The Mysterious Connection between Superluminous Supernovae and Gamma-Ray Bursts STScI Tue, May 24, 2016 2:00 PM Maryam Modjaz: Statistical analysis of host galaxies of SN Ic-BL (with and without GRBs) and spectra of the Ic family SN Zoo Type I - no H Type II - H Ib - no H, but He Ic - no H, no He are there differences between Ic-BL with and without GRBs? will look at hosts and SNe in the age of Big Data (samples of tens and twenties), how can we answer some of these questions? SN-GRB prefer to occur in low-metallicity hosts compared to those without GRB must be careful to account for selection effects why? low-metallicity progenitor? SN 2009bb was Ic-BL without GRB at high metallicity probably had central engine could it be dust? probably not could it be star-formation effect? some galaxies produce high-mass stars more efficiently probably not -- even after accounting for SF, still more GRBs in low-metallicity environments Is connection with metallicity a CORRELATION or a CAUSATION? Look at SN-GRBs compared to other stripped-envelope SNe oxygen abundances at SN sites must be careful to measure SN position properly trend with metallicity: Ic-BL with GRB at low metallicity, Ic-BL without GRB next Ib next Ic next at highest metallicity but careful: based on heterogeneous surveys, sometimes targetted need to move to homogeneous, untargetted survey so I'll be using PTF PTF - Palomar Transient Factory (now iPTF) measuring spectra of host galaxies for 90 SNe, un-targetted survey how to measure metallicity take spectrum at position of SN or GRB avoid gradients in metallicity choose SNe with secure ID -- need many spectra of each event important to get spectra at many phases, esp for Ib need wide wavelength range remove stellar absorption account for uncertainties use different and independent oxygen abundance diagnostics beware of PP04 method include SNe from galaxy-un-targetted surveys PTF Host galaxy project: pyMCZ derive metallicity and uncertainties from spectra look at HII region strong lines use Marino et al. 2013 techniques computes oxygen abundance, E(B-V), q ** please be responsible: don't drink and compute abundances! https://github.com/nyusngroup/pyMCZ goal: compare mass-metallicity relationship of hosts of GRBs to non-GRBs understand SN Ic vs. SN Ic-BL evolution over cosmic time results: 13 SN Ic-BL and 29 SNe Ic from PTF Ic-BL hosts have lower metallicities than Ic hosts we need SN rates as function of metallicity hosts of Ic-BL and GRBs tend to be overdense galaxies (that means compact?) high stellar mass and SRF densities this is a second parameter, not just due to metallicity is it possible this is related to binarity? is it due to a top-heavy IMF? SN Ic-BL with and without GRBs 407 spectra analyzed 40 objects Ic-BL with GRB tend to have higher velocities than those without GRBs by 6,000 km/s or so Velocities of SN Ic family from 22 SLSN Ic SLSN Ic have higher velocities and broader lines than normal Ic see poster at this meeting We have created an archive of all data http://www.cosmo.nyu.edu/SNYU/ also see Open Supernova Catalog Conclusions - is the era of statistical analysis - trends: SN-GRB have lowest metallicity, higher velocities - SN Ic-BL knows about GRB, doesn't care about GRB strength - SLSN Ic have higher velocities - next how to strip stars of envelopes at low metallicity? need more obs of relativistic SNe Ic-BL and off-axis GRBs need NIR spectra at 2 microns to look at He 2:25 PM Q: what are relative numbers of these classes? corr for selection effects A: selection effects are hard to correct better to do this with homogeneous dataset (PTF) Q: what are the velocities of ejecta? A: look at this graph, note changes with time Q: the statement that vels in SLSN are higher is a bit too simple it's more complicated, note plateau A: which lines used to measure velocity -- is important note that regular Ic have a plateau at late time, is just diffusion must be careful about claims of magnetar 2:29 PM Ranghild Lunnan: Host galaxies of Hydrogen-poor superluminous SNe the view from Pan-STARRS and HST aka SLSN-I see papers ApJ 787, 138 (2014) ApJ 804, 90 (2015) this talk will NOT cover SN PS1-14BJ: H-poor SLSN arXiv:1605.05235 Goal: constrain SLSN progenitor population what stars explode as SLSNe? probably never see one so nearby that we can see progenitor directly so look at host galaxy properties pair-instability and CSM interaction models require VERY massive stars and probably low metallicity magnetar model requires fast initial spin, but not necessarily a VERY massive star Early non-detections SLSNe only found in non-targetted surveys host galaxies tend not to be found in SDSS, they are faint search in GALEX and SDSS for host galaxies of SLSN confirms this: many non-detections, detections are faint and blueish cases of SN2010gx and PS1-10bzj very low metallicity high specific star formation rate young stellar population, approx 5 Myr first comprehensive SLSN-I host galaxy study 31 H-poor SLSNe, 15 from Pan-STARRS redshift 0.1 < z < 1.6 optical and NIR photometry optical spectroscopy for half the sample (low redshift) galaxy luminosities, masses, SF rates, metallicities overlaps between studies Lunnan et al. 2014 has 31 galaxies, median z=0.52 Leloudas et al. 2015 has 17 galaxies, median z=0.23 13 overlap with Lunnan Chen et al. 2016 15 galaxies, median z=0.16 11 overlap with Lunnan Angus et al. 2016 16 galaxies, median z=0.24 ... Results: SLSN-I are found in low-luminosity galaxies clearly different than regular core-collapse SNe many of the most extreme low-lum galaxies are found at low redshift (Q: is this due to selection effect?) SLSN-I are found in low-stellar-mass galaxies few x 10^(8) solar masses yes, "dinky" SLSN-I tend to be found in galaxies with high specific SF rate remember that these properties tend to be correlated in metallicity-mass diagram, SLSN-I hosts fall in corner 2 outliers, SLSN in high-mass hosts, found at large offsets HST imaging of SLSN environments the hosts looks like compact dwarfs lack of massive, grand-design spirals (which _do_ host normal core-collapse SNe) SLSN-I tend to be found in compact, irregular galaxies typical half-light radius 1 kpc how correlated are SLSN-I locations with star formation w/in host galaxies uses a pixel technique rank all pixels within a galaxy by brightness which pixel(s) hold SNe? GOODS CC SNe follow light Long GRBs are biased toward brighter pixels regions of star formation? dense environments? SLSN-I lie between the CC SNe and long-GRB distributions only 16 objects hard to draw conclusions, small statistics Conclusions: SLSN-I generally found in low-mass, low-metallicity hosts with high SF rate similar to trends of Long-GRB host populations local environments correlate with star formation need larger sample 2:51 PM Q: you mentioned CSM interaction requires very massive progenitors but it can occur with progenitors as low as 15-20 solar masses A: yes, thank you Q: are there biases in Pan-STARRS? A: yes, possibly host-less events are more exciting, get spectra but SLSN-II don't show some of these same effects Q: why different correlations with mass and blue lum? A: not so simple to connect blue lum and mass, depends on stellar population 2:54 PM Giorgos Leloudas: SUSHIES special thanks to Steve Schulze and Thomas Kruehler, and Javier Gorosabel SUSHIES I: MNRAS 449, 917 (2015) SUSHIES II: in prep based on larger sample, SED properties is there a driving factor behind SLSNe? SUSHIES I: results mass-metallicity relation for SLSN host galaxies H-poor SLSN hosts lie in low-mass, low-metallicity region GRB hosts also, to a lesser extent maybe low measured metallicities for SLSN hosts median 0.27 Z solar, 0.14 - 0.71 3 of them are < 0.07 Z solar no statistically significant difference with GRB hosts SLSNe II hosts are more metal-rich all are Type IIn metallicity 0.04 - 1.0 Z solar the most striking result: SLSN-I hosts have large equiv widths of 5007 e.g. are emission-line galaxies 50% are EELGs (equiv width > 100 Angstroms in O III 5007) very small prob of chance coincidence can see this same result in a emission-line-ratio diagram OIII / H-beta vs. NII / H-beta means stars have very high Teff specific star formation rate, measured with H-alpha once again, SLSN-I host galaxies lie far from the bulk of normal galaxies SUSHIES I conclusions: H-poor SLSNe and SLSNe IIn found in diff environments SLSN IIn progenitors are different (or a mixed bag) H-poor SLSNe show preference for young starbursts even more so than GRBs absolute ages are hard to measure, but we can use GRBs as a relative comparison SLSN-I represent the first explosions after a burst of star formation the presence of very massive stars is supported by hard ionization PTF12dam an EELG at z=0.107 host has very young stellar population age may be as small as 3 Myr, corr to > 60 solar mass (or even 120 solar) now, SUSHIES II, the followup paper based on 67 galaxies 51 H-poor SLSNe, 16 H-rich SLSNe can examine evolution with redshift for the first time slow (="R") and fast (="I") subsets of H-poor H-poor SLSNe seen at higher redshift brightness and colors H-poor hosts are faint 0.04 L-star at z < 0.5 H-rich are diverse H-poor are blue in (R-K) color -0.1 mag at z < 0.5 +0.5 mag for H-rich evolution above z > 0.7 survey bias or real? masses and SF rates SLSN-I hosts have low stellar masses SLSN-I hosts have high specific star formation rate quantifying the metallicity bias suggests that production of SLSNe drops sharply at metallicity > 0.4 solar Discussion is there one driving factor? metallicity, young age, star formation, etc. perhaps we don't understand stellar evolution very well I think that the correlation with starburst cannot be coincidence because dwarfs have burst SF, it is unlikely to catch them in starburst cycles of 1-2 Gyr, emission lines visible for < 100 Myr probability of 50% host galaxies are EELG is very small could low metallicity be result of young environment? points to unknown stellar evolution in these extreme environments What is next? need spectra at higher redshift have gone up to z = 2.2 so far 3:20 PM Q: comment: if young environments are important, would expect to see this in high spatial resolution HST images A: if SLSNe less correlated to UV light than GRBs, ... but UV light probes stars which are older than H-alpha light Q: keep in mind that binarity can be very important A: I agree Q: 6% of nearby galaxies are seen in starburst but how much of the mass is formed during these times? more than 6% of the stellar mass A: it helps to analzye measured properties - equiv width, rather than derived quantities I think we should explore effects other than metallicity Q: the transition between SLSN-I rates with metallicity is VERY sharp Q: what is the difference between spec of GRB-hosts and SLSN-I hosts? A: is largely a function of age Q: lack of H-rich SLSN at high z is a selection effect? A: yes, in part