging, Construction, Trains
-
Maintenance, Commissioning,
Calibration
10%
9%
Hardware and Software Failures
3.5%
7%
Lock Acquisition, Other
10%
5%
H1&H2&L1 = 45%
9%
-
10%
H1||H2||L1||G1 close to 100%
17
Triple Coincidence Accumulation
~ 61%
100%
~ 45%
Expect to collect one year of triple
coincidence data by summer-fall 2007
18
Sometimes You Get Lucky
 Large mirror (ITMY) was wedged into the earth quake stops
 Vented the vacuum and released it. Adjusted EQ stop.
 Noise improved!? 12->14 Mpc
Earth quake stop
Charge
Dissipation
on the optic?
19
Advanced LIGO
 LIGO mission: detect
gravitational waves and
initiate GW astronomy
 Next detector
 Should have assured detectability
of known sources
 Should be at the limits of
reasonable extrapolations of
detector physics and technologies
 Must be a realizable, practical,
reliable instrument
 Daily gravitational wave detections
 R&D is mature, prototypes exist
 Installation start in 2011
Advanced LIGO
20
Anatomy of the projected
Adv LIGO detector performance
 Seismic ‘cutoff’ at 10 Hz
 Suspension thermal
noise
 Test mass thermal
noise
Initial LIGO
10
10-22
-22
Strain Noise, h(f) /Hz1/2
 Newtonian background,
estimate for LIGO sites
10-21
-23
Advanced LIGO
10
10-23
NS-NS Tuning
 Unified quantum noise
dominates at
most frequencies for full
10
10-24
power, broadband
tuning
-24
1
10
10 Hz
2
10
Frequency (Hz)
100 Hz
3
10
1 kHz
21
Optical Configuration
Dual
Powerrecycled
Recycled
Michelson
Interferometer
with Fabry-Perot
Arm Cavities
End Test Mass
4 km Fabry-Perot
arm cavity
Power Recycling
mirror
2 kW
Input Test
Mass
500 kW
Laser
125 W
50/50 beam splitter
GW signal
Signal Recycling
mirror
22
Detuned Signal Recycling
•IFO Differential Arm mode is
detuned from resonance at
operating point
IFO DARM/CARM
500
200
d 0
100
SRC
DARM
fsig
50
LSB
-10000
d 0
d 0
FWHM
Carrier frequency
Sideband amplitude [a.u.]
1000
-5000
from R. Ward
USB
0
5000
10000
frequency offset from carrier [Hz]
•Responses of GW USB and GW LSB are
different due to the detuning of the signal
recycling cavity.
23
Opto-mechanical Spring
Radiation pressure:
F=2P/c
Detuned Cavity ->
dF/dx
• ½ MW in the arms ->
Measured Transfer Functions from the 40m
prototype
• ‘Optical Bar’ detector
• ~75 Hz unstable optomechanical resonance
• High Bandwidth servos
Optical Spring stiffness ~ 107 N/m
BMW Z4 ~ 104 N/m
Angular spring resonance ~ 2 Hz 24
The next several years
Other interferometers in operation (GEO, Virgo)
NOW
4Q
‘06
4Q
‘05
4Q
‘07
4Q
‘08
4 yrs
4Q
‘09
4Q
‘10
Adv
S5

~2 years
S6
LIGO
Between now and AdvLIGO, there is some time to
improve…
1) ~Few years of hardware improvements +
1 ½ year of observations.
1) Factor of ~2.5 in noise, factor of ~10 in event rate.
2) 3-6 interferometers running in coincidence !
25
NS/NS Binary
Area proportional to SNR
Most of the sensitivity comes
from a band around 50 Hz
50 Hz
26
30/30 M BH/BH
Area proportional to SNR
Most of the sensitivity comes
from a band around 30 Hz
30 Hz
27
Astrophysical Motivation
How does the number of surveyed galaxies increase as the
sensitivity is improved?
From astro-ph/0402091, Nutzman et al.
For NS-NS binaries
Power law: 2.7
Factor of 2.5 reduction in
strain noise,
factor of 10 increase in
# of sources
S4
Prop. to inspiral range
28
Baseline Goals
1. Dark Port Filter Cavity (Caltech/MIT)
1. Reduce amount of junk light, reduce shot noise
2. Reducing detected light power allows higher laser power
3. Upgrade the detection system to the Advanced LIGO style.
2. Higher power laser (Hannover)
1. Our (10 W) laser company was bought out by JDS Uniphase.
2. Collaborators at AEI/LZH are offering us 35 W lasers (for free!)
3. High Power Input Optics (UF, Gainesville)
4. Miscellaneous …
29
Seismic:
No modification in seismic
isolation systems
Thermal:
Shot Noise:
Good wires, good mirrors, and
control of “technical” noises
- New in-vac filter cavity
-- 4-5x more laser power
-- Advanced readout
30
Lower Thermal
Noise Estimate
Increased Power +
Enhanced Readout
31
The Plan
 Improvements on the 4km IFOs starting in Sep07
 Do Louisiana first (pathfinder). Start Hanford in Jan08.
 Some modest Suspension electronics fixes
 Then some more science running.
Not enough time/manpower to do all 3 IFOs.
A factor of 2.5 on H1/L1 is better than a factor of 2 on all three.
We don’t gain more AdvLIGO knowledge by doing 3 IFOs.
After the pumpdown, H2 can join Virgo in a science run.
32
Conclusions
 1 more year of R&D to support the upgrade
 Starting installation in Sep ’07
 Next Science Run (w/ improved sensitivity) starts
in Sep ’09.
 Reduces much technical risk for Advanced LIGO
 Its time to make the first gravitational wave
detection.
33
Short, Hard GRBs
 Several Short Hard Gamma-Ray Bursts
since last year
 Detected by Swift, HETE-2, then Hubble,
Chandra, and BATSE
 SHGRBs (< 2 s) different from ‘Long Soft
GRBs’ (supernova explosions)
 Candidates for progenitors of SHGRBs:
double neutron star (NS/NS) or neutron
star-black hole (NS/BH 1.4/10)
coalescences
 Optimistic rates for Initial LIGO (S5)
could be as high as a few/year
Nakar, Gal-Yam, Fox, astro-ph/0511254
Double Neutron Star Merger
Dana Berry / NASA
34
Beyond ‘Fixes’
 Suspension wire re-working




Change the clamp to reduce excess noise (no evidence so far)
Change the wire to reduce the intrinsic noise
Needs some serious coil driver redesigns capitalize on lower noise.
Need to know more about the excess noise first.
 Squeezed Light
 Implement on one IFO instead of the laser upgrade; more
speculative, but doesn’t require new IO equipment.
 An opportunity to commission another AdvLIGO system
 Signal Recycling
 No real sensitivity improvement; lots of work.
 Double Suspension
 Not directly applicable to AdvLIGO. Substantial reworking req.
 Not clear if we can get the technical noises out of the way.
35
Detection Strategy, point source
 Cross-correlation estimator
Point Spread Function
 Theoretical variance
 Optimal Filter
1  point(t , f ) H ( f )
Q (t , f )  N P ( f ) P ( f )
1
2
~
 point(t , f ) 
e

Dx ( t )
i 2pf 
c
A  ,
^
F1,At  F2A,t  

H
(
f
)

H
(
f
/
100
Hz
)
Strain Power:

Choose N such that:
 Y  H 
36
S4 Upper Limit map , H(f)=const
H 90%  (0.85  6.1) 10 48 Hz 1
37