Sources of Gravitational Waves
2
ndtalk
Astrophysical Sources of GWs
z
Binary systems (NS/NS, NS/BH, BH/BH)
z
Supernova
z
Bounce
z
Fall back
z
Oscillations & Instabilities
z
Old and Isolated NS
z
Cosmological origin
GW sources in ground-based detectors
Supernovae, BH/NS formation
Young Neutron Stars BH and NS Binaries
Spinning neutron stars in X-ray binaries
Sources in LISA
Galaxy mergers Galactic Binaries
Binary systems
(NS/NS, NS/BH, BH/BH)
The best candidates and most reliable
sources for broad band detectors
Coalescence of Compact Binaries
z During the frequency change from 100-200Hz GWs carry away5x10-3M~c2.
z In LIGOs band
z NS/NS (~16000 cycles)
z NS/BH(~3500 cycles)
z BH/BH(~600 cycles)
z The GW amplitude is:
z Larger total mass improves detection probability.
2 / 3 2 /
3
3
2 100
7.5 10 2.8 0
.7 100
M f Mpc
h M M Hz r
− ⎛ ⎞ ⎛ µ ⎞⎛ ⎞ ⎛ ⎞
≈ × ⎜⎝ ⎟ ⎜⎠ ⎝ ⎟⎜⎠⎝ ⎟ ⎜⎠ ⎝ ⎟⎠
: :
events/y ear
LIGO-I LIGO-II
NS/NS ~0.05 ~60-500
BH/NS ~0.02 ~80
BH/BH ~0.8 ~2000
Total 0.8 22000
•Phase effects are important, if the signal and the template get out of phase their cross correlation will be reduced.
Gravitational Waves from Binaries
Generically, there are 3 regimes in which black holes radiate:
z Orbital in-spiral: PN- approximations or point-particle orbits.
z Plunge/merger after the last stable orbit: numerical simulations or point- particle orbits.
z Ring-down of the disturbed black hole as it settles down to a Kerr hole: perturbation theory of black holes.
BH/BH coalescence
z The inspiral, merger, and ringdown waves from 50M: BH binaries as observed by initial and advanced LIGO.
z The energy spectra are
coming from crude estimates (10% of the total mass energy is radiated in merger waves and 3% in ringdown waves).
z We observe that the inspiral phase is not visible with initial LIGO, for this case Numerical Relativity is important.
Possible First Source:
Binary Black Hole Coalescence
z 10M~ + 10 M~ BH/BH binary
z Event rates based on population synthesis,
z mostly globular cluster binaries.
z Totally quiet!!
First gen
Second gen
100 M
pc inspiralRate: 1 / 300yrs to 1 / yr
merger
z=0.4
inspiral Rate: 2/mo to 10/day
merger
NS/BH Binaries
43 Mpc
NS-BH inspiral and NS Tidal Disruption
650 Mpc
NS-BH Event rates
z Based on Population Synthesis
z Initial interferometers
z Range: 43 Mpc
z 1/1000 yrs to 1per yr
z Advanced interferometers
z Range: 650 Mpc
z 2 per yr to several per day
Merging phase: NS/NS & BH/NS
z Tidal disruption of a NS by a BH
(Vallisneri)
z GWs could carry information about the EOS of NS eg. estimation of NS radius (15% error).
z The disruption waves lie in the band 300-1000Hz
z A few events per yearat 140Mpc (LIGO-II)
z Merging of NS-NS(Rasio et al)
z Imprint of the NS radii just before merging (ƒ11kHz)
z During the merging we could get important information about the EOS (ƒ21kHz)
Core-collapse Supernova
The most spectacular astronomical event
with exciting physics
Supernovae/gravitational collapse
Rate 1/30yr in typical galaxy Detection would provide unique insight into SN physics:
– optical signal hours after collapse – neutrinos after several seconds – GWs emitted during collapse
Supernova core collapse was the primary source of GW detectors.
GW amplitude uncertain by factors of 1,000’s?
Simulations suggest low level of radiation (<10-6 M:c2?), but
– rotational instabilities possible
– observational evidence for asymmetry from speeding final neutron stars (release of 10-6 M:c2 could explain 1000 km/s?)
– convective “boiling” observable to LMC
Core-Collapse Supernovae I
z Stars more massive than ~8M:end in core collapse (~90%are stars with masses ~8-20M:).
z Most of the material is ejected
z If M>20M: more than 10% falls back and pushes the PNS above the maximum NS mass leading to the formation of BHs (type II collapsars).
z If M>40M: no supernova is launched and the star collapses to form a BH (type I collapsars)
z Formation rate:
z 1-2per century in the Galaxy (Cappellaro & Turatto)
z 5-40%of them produce BHs through the fall back material
z Limited knowledge of the rotation rate! Initial periods probably <20ms.
z Chernoff & Cordesfit the initial spin with a Gaussian distribution peaked at 7ms. This means that 10%of pulsars are born spinning with millisecond periods.
Core-Collapse Supernovae II
Dimmelmeir, Font & Muller 2002 z Rotation increases strongly during the collapse.
•Multiple bounces are possible for a few models.
The ringing phase might last much longer
Core-Collapse Supernovae III
z GW amplitude
z Signals from Galactic supernova detectable.
z Frequencies ~1 kHz
z The numerical estimates are not conclusive. A number of effects (GR, secular evolution, non-axisymmetric instabilities)have been neglected!
(Axisymmetric collapse, Mathews-Wilson approximation…)
z Kicks suggest that a fraction of newly born NSs (and BHs) may be strongly asymmetric.
z Polarization of the light spectra in SN indication of asymmetry.
23 10
TT 10 Mpc
h d
−
Black-Hole Ringing I
z The newly formed BH is ringing till settles down to the stationary Kerr state (QNMs).
z The ringing due to the excitation by the fallback material might last for secs
z Typical frequencies: ~1-3kHz
z The amplitude of the ringdown
waves and their energy depends on the distortion of the BH.
z Energy emitted in GWs by the
falling material: ∆E>0.01µc2(µ/M)
3/10 2
9 0
20 1 1
/
[1 0.63(1 / ) ] 2(
3.2 Hz
) k
1
m M a
f
Q f
M π τ a
=
−
− − −
= ≈ −
d 1
ε µ
−
− ⎛ ⎞
⎛ ⎞
⎛ ⎞
≈ ×
Oscillations & Instabilities
The end product of gravitational collapse
Neutron Stars
z Suggested: 1932
z Discovered: 1967
z Known: 1070+
z Mass: ~ 1.3-1.8 Μ:
z Radius: ~ 8-14 Km
z Density: ~1015gr/cm3
Stellar pulsation primer
For spherical stars we can (in the Cowling approximation) write the Euler equations as 2
1
2 ( )
i
i i j
j
p
p A
t
∂ ξ δ ξ
∂ ρ ρ
⎛ ⎞ Γ
= −∇ ⎜ ⎟+ ∇
⎝ ⎠
Two main restoring forces, the pressure and the buoyancy associated with
internal composition /temperature gradients, lead to:
p-modes g-modes
2 2
2
( 1) s l l c
ω +r
≈
2
i
Ai p
ω gA ∇
≈− =
(δ p∝Ylm( , ))θ ϕ
NS ringing : Stellar Modes
z P-modes: main restoring force is the pressure
z G-modes: main restoring force is the buoyancy force
z F-mode: has an inter-mediate character of p- and g-mode
z W-modes: pure space-time modes (only in GR) (KK & Schutz) z Inertial modes (r-modes) :main
restoring force is the Coriolis force
z Superfluid modes: Deviation from
chemical equilibrium provides the main restoring agent
0.6 0.8 1.0 1.2 1.4 1.6 1.8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
w1-modes
p1-modes
f-modes
A B C D E F G I L G240 G300 WWF
M (in 1.4 solar mass units)
frequency (kHz)
Each type of mode is sensitive to the physical conditions where the amplitude of the mode is greatest.
Grav. Wave Asteroseismology
⎛ ⎞1/ 2 ⎡ ⎤
0.15 0.18 0.21 0.24 0.27 0.30 50
60 70 80 90 100 110 120 130 140 150
A B C D E F G I L G240 G300 WFF Q_1 Q_2 Q_3
M/R R ω w-mode
0.03 0.04 0.05 0.06 0.07 0.08 1.2
1.6 2.0 2.4 2.8 3.2 3.6 4.0
A B C D E F G I L G240 G300 WFF Q_1 Q_2 Q_3
(M/R3)1/2
ω f-mode
Grav. Wave Asteroseismology
⎤
⎡
1 M3 M
6 8 10 12 14 16
1.0 1.5 2.0 2.5
A B C D E F G I L WFF G240 G300
Radius (Km)
M (Solar Masses)
0.15 0.18 0.21 0.24 0.27 0.30 0.01
0.02 0.03 0.04 0.05
A B C D E F G I L G240 G300 WFF
M/R R4 / M3 τ f-mode
Unique estimation of Mass and Radius
Stability of Rotating Stars
Non-Axisymmetric Perturbations
Dynamical Instabilities
z Driven by hydrodynamical forces (bar-mode instability)
z Develop at a time scale of about one rotation period
Secular Instabilities
z Driven by dissipative forces
(viscosity, gravitational radiation)
z Develop at a time scale of several rotation periods.
z Viscosity driven instability causes a Maclaurin spheroid to evolve into a non-axisymmetric Jacobi ellipsoid.
z Gravitational radiation driven instability causes a Maclaurin
spheroid to evolve into a stationary but non-axisymmetricDedekind ellipsoid.
A general criterion is:
W
= T β
T : rot. kinetic energy W : grav. binding energy
27 .
≥ 0
β
The bar-mode instability I
For rapidly (differentially!) rotating stars with:
the “bar-mode” grows on a dynamical timescale.
If the bar persists for many (~10-100) rotation periods, the signal will be easily detectable from at least Virgo cluster.
–A considerable number of events per year in Virgo: ≤10-2 /yr/Galaxy
–Frequencies ~1.5-3.5kHz
dyn 0.27
T
β = W >β ≈
2
23 2
1.4 10
15 Mpc
9 10 0.2 3 kHz d
h − ⎛ ε ⎞⎛ f ⎞ ⎛ ⎞M R
≈ × ⎜ ⎟⎜ ⎟ ⎜ ⎟
⎝ ⎠⎝ ⎠ ⎝ ⎠
Remember mini-Grail: f0~3.2kHz
The pattern speed
d
dt m
ϕ ω
σ
= − =
z The pattern speed σ of a mode is:
z If a star rotates very fast, a backward moving mode, might change to move forward, according to an
inert rot
inert rot
m
σ
ω ω
σ
= + Ω
= +Ω
Chandrasekhar 1969: Gravitational waves lead to a secular instability
Friedman & Schutz 1978: The instability is generic, modes with sufficiently large m are unstable.
A neutral mode of oscillation signals the onset of CFS instability.
The CFS instability
•Radiation drives a mode unstable if the mode pattern moves backwards
according to an observer on the star (Jrot<0), but forwards according to someone far away (Jrot>0).
•They radiate positive angular
momentum, thus in the rotating frame the angular momentum of the mode increases leading to an increase in mode’s amplitude.
in rot
m m
ω = −ω + Ω
F-mode-(I)
z F-mode is the
fundamental pressure mode of the star
z It corresponds to polar perturbations
z Frequency for uniform density stars
z For ℓ=2 is ~2-4kHz
z Rotation breaks the symmetry:
the various -ℓ≤m≤ℓ decouple
z There is coupling between the polar and axial modes
z The frequency shifts:
2
3
1 3 4
2 ( 1)
2 1
growth time(if unstable)
( ) ~ 0.07 1.4 sec
10
l GW
l l GM
l R
R M R
t f l R
M M km
ω
+
= −
+
⎛ ⎞
⎛ ⎞⎟ ⎜ ⎟ ⎛ ⎞⎟
⎜ ⎜
≈ ⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠
:
inert( ) ( 0) m
ω Ω = ω Ω = +κ Ω
1 0 1 1 1 2 1 3 1 4 1 5 1 6
σ (kHz)
m = - 2 m = - 1 m = 0 m = + 1 m = + 2
The r-mode-(I)
z A non-rotating star has only trivial axial modes
z Rotation provides a restoring force (Coriolis) and leads in the appearance of the inertial modes.
z The l=m=2 inertial mode is called r- mode
z In a frame rotating with the star, the r- modes have frequency
z Meanwhile in the inertial frame
z r-modes are appear retrograde in the rotating system while in the inertial frame the prograde at all rotation rates!
rot
2 ( 1)
m ω = l l Ω
+
inertial rot 2
1 ( 1)
m m l l
ω = −ω + Ω = Ω −⎛⎜⎝ + ⎞⎟⎠
2
~
, , ~
r
u
u u
ϕ
θ
δ
δ δ δρ Ω
Ω
R-modes have:
Growth vs Damping
z Viscosity tends to suppress a GW instability.
z An instability is only relevant if it grows sufficiently fast that is not completely damped by viscosity
z Bulk viscosity: arises because the pressure and density variations associated with the mode oscillation drive the fluid away from beta
equilibrium. It corresponds to an estimate of the extent to which energy is dissipated (via neutrino emission) from the fluid motion as the weak interaction tries to re-establish equilibrium.
z Shear viscosity: in matter hotter than superfluid transition temperature T~109 K, due to neutron-neutron scattering, and for superfluids, due to electron-electron scattering
2
1 1 1
2
1
1
BV SV
GW
dE E dt
E dV
τ τ
ρ ξ
−τ
= + +
=
∫
Timescales
z Dissipation due to bulk viscosity
z Dissipation due to shear viscosity
z Dissipation/growth due to gravitational radiation
( )
SV
*
2 9
( )
2 2
,
0
2 ~
1
ab
ab
c
a b b a ab c
i m
dE dV
d
g T t
K
δσ ω ξ ξ ξ η
ηδσ δσ + Ω −
⎛ ⎞⎟
⎜ ⎟ = −
⎜ ⎟
⎜
⎛ ⎞⎟
⎜ ⎟
⎜ ⎟
⎜⎝ ⎠
= − ∇ ∇
⎠
∇ + −
⎝
∫
BV
6 9
2
, ( ) , ~
10 dE
d p K
T m p
t ζ σδ δσ i ω ζ
⎛ ⎞⎟
⎜ ⎟ =
⎜ ⎟
⎛ ⎞⎟
⎜ ⎟
⎜ ⎟
⎜⎝
= − + Ω ∆
Γ ⎠
⎜⎝ ⎠
∫
( )
( )
2 2
2 1 G
*
2
1/ 2
* W
( )
2
, 1
l
l B
l
l lm
l
m l
lm l
lm lm m
dE m N D
D r Y dV
d J
J l r Y dV
l t
δ δρ
ω ω δ
δ ρδυ υ
δ
δρ
∞ +
=
⎛ ⎞⎟
⎛ ⎞⎟
⎜ = − + Ω +
= ⎜ +
=
⎟
⎜ ⎟⎟
⎜⎝ + ⎠
⎜ ⎟
⎜⎝ ⎠
∑
∫ ∫
R-mode: Instability window
z For the r-mode (ℓ=2) we get:
z Instability window
z Many astrophysical
applications both on newly
5 9 6 2
BV
5 / 4 23/ 4 2
SV 9
4 0
8
W
1
G
1.4 10
10km 1ms sec
1.4 sec
10km 10
1 2.4 10
1.2 10
22 .4
10km 1ms
M R K P
M T
M R T
M K
M R P
M τ
τ τ
−
⎛ ⎞
⎛ ⎞⎛⎟ ⎞⎟ ⎜ ⎟ ⎛ ⎞⎟
⎜ ⎜ ⎟ ⎜
≈ ⎜⎜⎝ ⎟⎟⎟⎠⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠
⎛ ⎞ ⎛⎟ ⎞⎟ ⎛ ⎞⎟
⎜ ⎜ ⎜
≈ ⎜⎜⎝ ⎟⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠ ⎜⎜⎝ ⎟⎟⎠
⎛ ⎞⎛⎟ ⎞ ⎛⎟
⎜ ⎜ ⎜
≈ ⎜⎜ ⎟⎟⎟⎜⎜ ⎟⎟ ⎜⎜
×
⎝ ⎠ ⎝
⎝ ⎠
×
−
:
:
:
6
⎞⎟ sec
⎟⎟⎠
R-modes
(astrophysics)z GW amplitude depends on α (the saturation amplitude).
z Mode coupling might not allow the growth of instability to high amplitudes (Schenk etal)
z The existense of crust, hyperons in the core, magnetic fields,
affect the efficiency of the instability.
z For newly born neutron stars might be quite weak; unless we have the creation of a strange star
z Old accreting neutron (or
strange) stars, probably the best source! (400-600Hz)
22 1Mpc
( ) 10 Hz
h t 1 k
α d
− ⎛ ⎛ ⎞
≈ ⎜⎝ ⎟
Ω ⎞
⎜ ⎟
⎠ ⎠
⎝
2 3
10 10
α
−−
−Lindblom-Vallisneri-Tohline
F-mode (astrophysics)
z F-mode is naturally excited in any process.
• In GR the m=2 mode becomes unstable for Ω>0.85ΩKepleror β>0.06-0.08
• The instability window significantly smaller than the r-mode
• Detectable from as far as 15Mpc (LIGO-I), 100Mpc (LIGO-II) (depending on the
saturation amplitude).
• Differential rotation affects the onset of the instability
• Recent non-linear calculations by Shibata
& Karino (2004) suggest that:
• Up to 10% of energy and angular
momentum will be dissipated by GWs.
• Amplitude (ar ~500Hz): ⎛ ⎞ ⎛1/ 4 ⎞ ⎛3/ 4 ⎞
Isolated & Old NS
Isolated NS
z Wobbling or Deformed NS (many
interesting features but highly uncertain the degree of deformation)
z LMXBs : if accretion spin-up torque on NS is counterbalanced by GW emission then Sco X-1 and a few more might be detectable around 500-700 Hz.
⎟⎟⎠
⎜⎜ ⎞
⎝
⎟⎟ ⎛
⎠
⎜⎜ ⎞
⎝
× ⎛
≥ −
kpc r f
kHz
10 10 1
2
2
ε 8
LMXBs might be as robust source of GWs
Slowdown from pulsar
zUpper limits on amplitudes from known pulsars, set by assuming spindown due to the emission of gw energy. The points represent all pulsars with gravitational wave frequencies above 7 Hz and amplitudes above 10-27.
zExpected sensitivities of three first-generation interferometers in a one-year
The Wagoner mechanism (1984) Papaloizou &Pringle (1978)
Possible GW mechanisms:
– accretion induced asymmetry
– unstable r-modes: strong bulk viscosity may shift instability window to lower temperatures;
accreting stars can reach quasi-equilibrium state Key idea: Emission of GW balances accretion torque.
Strength of waves can be inferred from X-ray flux.
Requires deformation:
Observational evidence (?):
clustering of spin-frequencies in LMXB (250-590 Hz)
1/ 2 5 / 2
8
9
300 Hz 4.5 10
10 / s
M
M yr
ε ν
−
−
⎛ ⎞ ⎛ ⎞
= × ⎜ ⎟ ⎜ ⎟
⎝ ⎠
⎝ : ⎠
LMXBs & r-modes
UNSTABLE
5ms
Period
clustering of ms pulsars
STABLE Limiting Period 1.5ms
Fastest known pulsar 1.56ms
LIGO narrow banding
•LIGO-I phase
•The only detectable source is BBHs (10M:)
•LIGO-II phase (2006)
•Many sources…
Narrow banding for LMXBs
Stochastic Background
GW from the Big Bang
2
3H0
ρ =
gw gw
c
f d df
ρ Ω = ρ
d gw
f ρ
Ω = Stochastic background reflecting
fundamental physics in the early universe;
- Phase transitions - Inflation
- Topological defects
- String-inspired cosmology - Higher dimensions
After the Big Bang, photons
decoupled after 105 years, neutrinos after 1s, GWs before10-24 s!
Strength expressed as fraction of closure energy density;
14 5
10− ≈ Ω <10−
Detection: Requires cross-correlation of detectors.
Best window, free of “local” GW sources, is around 0.1-1 Hz.
Need LISA follow-on mission?
18 1Hz 2
10 ( )
h − ⎛ ⎞ h f
⎜ ⎟ Ω
GW from Inflation
The Dark Side of the Universe
-- Kip Thorne
z Our present understanding of the Universe is based almost entirely on electromagnetic radiation.
z Black holes can emit only gravitational radiation.
z More than 90% of the Universe is dark, but it still interacts by gravity.
z There are 5-10 times as many dark baryons as luminous ones.
z If part of the dark matter forms compact clumps, then
gravitational wave detectors will be the only way to see it directly.
The Rewards Are Huge
(Very) Early Universe
Gravitational Wave Observations
Stellar interiors Cosmology
Quantum theory
Astrophysics
Fundamental physics Extreme Gravity