Progress and Preliminary Results of TAMA Data Analysis

Progress and Preliminary Results of

TAMA Data Analysis8th GWDAW,

Milwaukee WI, USA, 16th Dec. 2003

Nobuyuki KandaDepartment of PhysicsOsaka City University

and

the TAMA collaboration

2

Outline1. Detector status

Search for GW :2. Burst GW3. Inspiral Gravitational Wave4. Black-hole QNM ringdown GW5. Continuous GW from SN1987 remnant

Data Qualify :6. Online veto study

Cooperation :7. LIGO-TAMA coincidence analysis

8. Remarks

3

Detector status (briefly)

4

Detector Statusthe TAMA collaboration

National Astronomical Observatory (NAOJ), Institute for Cosmic Ray Research (ICRR), The University of Tokyo,

High Energy Accelerator Research Organization (KEK), University of Electro-Communications, Osaka City University, Osaka University,

Yukawa Institute for Theoretical Physics, Kyoto University, Niigata University, Hirosaki University, Tohoku University, Hiroshima University,

Tokyo Denki University, National Institute of Advanced Industrial Science and Technology, Tokai University

5

Latest Sensitivity

10-21

10-20

10-19

10-18

10-17

10-16

10-15

10-14

10-13

h e

quiv

alen

t no

ise

spec

trum

[/s

qrt(

Hz)]

101 102 103 104 Frequency [Hz]

h equivalent noise spectrum of TAMA300

2001/06 (DT6) 2002/08/31 (DT7) 2003/02/20 (DT8) 2003/11/04 (DT9)

h ~ 2 x 10-21 [/√Hz] @ 1kHz

6

Observable Range

* for optimal incident direction

567

1

2

3

4

567

10

2

3

4

567

100

2

3

Obse

rvab

le D

ista

nce

with

SNR

=10

[kpc

]

0.1 1 10 100mass of accompanying star [Msolar]

Distance of detecting inspirals with SNR=10 2003/11/04 (DT9)

Inspiral QNM ringdown

0.5Msolar-32.6kpc

1.4Msolar-72.5kpc

2.7Msolar-96.3kpc

10Msolar-21.9kpc

Range with SNR = 10 for inspiral GW and BH ringdown GW

7

CommissioningData Taking period actual data amount take note

DT1 8/6 - 7/1999 ~3 + ~7 hours continuous lock first whole system test

DT2 9/17 - 20/1999 31 hours first Physics run

DT3 4/20 - 23/2000 13 hours

-- 8/14/2000 World best sensitivity h ~ 5x10-21 [1/√Hz]

DT4 8/21 - 9/3/2000 167 hours stable long run

DT5 3/1 - 3/8/2001 111 hours

Test Run 1 6/4 - 6/6/2001Longest stretch of continuous lock

is 24:50keep running all day

DT6 8/1 - 9/20/20011038 hours

duty cycle 86%full-dressed run

DT7 8/31 - 9/2/2002 24 hours with duty cycle 76.7%Recycling,

h ~ 3x10-21 [1/√Hz],Simultaneous obs with LIGO & GEO

DT8 2/14 – 4/14/20031168 hours,

duty cycle 81.1%coincidence obs with

LIGO S2

DT910/31(Actually 11/

28)/2003 – 1/5/2003

weekday: night timeweekend: full time

partial coincidence run with LIGO S3trying ‘crewless’ operation

8

DT9 – on going –

9

Search for GW events

10

Search for GW:

Burst Gravitational Wave

1. Target SourceSupernova core collapseFrequency band : a few 100 Hz – a few kHzWithout strict waveform assumption

2. Excess power filterSpectrogramIntegration : Df - Dt

3. Non-Gauss noise rejectionSpike like <–> level drift

11

Burst GW:

Excess power filterraw data signal Spectrogram (t-f plane)

Integration for Frourie domainDf = 500 Hz, Dt =200 msec

12

Noise behaviormean power VS 2nd moment of power fluctuationraw data -> time slicej-th time slices -> parameter

mean power of trend:

2nd moment of power fluctuation

Burst GW :

Non-Gauss noise rejection

C2 =12

(< P 2

j >

< Pj >2− 1

)C1 =< Pj >

See the talk by Masaki Ando : “Search results for burst gravitational waves with TAMA data”at Thursday 18th, session “event Search III : Burst”

13

1. Known wave formcoalescence of compact binaries ;NS-NS, NS-BH, BH-BH, PBMACHO

2. Known noise spectrum in Fourier domain3. Linear system

signal: s(t) = n(t) + a h(t)noise component :n(t), GW signal: a h(t)average noise power spectrum: Sh(f)

template waveform: h(t)signal-to-noise ratio:

chi^2 test

Search for GW:

Inspiral Gravitational Wave

ρ(τ ; parameters) = 2∫ f2

f1

h̃∗(f) · s̃(f)Sh(f)

e−i2πfτdf

SNR = ρ/√

2

14

Observable Range for Inspiral GW

56

1

2

3456

10

2

3456

100

2

3

Obse

rvab

le D

ista

nce

with

SNR

=10

[kpc

]

0.1 1 10 100mass of accompanying star [Msolar]

Distance of detecting inspirals with SNR=10 2003/11/04 (DT9) 2003/02/20 (DT8) 2002/08/31 (DT7) 2001/06 (DT6)

0.5Msolar-32.6kpc

1.4Msolar-72.5kpc

2.7Msolar-96.3kpc

10Msolar-21.9kpc

SNR =√

2 A

[4

∫f− 7

3

Sn(f)df

] 12

A = T!c

d

(5µ

96M!

) 12

(M

π2M!

) 13

T− 1

6! T! =(

G

c3

)M!

15

Event (r/√c2) histogram

DT8 search Preliminary result

16

Efficiency for Galactic event

17

Upper limit1. DT2

Range (SNR=10): 3.4 kpcMass region: 0.3 - 10 Msolar Upper limit: 0.59 event/hour (C.L.90%)

2. DT4Range (SNR=10): 17.9 kpcMass region: 1-2 MsolarUpper limit: 0.027 event/hour (C.L.90%)

3. DT6Range (SNR=10): 33.1 kpcMass region: 1-2 Msolar ,Upper limit: 0.0095 event/hour (C.L.90%)

=83 event/yr4. DT8

Range (SNR=10): 42.2 kpc, Detection Efficiency ~60% for Galactic eventMass region: 1-2 Msolar ,Upper limit: 0.0056 event/hour (C.L.90%)

=49 event/yr 1-3 Msolar ,Upper limit: 0.0033 event/hour (C.L.90%)

=29 event/yrSee the talk by Hirotaka Takahashi : “Search for gravitational waves from inspiraling compact binaries

using TAMA300 data”at Wednesday 17th, session “event Search I : Inspiral”

50

40

30

20

10

0

Obs

erva

ble

Rang

e [k

pc]

20032002200120001999year

2

4680.01

2

4680.1

2

468

Uppe

r Lim

it C.

L.90

% [e

vent

/hou

r]

Range Upper Limit for Glactic event Upper Limit for evidence

18

Search for GW:

Black-hole QNM ringdown GW

1. BH formation (by compact binary, etc.)-> quasi-normal mode GW• dumped sinusoidal waveform “ringdown”• mass and Kerr parameter determine the waveform

QNM

h(t) = Ae−π fctQ sin(2πfct)

fct ∼ 3.2 × 104

M[1 − 0.63(1 − a)0.3][Hz]

Q ∼ 2.0(1 − a)−0.45

19

BH ringdown: Observable rangeAssuming the BHs formed from binary coalescence -> Flanagan & Hughes, Phys.Rev.D57

(* perturbation theory may not predict the amplitude... )

567

1

2

3

4567

10

2

3

4567

100

2

3

Obse

rvab

le D

ista

nce

with

SNR

=10

[kpc

]

3 4 5 6 7 8 91

2 3 4 5 6 7 8 910

2 3 4 5 6 7 8 9100

2 3

mass of accompanying star [Msolar]

Distance of detecting QNM ringdown with SNR=10 2003/11/04 (DT9) 2003/02/20 (DT8) 2002/08/31 (DT7) 2001/06 (DT6)

target mass range

20

Templates for BH ringdown

Design the template bank• Strictly orthogonal and

normalized parameters, which correspond to Physics meaning

• Efficient for Matched Filter algorithm

Minimal match ~98%,# of templates ~800

Nakano et al. (gr-qc/0306082, PRD 68, 102003(2003).)

See the talk by Hiroyuki Nakano : “Effective Search Method for Gravitational Ringing of Black Holes” in ‘poser session’

21

Search for BH ringdown

1. Matched Filter techniquesimilar to ‘Inspiral search’

2. Detection efficiency estimationassumption: amplitude, radiation pattern of fundermental (l=m=2) mode, glactic distribution.Monte-Carlo (embed ringdown GW in real TAMA data)

3. Veto studyreject spurious signals due to noises (spikes, glitch, etc.)

22

Search for BH ringdown

Typical signal amplitude

Detection efficiency by Monte-Carlo

See the talk by Yoshiki Tsunesada : “earch for black hole ringdown gravitational waves in TAMA300 data”at Wednesday 17th, session “event Search I : Inspiral”

0.01

2

4

68

0.1

2

4

68

1

Dete

ction P

robability

1002 3 4 5 6 7 8 9

10002

Ringdown frequency fc [Hz]

SNR>10 SNR>20 SNR>30 SNR>40 SNR>50 SNR>100

23

Search for GW:

Continuous GW from SN1987A remnant

1. Assumptions:SN1987A remnant pulser

Large spindown rate 2–3 x10-10 Hz/sSearch range: 934.908 ±0.05 Hz

2. 1200 hours TAMA data (dt4,dt6)3. Upper limit:

h ~ 5 x 10-23 (C.L> 99%)

-> Soida et al. Class. Quant. Grav.Vol.20. No.17(2003)S645

24

Data quality evaluation

25

Data Quality:

Online veto study

Various noise induce at anywhere in the control servo loop

check by calibration signal injectiononline evaluation

26

Online ‘noise budget’ estimation

See the talk by Daisuke Tatsumi : “Online Veto Analysis of TAMA300”at Friday 19th, session “Detector Characterization III”

27

Cooperation : LIGO-TAMA

28

LIGO–TAMA coincidence

1. MOU was approved at Dec.2002for joint analysis, exchange the operation informations, and share some resources.

2. Coincidence observation for S2–DT8overlap duration of all x4 detectors: 250.7 hrs

3. Advantage of Multi-DetectorSky coverage improvementSource direction determination

4. Physics TargetCompact binary coalescenceBurst GW from super-novaeTrigger by external observation as GRB

5. Joint working group kicked off

29

STEP 2

STEP 1

LIGO–TAMA coincidenceCoincidence Schematics

TAMA LIGO(LHO1, LHO2, LLO)

event candidates lists event candidates lists

AND(=coincidence)

event behavior(waveform, amplitude, -> coherence)

upper limit / significancy

upper limit / significancy

search by own data filter evaluatation(efficiency, fake

rate, etc.)

search

See the talk by Patrick Sutton : “Status and Plans for the LIGO-TAMA Joint Data Analysis”at Friday 19th, session “Multi-Detector Analysis”

30

Remarks1. TAMA’s sensitivity and stableness of operation

have been progressed steadily. 2. Data Taking 9 is now held with trying ‘crewless’

operation.3. Following event search tasks are going in

TAMA ;Burst GWInspiral Gravitational WaveBlack-hole QNM ringdown GWContinuous GW from SN1987 remnant

4. Data Qualification is trying as online issue. for noise budget and veto.

5. LIGO-TAMA coincidence analysis are going.