The Mysterious Connection between Superluminous Supernovae and Gamma-Ray Bursts STScI Wed, May 25, 2016 Announcement: workshop in September 12-16, 2016, in Munich 11:10 AM Cosimo Inserra: Spectropolarimetry of Superluminous Supernovae: a Step Towards Understanding their Progenitor Scenario Polarization and spectropolarimetry use polarization of light from explosion to gauge asymmetry Stokes vector S (I, Q, U) I = total flux Q = horiz/vert U = 45/135 angle polarization P = combination of Q, U, I = 0 for no preference in angle electron scattering can cause polarization line scattering leads to de-polarization in small spectral region if large-scale asymmetry of material, can get overall polarization SN 2015bn slow-evolving SLSN I or Ic z = 0.1136 risetime of 70 days in r-band 2 VLT spectra at -24d and +27d mag < 17.4 so high S/N = 400 E(B-V) = 0.02, very little ISM Polarization P(ISM) < 0.2 percent, good Results: before peak, polarization increases to read after peak, polarization increases to blue no obvious small-scale spectral features look in Q/U plane rough straight line in both cases so there is a preferential axiso how to explain? hard to explain dominance of electron scattering in the red combination of electron scattering and line pre-peak: photosphere almost spherical inner region beyond photosphere ellipticity 0.88 outer region beyond photosphereellipticity 0.88 post-peak: photosphere almost spherical inner region beyond photosphereellipticity 0.68 <-- different outer region beyond photosphereellipticity 0.88 similar behavior seen in SNe assoc with X-ray flashes mass of ejecta 6-11 solar masses, similar to 8-15 from Nicholl et al. conclusion: - SLSNe I show increase in polarization with time - have preferred axis and asymmetric ejecta - wavelength-dependent opacity -- most challenging to model - could be an increase in asymmetry with time in the inner layers 11:26 AM Q: wavelength dependence - lines are de-polarizing, lots of iron lines in blue, yet polarization increases there? A: yes, that's true, but we see overall combination of many lines decrease in polarization due to line is local effect, not global yes, this problematic Q: what about interstellar polarization? A: less than 0.2 percent in this case, used standard stars within 3 arcmin to the SN Q: but what about ISM polarization from host galaxy? A: true, but data in this case suggests little very little reddening strongly suggests little ISM polarization Q: how does this compare to normal SNe? A: some SNe have polarizations up to 3 or 5 percent, but this is similar to typical SNe Q: what about X-ray polarization? A: can't predict, but would guess one would see something Q: if ISM did polarize, wouldn't points lie in same location of Q-U plane at both epochs? A: not necessarily, contribution from various bits of ejecta changes over time Q: SN 1987A had very asymmetric explosion A: yes 11:33 AM Philipp Podsiadlowski: Binary Progenitors for Long-Duration Gamma-Ray Bursts and Superluminous Supernovae Long-GRBs and SLSNe are rare (1:1000 to 1:10,000 rel to normal CC SNe) need rare/exotic channel large diversity of binary interactions produce unusual events at such rates LBV SNe, pair-instability SNe, etc. Overview: - setting the scene - binary interactions and SNe - lessons from binary simulations - how to make progress Collapsars and magnetars rapidly rotating disc -> high accretion -> MHD jets extract rot KE if rotation required for LGRBs and SLSNe, then there is nothing mysterious about similarities in galaxy properties, but doesn't mean a causal connection but what distinguishes the types? the Rotation Problem massive stars lose ang mom efficiently lost in stellar winds strong coupling between envelope and core stellar wind mass loss depends on metallicity what makes some single stars rotate very rapidly? star formation process? pre-main-sequence merger? accretion from unseen companion? Binary Interactions can store lots of ang mom in orbital motion most stars are members of binary systems 70% for O stars with mass > 15 solar mass transfer more likely for post-MS systems binary interactions mass loss -> CSM mass accretion -> rotation common-envelope evolution -> CSM mergers -> rotation/CSM tidal interaction -> rotation wind Roche-lobs overflow -> CSM changes in core evolution/final fate tidal-spin up models requires compact binary: period < 10 hours ex: Cygnus X-3: WR + NS/BH core spin up in 10,000 yr but - at solar metallicity mass loss causes orbital widening need lower metallicity -> lower mass loss need later initial interaction channel to produce WR + NS/BH mergers? binary merger scenarios A: mergers of compact objects B: mergers of non-compact objects C: merger driving by mass transfer D: ... some models don't yield rapid rotating merger product better CO + He than He+He explosive common-envelope ejection slow merger of massive stars one star inside envelope of the other mass transfer from secondary to core of primary H-rich stream falls into He-burning layer of primary can create runaway nuclear reaction then eject He-rich shell and H-rich envelope can create clean C/O star, SN-Ic supernova in LBV phase? we don't expect it from single stellar models but different in binary systems if two O-stars merge after MS if core mass small enough, He burning as blue SG and might explode as blue SG, not extended envelope perhaps 0.001-0.01 of CC systems explode as blue SG could favor production of NS over BH large variety of outcomes SNe IIn, interaction SNe, PISNe tidally induced homogeneous evolution if very close binary, tidal locking can lead to homogeneous evolution can favor binary BH systems metallicity bias lower mass loss keeps fast rotation late mass transfer (case C) more common for low metallicity in general, low metallicity helps binary models, but is not required circumstellar medium do not assume steady wind in binaries! equatorial outflow could have circumbinary disks CE/LBV ejections can create circumstellar shells spiral-in of compact objects in common-envelope systems lead to ejections final fate of massive stars depends on previous binary interactions even very massive stars more likely to produce NS vs. BH in binaries cosmological simulations of exotic SNe simulate rates of GW sources as function of redshift and metallicity fit key observables: galaxy mass-metallicity, metallicity gradients can predict rates of pre-SN progenitor systems can predict rates of transients events connecting the pieces progenitor metallicities: host galaxies, locations of SNe progenitor masses: distance to HII regions, oxygen ejecta masses ejecta masses and compositions: explosion modelling, Ib vs. Ic rates of various events dynamical environment (see Blanchard's talk of yesterday) 11:58 AM Q: what about NS vs. BH in binary systems A: depends on when a star loses its envelope can produce WD as well if star loses envelope too early around 10 solar masses, second dredge-up makes complications Q: Is there any hope for single-star models? Ejecta masses are too small A: even in homogeneous models, some models have too much He (comment) regarding Ic vs. Ib, how to hide He, systems aren't hiding lots of He Q: why are Ib more or less common than Ic? A: that observed ratio has changed quite a bit over time Q: does explosive He-shell ejection work in low-mass systems? A: at very low masses, He can be transfered to companion stars but doesn't work for massive systems 12:03 PM Norbert Langer: the message from LIGO Talk about progenitors of GRBs, SLSNe and BH-mergers Let me convey a few messages from perspective of LIGO The problem how to get 10^(16) cm^2/s ang mom in pre-collapse core? start with massive stars survey of 606 OB stars specific SURFACE ang mom peaks at 10^(19) cm^2/s so initially, plenty of ang mom available but stellar mass loss causes spin down loss of ang mom can be very large for small mass loss rel loss of ang mom can be 10x mass loss if there is efficient ang mom transport inside the star and we think there is such transport, even without B-fields works for stellar winds and binary mass transfer the role of magnetic fields observations show detected mag field in about 10% of all massive stars generally slowly rotating so probably have spun-down due to B-field effect on evolution? no GRB or SLSN probably results after spindown but what about MW magnetars? post-MS expansion is another problem as envelope expands, core contracts envelope can brake core rotation especially if B-field present and significant we observe white dwarfs and NS which suggest that much of the ang mom IS lost, at least in some (most?) cases core spin-down is observed in low-mass red giants mass 1-3 solar masses can measure core rotation core rotation increases as envelope size increases so, expanded envelope is bad news for fast-rotating core how to keep rapidly-rotating cores? avoid mass loss avoid core/envelope structure solution: mix envelope with core burn _all_ H to He end up with compact He star does this happen? Chemically homogeneous evolution as rotation speed goes up, internal mixing time goes down for mass > 10-20 solar, mixing time similar to nuclear timescale this means chemically homogeneous evolution leads to more blue SG, rather than red SG (because compact) GRBs and SLSNe it is not coincidence that GRBs and SLSNe are type I in this picture, rapid rotation requires loss of H-rich envelope SLSNe-II and SLSNe-I with H are unlikely to be powered by magnetars alternative for H-rich events stationary ionization-confined shells confined by external ionization ex: Betelgeuse shells can contain several solar masses of material see Mackay et al., 2014 Nature Rates if strong biases exist, then ... to get 1 GRB out of 1000 SNe, only 10% of "suitable" stars must turn into GRB so this may suggest that some models are more likely than others Merging massive black holes and LIGO massive close binaries: tidally induced spin-up and mixing upgrade MESA code to include diff rotation and tides full-scale binary evolution grids 1-zone model grids and LIGO rates population synthesis no mass transfer, each star evolves at same rate create He stars each star goes SN alone leaves 2 BH which can merge metallicity must be low for this to work produces binary BH in range of 30-60 solar, then gap, then > 120 solar gap due to PISN which leave no BH BH mass ratios close to 1 Kerr parameters > 0.3 results very different from predictions of common-envelope evolution should be testable soon connection to GRBs and SLSNe each CHE-binary could provide 2 GRBs mass range of 60-120 solar could produce PISN summary: - chemically homogeneous evolution is necessary to produce GRBs, SLSNe - rates have potential to distinguish which channels are most common - LIGO has potential to show is if this chemically homogeneous evolution does happen 12:31 PM Q: what is range of expectations for delays between collapses in these chemically homogeneous systems? A: 10,000 - 100,000 yrs, too long to be observed Q: if one star goes SN, how does it not unbind the system? A: in chem homog system, very hard to unbind stars, not much mass loss (ex: pre-SN 51 solar masses -> post-SN 41 solar masses BH) Q: in chem homog evolution, is there obvious way to make strong mag fields? A: the core should remain rapidly-rotating, at least 12:35 PM Fruchter: Items from the Box of Mystery Not all superluminous SNe are superluminous Years ago, voted on "are GRBs and SNe related?" (actually, was specific to one specific GRB/SN association) result was split 50/50 Let's vote here: "Do magnetars explain EVERYthing (are main factor)?" we change the question to ... "Does same mechanism produce long-GRBs and SLSNe-I?" Yes: 17 No: 13 No Idea: 10ish Many thanks to people -- [I can't get names right] esp to Flory Hill Flory Hill has pictures -- in hardcopy and will E-mail to participants and thanks to Andrew Fruchter as well 12:45 PM lunch