Optical Counterparts of Neutron Star Mergers
Powered by the Decay of R-Process Nuclei
Brian Metzger
Princeton University
NASA Einstein Fellow
In Collaboration with
Eliot Quataert, Siva Darbha, & Daniel Perley (UC Berkeley)
Almudena Arcones & Gabriel Martinez-Pinedo (GSI; Darmstadt, Germany)
Dan Kasen (UCSC), Roland Thomas & Peter Nugent (LBNL)
Binary Compact Object Mergers
NS
NS
BH
NS
10 Known Galactic NS-NS Binaries
(Lorimer 2008)
Hulse-Taylor
Pulsar
Tmerge = 300 Myr
-5
-4
-1
˙
N
~
10
-10
yr
merge
(e.g. Kalogera et al. 2004)
Gravitational Waves from Inspiral and
Merger
Credit: Kip Thorne
“chirp
”
Ground-Based
Interferometers
LIGO 5th Science Run
(2007) Range ~ 10-30 Mpc
“Advanced” LIGO+Virgo
(~2015) Range ~ 300-600 Mpc
LIGO (North America)
Virgo (Europe)
Electromagnetic Counterparts of NS-NS/NS-BH Mergers
Importance of EM Detection:
• Place Merger into Astrophysical Context
 Host Galaxy, Local Environment, & Binary Properties
• Improve Effective Sensitivity of G-Wave Detectors
(Kochanek & Piran 93)
 Advanced LIGO Detection Rates Uncertain (~ 1 - 103 yr-1 )
• Cosmology: Redshift  Measurement of H0
(e.g. Krolak & Shutz 87)
Electromagnetic Counterparts of NS-NS/NS-BH Mergers
Short-Duration
Gamma-Ray Burst
Blinnikov+84, Paczynski 86;
Goodman 86; Eichler+89
Supernova-Like
Transient Powered by
Radioactive Ejecta
Li & Paczynski 98; Kulkarni 05;
Rosswog 05; Metzger+08, 10
Bright, but Beamed
Dimmer, but Isotropic
Importance of EM Detection:
• Place Merger into Astrophysical Context
 Host Galaxy, Local Environment, & Binary Properties
• Improve Effective Sensitivity of G-Wave Detectors
(Kochanek & Piran 93)
 Advanced LIGO Detection Rates Uncertain (~ 1 - 103 yr-1 )
• Cosmology: Redshift  Measurement of H0
(e.g. Krolak & Shutz 87)
•
B2FH: Type I SN light curves powered by 254Cf
•
Today: Type Ia SNe powered by 56Ni & 56Co
•
B2FH: Type I SN light curves powered by 254Cf
•
Today: Type Ia SNe powered by 56Ni & 56Co
Similar to a Supernova, but …
Faster Evolving
NS Merger Ejecta
How Supernovae Shine
(Arnett 1982)
Spherical ejecta w mass M, velocity v, thermal energy E = f Mc2, & opacity 
}
R
M
r=
R=vt
V
t ~ krR
M
4p R 3
3
t diff ~ t R
(
)(
)
c
1/ 2
æ
ö
çM
÷
Emission peaks when t = tdiff t peak ~ 2 weeks v 4
-1
è M8ø
10 km s
1/ 2
1/ 2
E(t peak )
æ
ö
43
-1
f -5 v 4
çM
÷
Lpeak ~
~
10
ergs
s
-1
t peak
M
è
10
10 km s
8ø
(
-1/ 2
)
Type Ia SN:
v ~104 km s-1, Mej ~ M, fNiCo ~ 10-5  tpeak ~ week, L ~ 1043 erg s-1
NS Merger Ejecta:
v ~ 0.1 c, Mej ~ 10-2 M, f ~ ?  tpeak~ 1 day, L ~ ???
How Supernovae Shine
(Arnett 1982)
Spherical ejecta w mass M, velocity v, thermal energy E = f Mc2, & opacity 
}
R
M
r=
R=vt
V
t ~ krR
M
4p R 3
3
t diff ~ t R
(
)(
)
c
1/ 2
æ
ö
çM
÷
Emission peaks when t = tdiff t peak ~ 2 weeks v 4
-1
è M8ø
10 km s
1/ 2
1/ 2
E(t peak )
æ
ö
43
-1
f -5 v 4
çM
÷
Lpeak ~
~
10
ergs
s
-1
t peak
M
è
10
10 km s
8ø
(
-1/ 2
)
Type Ia SN:
v ~104 km s-1, Mej ~ M, fNiCo ~ 10-5  tpeak ~ week, L ~ 1043 erg s-1
NS Merger Ejecta:
v ~ 0.1 c, Mej ~ 10-2 M, f ~ ?  tpeak~ 1 day, L ~ ???
Credit: M. Shibata (U Tokyo)
Tidal Tails (Dynamical Ejecta)
(e.g. Janka et al. 1999; Lee & Kluzniak 1999; Ruffert & Janka 2001; Rosswog et al. 2004;
Rosswog 2005; Shibata & Taniguchi 2006; Giacomazzo et al. 2009; Rezzolla et al. 2010;
Chawla et al. 2010)
Full GR / Simple EOS
Current Sims:
Mej ~ 0 - 10-1 M
Newtonian / Realistic EOS
Lee & Ramirez-Ruiz 07
Sources of Neutron-Rich
Ejecta
Tidal Tails (Dynamical Ejecta)
(e.g. Janka et al. 1999; Lee & Kluzniak 1999; Ruffert & Janka 2001; Rosswog et al. 2004;
Rosswog 2005; Shibata & Taniguchi 2006; Giacomazzo et al. 2009; Rezzolla et al. 2010;
Chawla et al. 2010)
Full GR / Simple EOS
Current Sims:
Mej ~ 0 - 10-1 M
Newtonian / Realistic EOS
Accretion Disk Outflows
Neutrino-Driven Winds (Early)
(McLaughlin & Surman 05; Surman+ 06, 08; BDM+08)
Thermonuclear-Driven Winds (Late)
(Metzger, Piro & Quataert 2008; Lee et al. 2009)
Mej ~ Mdisk/3 ~ 10-3 - 10-2 M
Neutron-Rich Freeze-Out Ye ~ 0.1-0.4
(BDM + 2009)
Lee et al. 2004
Lee & Ramirez-Ruiz 07
Sources of Neutron-Rich
Ejecta
How Supernovae Shine
(Arnett 1982)
Spherical ejecta w mass M, velocity v, thermal energy E = f Mc2, & opacity 
}
R
M
r=
R=vt
V
t ~ krR
M
4p R 3
3
t diff ~ t R
(
)(
)
c
1/ 2
æ
ö
çM
÷
Emission peaks when t = tdiff t peak ~ 2 weeks v 4
-1
è M8ø
10 km s
1/ 2
1/ 2
E(t peak )
æ
ö
43
-1
f -5 v 4
çM
÷
Lpeak ~
~
10
ergs
s
-1
t peak
M
è
10
10 km s
8ø
(
-1/ 2
)
Type Ia SN:
v ~104 km s-1, Mej ~ M, fNiCo ~ 10-5  tpeak ~ week, L ~ 1043 erg s-1
NS Merger Ejecta:
v ~ 0.1 c, Mej ~ 10-2 M, f ~ ?  tpeak~ 1 day, L ~ ???
Rapid Neutron Capture (R-Process)
Nucleosynthesis
Decompressing NS Matter  A ~ 100 Nuclei + Free Neutrons
(Lattimer et al. 1977; Meyer 1989; Freiburghaus et al. 1999; Goriely et al. 2005)
Protons
Chart of the Nuclides
Neutrons
Rapid Neutron Capture (R-Process)
Nucleosynthesis
Decompressing NS Matter  A ~ 100 Nuclei + Free Neutrons
(Lattimer et al. 1977; Meyer 1989; Freiburghaus et al. 1999; Goriely et al. 2005)
R-Process Network
Chart of the Nuclides
Protons
(Martinez-Pinedo 2008)
• neutron captures
(Rauscher & Thielemann 2000)
• photo-dissociations
• - and -decays
• fission reactions (Panov et al. 2009).
3rd
Abundance Peaks at
A ~ 130 and A ~ 195
2nd
BDM et al. 2010
Neutrons
Nucleosynthesis Calculations by G. Martinez-Pinedo & A. Arcones
Radioactive Heating of NS Merger Ejecta
Ye = 0.1
@ t ~ 1 day :
Ye = 0.1
 t-1.2
• R-process & Ni heating similar
• ~1/2 Fission, ~1/2 -Decays
• Dominant -Decays:
132,134,135 I, 128,129Sb,129Te,135Xe
Nucleosynthesis Calculations by G. Martinez-Pinedo & A. Arcones
Radioactive Heating of NS Merger Ejecta
Ye = 0.1
@ t ~ 1 day :
Ye = 0.1
 t-1.2
• R-process & Ni heating similar
• ~1/2 Fission, ~1/2 -Decays
• Dominant -Decays:
132,134,135 I, 128,129Sb,129Te,135Xe
fLP = 3 x 10-6
Results Robust to:
• ejecta composition
(Ye = 0.05 - 0.3)
• nuclear mass model
• outflow trajectory
(dynamically-ejected or wind-driven)
Light Curves
Color Evolution
Bolometric Luminosity
Blackbody Model
Monte Carlo Radiative Transfer (SEDONA; Kasen et al. 2006)
Peak Brightness MV= -15 @ t ~ 1 day for Mej = 10-2 M
Red Transient (Line Blanketing), Reddens in Time
CAVEAT: Fe composition assumed for opacity
does a pure r-process photosphere look like?
What
Metzger et al. 2010
“kilo-nova”
Three Detection Methods
1) Gravitational-Wave Triggered Follow-Up (See talks by Hughey,
Price & Kanner)
V < 22-24 to probe entire Advanced
LIGO merger volume (for MV = -15)
Positional Uncertainty ~ degrees
}
Wide-Field, Sensitive
Telescope (e.g. LSST)
Three Detection Methods
1) Gravitational-Wave Triggered Follow-Up
V < 22-24 to probe entire Advanced
LIGO merger volume (for MV = -15)
Positional Uncertainty ~ degrees
}
Wide-Field, Sensitive
Telescope (e.g. LSST)
Upper Limits
2) Short Gamma-Ray Burst Follow-Up
GRB 070724A
(Kocevski et al. 2009)
GRB 050509b
(Hjorth et al. 2005)
M ej < 0.1M8
M ej <10-3 M8
GRB 080503
(Perley, BDM, et al. 2009)
Possible Detection
Fundamental Obstacle? Bright Optical Afterglow
Early Follow-Up Observations of Short GRBs (courtesy Edo Berger)
Triangles - Upper Limits
Solid Squares - Detections
(known redshift)
Open Squares - Detections
(with likely redshift)
Open Squares w lines - Detections
(unknown redshift range)
M ej =10-2 M8
M ej =10-3 M8
GRB 080503:
Candidate Kilonova
(Perley, BDM et al. 2009)
Optical
Rebrightenin
g @ t ~ 1 day
Wheres the Host
Galaxy?
z = 0.561
Kilonova Parameters: v ~ 0.1 c, Mej ~ few 10-2 M , z ~ 0.1
Three Detection Methods
1) Gravitational-Wave Triggered Follow-Up
V < 22-24 to probe entire Advanced
LIGO merger volume (for MV = -15)
Positional Uncertainty ~ degrees
}
Wide-Field, Sensitive
Telescope (e.g. LSST)
Upper Limits
2) Short Gamma-Ray Burst Follow-Up
GRB 070724A
(Kocevski et al. 2009)
GRB 050509b
(Hjorth et al. 2005)
M ej < 0.1M8
M ej <10-3 M8
GRB 080503
(Perley, BDM, et al. 2009)
Possible Detection (but no redshift)
Fundamental Obstacle? Bright Optical Afterglow
3) “Blind” Optical Transient Surveys
(e.g. Palomar Transient Factory, Pan-STARRs, LSST) See talks by Quimby, Kasliwal
N˙ merge ~ 10-4 yr -1, Mej =10-2 M8  PTF ~ 1 yr
-1
& LSST ~ 103 yr-1
Conclusions
 Direct gravitational wave detection plausible within the decade
 Maximizing science requires identifying EM counterpart
 Remnant accretion disk may power short GRB
 but… beamed emission may limit use as gravitational wave beacon
 Two sources of neutron-rich ejecta from NS Mergers
 Dynamically-Ejected Tidal Tail (uncertain, but may be small < 10-3 M)
 Recombination Disk Winds (~30% of remnant disk mass or ~10-3-10-2 M)
 Optical transient powered by decaying r-process elements
 Nuclear Reaction Network: fLP ~ 3 x 10-6 , Robust to Uncertainties
 Radiative Transfer: Short Duration (~1 day), Low Luminosity (MV ~ -15)
& Red Colors (UV Line Blanketing)
 Detection & follow-up will require carefully-planned search strategy
& close cooperation btw astronomy & gravitational wave community