Synergy of GWs and EM signals

Transcription

1 Leiden, Feb Synergy of GWs and EM signals Gravitational wave facilities (LIGO, VIRGO ) Transient facilities (PTF, ZTF ) Stephan Rosswog

2 I. Gravitational wave detection LIGO & VIRGO detectors currently upgraded, increase sensitivity by factor > 10 (to h ) accessible volume enhanced by > factor 1000 expected to be online 2016 nsns and nsbh binary systems are main targets expected detection rates (Abadie et al. 2010): yr -1 nsns yr -1 nsbh large uncertainties!! first detections may be ambiguous/near detection threshold, additional signatures may give confidence, enhancement of detection efficiency

3 very difficult to localize accurately (~100 square degrees) Complementary information: Gravitational waves: physics of the binary system - masses - radii - neutron star equation of state - collapse to black What hole are the EM signatures -... of the main GW-sources? EM detection: astrophysical environment - redshift - type of host galaxy - location with respect to host galaxy kicks binary evolution - ambient medium -...

4 II. Expected EM signatures I. Short GRBs - collimated into ~8 o see ~ 1 out of 70 bursts II. Radioactively powered transients: macronovae / kilonovae (diff. kinds) - from dynamic ejecta - from neutrino-driven wind - from disk dissolution close to isotropic III. Radio flares (ejecta contain few erg of kinetic energy; decelerated by ambient medium)

5 III. Time scales I. Inspiral due to gravitational wave emission GR prediction: yields approximately talks by Wex van den Broeck Berti Sesana very broad range very sensitive to orbital period and eccentricity; set by stellar evolution last 100 km in < 0.7 s (Rosswog 2015, arxiv: )

6 II. Tidal interaction (a ~ 10 Rs) tid 0.5 s no tidal locking (Bildsten & Cutler 92, Kochanek 92) III. GW frequency time to merger chirp mass M chirp = µ 5/3 M 2/5 last ~ 17 minutes produce GWs with fgw> 100 Hz

7 III. Hypermassive neutron star (HMNS) merger dynamics produces differentially rotating remnant collapse only for ( Shibata+ 06, Hotokezaka+ 13, ) M>1.35 M TOV,T=0 > 2.7M the majority of systems passes through a hypermassive neutron star phase a fraction may possibly avoid collapse to bh dynamical time scale life time??? may depend on dirty details (angular momentum transport by B-fields, nu-winds etc.)

8 IV. Accretion disk dynamical time scale viscous time scale V. Neutrino escape diffusion time HMNS luminosity L E grav di 53 erg 10 s VERY optically thick ( 10 4 ) disk optical depth 1 luminosity L,disk L,HMNS

9 IV. Short GRBs: Why do we think they are caused by Compact Binary Mergers? standard model : GW-driven merger of either ns+ns or ns+bh ( compact binary merger ; CBM) (Blinnikov 84, Paczynski 86, Goodman 86, Eichler et al 1989, Narayan et al 92..) a) black hole + disk b) massive neutron star + disk sgrb property (e.g. reviews Lee & Ramirez-Ruiz 07, Nakar 07, Gehrels et al. 2009, Berger 11, 14; Zhang & Kumar 14) energy: E,iso erg E,true erg 8 2 CMB: E grav few erg variation: t few ms CMB: dyn. time scale ns+ns: dyn 1 p G =0.4 ms ns+bh: dyn g/cm 3 2 MBH 1 ms! K,ISCO 3M 1/2

10 duration: sgrb 0.3s CMB: viscous accretion visc 0.3 s 0.1 r 300 km 3/2 3M M 1/2 r/h 2 2 host galaxies: a) some in dead ellipticals b) majority in SF galaxies (but different from lgrb hosts) c) typical projected offsets 5 kpc CMB: GW-driven inspiral time GW = years P hr 8/3 M M 2/3 µ M 1 (1 e 2 ) 7/2 set by binary evolution from ns kicks during SN-explosion: a) large eccentricities e b) binary centre of mass motion broad distribution expected

11 event rates large uncertainties, but same order of magnitude NIR emission days after GRB (GRB130603B) radioactively powered el.mag. transients have been a prediction of the CBM model (Li & Paczynski 98, Kulkarni 05, Rosswog 05, Metzger et al. 10, Roberts et al. 11, Kasen et al. 13.) though originally for optical/uv peak after several days + NIR: consistent with high-opacity material from heavy r-process from dynamic ejecta (Freiburghaus, S.R., Thielemann 99, S.R. et al. 1999, Goriely et al. 11, Roberts et al. 11, Korobkin et al. 12, S.R. et al. 14, Mendoz-Temis et al. 14 )

12 Lorentz factors: > 100 set by energy-to-rest-mass-ratio m bar < M baryonic pollution E CMB generally non-trivial; easier for ns+bh systems, cleaner? late-time activity : a) extended emission b) X-ray flares time scales s CMB natural time scales (fractions of s) hard problem for CBM, suggestions (fallback? magnetar? ), but no consensus (Perley et al. 2009)??

13 S.R., Liv. Rev. Comput. Astrophys ns+ns-binaries vs. ns+bh-binaries observed systems: ns+ns: 10 (e.g. Lorimer 08) ns+bh: 0 est. coalesc. rates: ns+ns: MWEG -1 yr -1 ns+bh: MWEG -1 yr -1 for rotating bh: both horizon and innermost stable orbit move closer to bh innermost stable circular orbit (abh= 0) separation where mass-transfer sets in innermost stable circular orbit (abh= 1)

14 simple illustration: mass dependence ns: 1.3 Msol, Schwarzschild bh: 6 Msol ns: 1.3 Msol, Schwarzschild bh: 8 Msol a) low bh-masses and b) high bh-spins desirable mass and spin distributions in ns+bh binaries? which fraction of ns+bh systems produces a sgrb?

15 Late-time emission due to surviving magnetar? unclear whether/how baryonic pollution can be avoided recent work (Murguia-Berthier et al. 2014) suggests that HMNS should collapse early (<0.1 s) to avoid choking of jet by baryonic pollution currently unclear how one can have both a sgrb and activity at ~10 5 dynamical time scales

16 V. Radioactively powered transients ( macronovae / kilonovae ) closely related to Heavy element nucleosynthesis Solar system abundances: - from solar photosphere + meteoritic abundances - indicative of gas cloud from which solar system formed - contributions from different processes essentially two neutron capture processes in nature: slow n-capture ( s-process ) rapid n-capture ( r-process ) Big Bang stellar burning neutron captures half of the elements heavier than iron!

17 Compact binary mergers as r-process factories? suggested by Lattimer & Schramm 1974, Eichler et al. 1986, enough ejected : Rosswog abundances similar to solar system : Freiburghaus should produce EM-transient : Li & Paczynski 1998 all recent work finds ejecta very promising for r-process Why does it work? stability against ß-decay, condition on chemical potentials µ n µ p = µ e n! p + e + e, ß-equilibrium provides ) electron fraction Y e makes neutron star matter extremely neutron rich (> 90% neutrons, i.e. Ye < 0.1)

18 There are at least three different mass loss channels: a) dynamic ejecta (grav. torques, hydrodynamic interaction), b) neutrino-driven winds c) (late-time) disk dissolution d) ejecta differ in their properties (neutron-richness, time scales) different nucleosynthesis different heating rates different opacities different types of EM-transients

19 Channel 1: Dynamic ejecta typical merger case: 1.3 & 1.4 Msol, no spin visualized: Ye value at given optical depth total amount: Msol extremely neutron rich: Ye 0.03, with small crust contaminations velocity v 0.1 c

20 r-process calculations for dynamic ejecta (Korobkin, SR, Arcones, Winteler, MNRAS 426, 1940 (2012) ) delivers: - nuclear abundances: - solar-system like - extremely robust - only the heaviest elements (A>130) - heating rates

21 Long-term evolution of merger debris (Rosswog et al., 2014) typical merger simulations restricted to 20 ms, sound speed in neutron star 0.3c, CFL condition: Δt < Δx/cs 10-7 s cut out central remnant, replace by potential, follow ejecta include heating by radioactive decays follow evolution up to 100 years 1 day 1 year 100 years 100 years, but still in shape 2 1.4M z 5x10-4 pc 0.15 pc 15 pc x

22 Electromagnetic signals from ejecta: Macronovae ( Li & Paczynski 1998, Kulkarni 2005, Rosswog 2005, Metzger et al. 2010; Roberts et al. 2011;... Goriely et al. 2013; Bauswein et al. 2013; Rosswog, Piran, Nakar, 2013; Piran, Nakar, Rosswog, 2013; Rosswog et al., 2014; Grossmann et al. 2014; Tanaka et al. 2013,. ) similar to type Ia supernova: radioactive decays when matter becomes transparent el.magnet. transients supernova-like, but evolve faster and are dimmer from dynamic ejecta properties (κ= 10 cm/g; Kasen et al. 2013) infrared transient peak after 5 days, duration 10 days isotropic (from Grossmann et al. 2014; arxiv: )

23 The probably first macronova detection ( E. Berger et al., Smoking Gun or Smoldering Embers? A Possible r-process Kilonova Associated with the Short-Hard GRB B N. Tanvir et al., A kilonova associated with sgrb B, Nature ) June : - short Gamma-ray Burst GRB130603B, T s, z= nir-transient, present at 9 days, but faded away after 30 days 9 days 30 days after burst optical nir most natural explanation : macronova event If true: - short Gamma-ray Bursts caused by compact binary mergers - compact binary mergers are a major source of rapid neutron capture elements - isotropic macronovae promising accompanying signature for chirp GW signals

24 Channel 2: Neutrino-driven winds in HMNS phase similar to neutrino-driven winds in Core-collapse supernova/proto-neutron star merger remnant end of SPH simulation (from Perego et al. 2014) gravitational binding energy per nucleon: E grav 40 MeV M 2.6M 100 km r keep in mind: typical energies E e, e 20 MeV few captures can lift nucleons out of gravitational potential

25 Direct, 3D modelling of neutrino-driven winds (Perego et al b) Step 1: 3D SPH simulation, 2 x 1.4 Msol, opacity-dep., multi-flavour neutrino cooling, but NO heating [ color-coded: electron fraction Ye Ye= #electrons / #(neutrons + protons) ]

26 Step 2: nu-wind evolution (3D, Eulerian hydrodynamics code FISH (Kaeppeli et al 2012) + Advanced Spectral Leakage Scheme (ASL) (Perego et al 2014))

27 outflow properties depend on latitude high latitude (pole) more proton-rich (Ye~ 0.3) weak r-process (80 < A < 130) lower opacities low latitude (disk) more neutron-rich (Ye~ 0.25) weak r-process (80 < A < 130) + strong r-process (A> 130) larger opacities early ( 8 hours), bright (8 x erg/s), UV peak from high-latitude neutrino-wind ejecta + later ( 5 days), dimmer (2 x erg/s), IR peak from dynamic ejecta

28 VI. Conclusions dynamics that is so far testable seems well described by standard GR compact binary mergers are still the best model for sgrbs compact binary mergers are likely the source of the heaviest r-process elements; possibly produce even the whole mass range there are different channels of mass/loss nucleosynthesis they should produce macronovae June 2013 event there should be several flavours of such EM transients ejecta should produce radio flare (time scale years)

29 Gravity Open issues do we need alternatives to standard GR? what can we realistically learn about the ns-eos/nuclear matter properties from gravitational waves? EOS is anything more exotic than n-p-e/muons needed? maximum mass? could ns and bh co-exist in some mass range? Short GRBs how is the ultra-relativistic ( 300) launched? annihilation? MADs? Combination of different effects? is black hole formation needed? if not: how can baryonic pollution be avoided? what causes activity at very late times? which fraction of nuns-mergers can survive as HMNS? what is the role of nsbh binaries?

30 Nucleosynthesis production of all r-process elements? consistent with chemical evolution of galaxies? How? What are the opacities of the ejected material? Macronovae was observed event really caused by dynamic ejecta? different components: interaction? obscuration? What are the opacities of the different ejecta components?