The Mysterious Connection between Superluminous Supernovae and Gamma-Ray Bursts STScI Tue, May 24, 2016 9:02 AM Andrew Levan: Ultra-long gamma ray bursts Connection between long GRBs and SNe continuum of GRBs as function of t90? short GRB, long GRB, ultra-long GRB, and then TDEs = NS/NS, SN Ic, interesting ^^, SMBH + passing star it's the category of ultra-long GRBs which is topic for today is this a continuum, or a series of distinct physical mechanisms? does the duration impact the properties of the associated SN? The Christmas Burst, GRB 101225A lasted 20,000 sec in gamma-rays long-lived at other wavelengths could it be an asteroid falling onto NS? turned out to be a SN at z=0.85 with a very faint GRB-SN, mag -17 or so Host galaxies for ultra-long GRBs ex: GRB 111209A tend to be low-lum galaxies with low metallicity How does one define "duration" of a GRB T90 is defined to be duration containing 90% of total fluence easy to define for very bright sources not so easy for fainter sources, or with SWIFT SWIFT, for example, operates in a way that places interruptions in the datastream and automatically computed T90 values Lien et al. ApJS, submitted, have re-computed T90 in a different way their long-duration list does include the usual suspects, but some others, too their T90 values can be order(s) of mag longer than automatic T90 T(burst) Zhang et al. 2014 ApJ 787, 66 a duration of rapid activity before the standard afterglow provides useful information in some cases but not a silver bullet SWIFT "ultra-longs" tend to be 10x brighter than standard long GRBs example: T90 > 3000-5000 seconds continued gamma-ray detections rapid X-ray variability shows examples of events in this category GRB 051117A GRB 060607A, z=3.08 GRB 050904, z=6.29 even at z=6, HST could detect a UV-bright SN in IR Supernova connections GRB 091127 = SN 2009nz, BAT detection to 5500 sec host galaxy I = 22.5 looks like SN 1998bw GRB 101225A GRB 111209A Conclusions: Long GRBs are "always" associated with broad-line SNe Ic Ultra-long GRBs are not necessarily associated with SLSNe how to make a luminous SN from an ultra-long GRB late energy injection is important for SN luminosity need to worry about afterglow contamination, beaming, bolometric correction, etc. ultra-long GRBs have integrated fluxes up at 10^(56) erg-sec the longest events: TDE or SLSN? rapid switch-off around 100 days can occur in both models bumps in light curve sometimes appear detectability and rates of ultra-long GRBs long bursts tend to be spread out, so peak flux tends to be lower so harder to detect? not so easy for automated routines to notice the bursts SWIFT scans the sky, doesn't often sit and watch one spot for 1000s of sec so calculating rates can be difficult Summary - there's a population of very long duration transients - probably more common than expected - ultra-long GRBs which significantly impact their SNe are probably rare 9:26 AM Q: what was lum of SN assoc with Christmas burst? A: around -19.5, but tricky to compute Q: don't get excited by t^(-5/3) behavior, not nec TDE Q: do we expect diff chem environment for ultra-long GRBs? A: abs lines in spectra tend to be from not-so-local material Q: are there host galaxies for these ultra-long GRB events? A: not all of them are in tiny little galaxies not obviously different from regular GRBs 9:29 AM Elena Pian: GRB 111209A: A very luminous magnetar powered SN associated with an Ultra-long GRB SN at z=0.677 = SN 2011kl observations from ground over 70 days also SWIFT and HST data X-shooter spectrum, taken just around max light blue continuum weak lines due to line blending weak abs lines C II, CaII, TI III, Fe II model: magnetar with B = 6-9 x 10^(14) Gauss and P = 12 ms more similar to superluminous SNe than GRB-SN collapsar accretion rates (0.1 Msolar/sec) are incompatible with GRB duration: would need to accrete 100-1000 solar masses of material light curve compared to SLSNe and GRB-SNe 111209A lines in between the typical curves of these two types late time fades more rapidly than 56Ni, though not definitive peak mag -20 bolometric conclusions 111209A = SN 2011kl, H-poor SN luminosity intermediate between SLSNe and GRB-SN standard collapsar model has difficulties due to long duration spectrum at max is similar to those of SLSNe if powered by magnetar, P = 12 ms and B approx 10^(15) G 9:39 AM Q: is this intermediate in gamma-ray energy output? A: is on the high side, 6 x 10^(53) erg if isotropic Q: magnetar model doesn't explain observed energy of GRB and SN at same time A: yes, that's true Q: I'm not sure that there's a break in SN light curve A: yes, the feature is not terribly significant Q: do accretion rates need to be so high if engine is BH? A: good question, I'd like to talk with people about this I'm not sure if one can lower the accretion rates very much 9:45 AM Paolo Mazzali: GRB/SNe, SLSNe, SNe Ib/c: different but similar GRB/SNe are broad-lined SNe Ic ex: GRB 030329 = SN 2003dh GRB 980425 = SN 1998bw broad lines = large kinetic energy typically 10ish solar massses of ejecta, 0.4-0.5 solar mass 56Ni strong nebular Fe lines driven by 56Ni decay can look at 2-micron He I line to distinguish Ib (have it) from Ic (don't have it, no He) it takes only a little He to make a Ib only 0.1 solar masses of He can be seen it takes only a little H to make a IIb only 0.02 solar masses of H can be seen light curves very useful for 56Ni, spectra for Mej/Ek estimates Ek and Mej are correlated, 56Ni less so suggests more massive stars (> 35, 40 M solar) become SN Ic and GRB-SN What is the driving force? SN energy may dominate, esp after correcting GRB for beaming total energy of GRB and SNe of GRB-SNe are similar-ish SN Ib 2005bf showed a bright, late second LC peak did it have a magnetar? SN bl-Ic 2006aj more luminous than other GRB-SNe small 56Ni = 0.2 solar remnant likely to be a NS extra energy from magnetar GRB 111209A = SN 2001kl spectrum shows not-so-much suppression in near-UV from metal lines consistent with high-velocity SLSN SLSNe: ULGRB = SNeIc: GRB/HNe peculiar OII lines are result of non-thermal at high Temp He I lines appear later when temps are lower O II lines require non-thermal excitation (> 22 eV) perhaps X-rays could do it maybe a magnetar could create such X-rays velocity evolution SLSNe and HNe (Hyper-novae?) have high velocities but in SLSNe, high velocities are sustained over a long time (50-100 days) slow decline suggestive of magnetar powering SLSNe comparing properties SNe Ic --- SNe Ic/BL --- GRB/HNe from lower Mej to higher from lower Ek to higher SLSNe-"I" go with SNe Ic sequence [?] ULGRB/SN go with GRB/HNe sequence [?] these are all cores of massive stars GRB/SNe and ULGRB/SNe have highest Ek, Ek/Mej ULGRB/SNe not at the massive end of these ranges magnetar parameters GRB, XRF require rapid energy injection, large E/M SLSNe powered by interaction, need late injection 10:01 AM Q: could you repeat what you said about He lines? A: we see these He lines Q: other people don't identify these as He lines Q: ? A: ? Q: are you saying that the stars are different in these classes? A: actually, stars can be similar comment: there can be selection effects here A: my point is that one doesn't need super massive star Q: proprietary codes are difficult for others ... 10:06 AM Matt Nicholl: SLSNe with double-peaked light curves SN 2006oz: earliest example clear bump in g-band LSQ14bdq z = 0.345 spectroscopically typical SLSN-I broad LC, large ejecta mass zoom in this object's early LC early bump had peak -20 mag narrow, 15-day width combination makes it unlike normal SNe can't fit with 56Ni so early bump not a normal SN similar to double-peaked CC SNe SN 1993J, etc. shock-wave heating of outer layers, then cooling the second peak driven by different energy source try shock-cooling model for LSQ14bdg depends on progen radius, Ek/Mej use second peak to estimate Mej size was maybe 500 Rsolar, strangely extended Piro 2015 model 30 solar mass compact core with low-density material outside radius of 500-5000 solar any H would be transparent, so wouldn't see it interaction-powered bumps? CSM? no, difficult to power _both_ peaks Moriya and Maeda 2012 suggest dip due to ionization model doesn't explain fast rise time of early bump Kasen et al. 2015 model magnetar drives second shock in pre-expanded material how common are early bumps? 6 SLSNe show early flux excess in UV/optical sometimes dismissed as noise some show multi-filter deviations if common and similar, may argue against CSM interaction in fact, hard to exclude bumps for ANY SLSN often coverage not good enough SED of bump slope of UV SED consistent with blackbody 25,000 - 30,000 K consistent with shock-cooling models consistent with magnetar-powered shock breakout how often could we detect bumps? DES could detect out to z=2 in fact DES is already finding bumps see Smith et al. 2016, DES14X something Conclusions - many SLSNe exhibit fast, bright early peaks - requires extended material at high temperatures - SLSNe also show bumps, wiggles in later evolution need to continue high-cadence observations at late times! 10:18 AM Q: radio data also shows bumps and wiggles Q: some SNe IIn show similar early bumps and we know they are driven by CSM A: the SNe IIn have bigger separation between bumps SLSNe seem more homogeneous then SNe IIn we need a spectrum during early bump 10:20 AM Alexey Tolstov: Numerical simulations of multicolor light curves and spectra for SLSN PTF12bam three scenarios pair-instability magnetar CSM in CSM model, shocks form forward shock and reverse shock SLSN-I PTF12dam has slow-fading light curve "magnetar" fits are based on simplified one-zone models spin-down energy is converted to shell kinetic energy -- NOT into luminosity model: explosion inside an extended envelope ejecta 5 solar mass, "wind" 48 solar wind of He SN 2007bi: PISN and CCSN models exist long tail to light curve powered by 56Ni PTF12dam: combine PISN and CCSN models use numerical code STELLA (Blinnikov et al. 1998) 1-D hydro plus radiation opacity includes ... has been used to model many SNe use pre-SN C+O star (43 solar masses) plus extended envelope compare results of simulations to observations change parameters in fits, lower 56Ni mass, fainter tail larger energy, brighter peak good model has 42 solar mass Mej 38 solar mass M envelope 6 solar masss 56 Ni E = 20 x 10^(51) erg be careful when comparing bolometric curves to observed curves model fits temperature evolution well and velocity evolution well, though some issues at early times conclusions - combined model shock-wave interaction with non-H CSM envelope plus 56Ni - E51 = 20-30, Mej+env = 40 + 20-40 solar, M(Ni) = 6-7 solar, Radius = 10^(16) cm - open questions CO/He composition time scale of formation of envelope 10:32 AM Q: could you expand on magnetar aspects of model? A: not my specialty Q: can this model explain other SLSNe? A: yes, can use this approach for others but so far, only this event 10:33 AM - coffee break