Explaining extended emission GammaRay Bursts using accretion onto

Explaining extended emission Gamma-Ray Bursts using accretion onto a magnetar Paul O’Brien & Ben Gompertz University of Leicester 1 (with thanks to Graham Wynn & Antonia Rowlinson et al. )

GRB progenitors Long GRB: Collapsar Short GRB: Binary Merger LGRB: Collapsar model – occurs in region of massive (hence recent) star formation. Several examples known of associated super/hypernova signature SGRB: Merger model (e. g. NS-NS) – can occur in any type of galaxy, and also off of a galaxy due to natal dynamic kick and long merger time The “central engine” produced may be a either black hole or a “magnetar” 2

Extended emission GRBs Example: GRB 060614 T 90 = 103 s Redshift = 0. 125 No supernova detected – short? BAT Lightcurves Pluses: - Hard short episode followed by long softer hump - Short spectral "lag" (Norris & Bonnell) Minuses: - 5 s duration of hard episode - Brighter & more variable hump emission than others Could GRB 060614 be a new class (e. g. WD+NS, King et al. 2006) 3 •

Swift extended emission GRBs (Gompertz, O’Brien, Wynn, Rowlinson 2013) Similar luminosity extended “tail” Late plateau Swift EE GRB sample: look for >30 s of Extended emission (EE) (at 3 ) following a short (<few second) initial emission spike. The “Extended emission” looks similar in shape, duration and luminosity, suggesting a common physical process. Also see “late plateaus” (as in other SGRBs/LGRBs) 4

Example magnetar spin down fits (Rowlinson et al. 2013; Gompertz et al. 2013) SGRB examples Magnetar spin-down component Prompt decay PL component Relations between the initial spin period (P 0), dipole field (Bp), plateau luminosity (L) and magnetar spin-down time (Tem): L B p 2 / P 04 Tem P 02 / Bp 2 (Zhang & Mészáros 2001) 5 EE GRB example Model can fit the “late-time” plateau in EE GRBs But what about the EE tail?

Propellering and accretion A B C D E Schematic model: red circle = Alfvén radius (rm), green circle = co-rotation radius (rc). These depend on the magnetic dipole field (B) and spin period (P) respectively. A) High accretion rate suppresses rm – magnetar is spun up and rc shrinks B) As the accretion rate declines, rm expands C) When rm > rc matter outside rm is propellered away (producing EM emission) D) As accretion rate drops, rm expands, but rc also expands due to loss of ang. mom. E) When disk depleted, rc slowly increases as spin lost to dipole emission 6

Example fits using propellering (P) plus dipole spin-down (D) P D Poor fit at late times; maybe B varies? Assumed 40% EM propeller efficiency; 5% for dipole; <0. 9 c ejection velocity; exponential fallback rate fits better than power-law (Fernández and Metzger 2013) 7

EE GRB magnetar fit results Derived disk masses of 3 x 10 -3 to 3 x 10 -2 M and outer radii of 400 -1500 km (consistent with predictions for fallback disks, e. g. Lee et al. 2009). Initial spin period ~1 ms; B field strength ~1015 G 8

Magnetar results (Gompertz et al. 2014; Rowlinson et al. 2013) EE GRBs Filled symbol: use known z Open symbol: use average z Magnetic field strength <1017 G (approx limit based on speed of sound on surface of NS) Spin break up period for a 1. 4 Msolar NS (Lattimer & Prakash 2004) Not clear if such strong magnetar B fields or long lifetimes can occur Warning: points on this diagram from papers which assume different radiative efficencies 9

Summary • Generally get a good fit to the EE GRBs using a self-consistent combination of propellering and dipole spin-down emission for a magnetar+fallback disk model • To work, propellering requires the efficient conversion (>10%) of K. E. into EM emission during the propeller phase • Derived disk masses and sizes consistent with theoretical fallback discs • Best fits require exponential rather than powerlaw accretion rates – as expected in presence of strong outflows (Fernández and Metzger 2013) • Why do only some GRBs show an EE tail? Maybe these objects require a more unequal mass merger? • May be able to test magnetar model using predicted radio emission (i. e. detect the energy injected), or use GW (extra signal if magnetar collapses) 10

11

Outcomes from NS-NS merger (Usov 1992; Duncan & Thompson 1992; Dai et al. 2006 Metzger 2009; Metzger et al. 2011; etc) Expect a relation between the pulsar initial spin period (P 0), dipole field strength (Bp), luminosity (L) and the characteristic timescale (Tem) for spin-down: L Bp 2 / P 04 and Tem P 02 / Bp 2 12