Since we do not know how core-collapse supernovae explode there is no requirement that the core collapses of all massive stars result in luminous supernovae, and, in fact, there are now several lines of evidence that suggest that 10-30% of core collapses fail and form a black hole without a dramatic external explosion. First, searches for supernova progenitors have failed to identify progenitors more than 17 solar masses (Smartt 2009). Second, the supernova rate appears to be lower than the massive star formation rate (Horuichi et al. 2011). Third, a population of failed core-collapse supernovae would naturally explain the observed distribution of neutron star and black hole masses (Kochanek 2014). Finally, although theoretical studies are unable to reliably make stars explode, they do tell us that stars more than 17 solar masses are intrinsically harder to explode (e.g., Pejcha & Thompson 2015).
Since 2008 we have been monitoring a million massive stars in 27 nearby galaxies with the Large Binocular Telescope to see if any vanish without producing a final luminous transient. The first four years of data has already yielded the first promising candidate (Gerke et al. 2015). We are following up this candidate in the hopes of confirming the birth of a new black hole and will continue to search for additional candidates in order to place tighter constraints on the fraction of core-collapses that result in failed supernovae.
Dr. Chris Kochanek and I are investigating the true nature of supernova "impostors". The most-widely accepted picture is that these transients are extragalactic analogs of the Great Eruption of Eta Carinae, in which a non-terminal eruption ejects a significant amount of mass that then may obscure the surviving star. However, in a recent paper, we show that the late-time evolution of the archetypical impostor, SN 1997bs, is inconsistent with this picture. Instead it appears the SN 1997bs might be a low-energy supernova. I am currently extending this analysis to other similar impostor events.
No supernova in the Milky Way has been observed since the invention of the telescope. It would be a tragedy to miss the opportunity to fully characterize the next one. To aid preparations for its observation I modeled the distance, extinction, and magnitude probability distributions for a Galactic core-collapse supernova, its shock breakout radiation, and its massive star progenitor. We found that a relatively modest IR camera would be capable of observing the supernova shock breakout even during the day, if the Super-Kamiokande neutrino detection experiment promptly disseminates pointing information when the next Galactic supernova occurs. We also showed that progenitor imaging will very likely exist in the IR, but that nearly half of potential progenitors are missing optical data. We further suggested that precursor variability could likely be easily constrained with a very shallow, synoptic IR survey of the Galactic plane. Finally, we used historical supernovae to infer a Galactic supernova rate.
While large numbers of field galaxies at redshifts as high as ~7 have been identified, only a small handful of galaxy clusters have been observed past z~2. Ordinary techniques for identifying clusters become ineffective at high redshifts due to various observational constraints. I searched for high-redshift galaxy cluster progenitors with Dr. Paul Martini by targeting 9 of the most luminous z~4 quasars, and successfully discovered an overdensity of Lyman-alpha emitters around SDSSJ 114514.18+394715.9.
I worked with Dr. Dennis Zaritsky at the University of Arizona to measure the rates of tidal features within galaxy clusters. Galaxy mergers drive the growth and evolution of galaxies. Tidal features are the signatures of past and ongoing mergers. We measured the rates of tidal signatures in a sample of 3500 galaxies from 54 nearby clusters to study the effects of the local and global environment on the incidence and lifetimes of these features. We found a deficit of tidally disturbed galaxies with decreasing clustercentric radius that is most pronounced inside of ~0.5 R200.