Most of my research focusses on observational studies of the rare class of superluminous supernovae. I use data primarily at optical wavelengths (where supernovae emit most of their electromagnetic energy) to analyse the spectroscopic evolution and model the light curves of these events. This is done together with a wide range of collaborators in the US, UK, and around the world. I was involved in both the Pan-STARRS and PESSTO surveys.
One of the most significant results came from the first detection of a “slowly fading” superluminous supernova during the early phase before it reached maximum luminosity. We showed that the rate of brightening and the blue spectral energy distribution were incompatible with radioactively-powered explosion models, resulting in a letter to Nature. The data could instead be reproduced by the spin-down power of a millisecond magnetar.
A particularly surprising result was the discovery of pronounced precursor ‘bumps’ in the light curves of these objects. We were the first to resolve a double-peaked light curve in detail, and further showed that these may in fact be common. This morphology seems to indicate rapid cooling of extended material, either due to an inflated low-density envelope or an engine-driven shock breakout.
In 2015-2016, we carried out the most extensive monitoring campaign for any superluminous supernova to date. Nebular-phase spectra of SN 2015bn, obtained at almost 500 days after explosion, showed a remarkable similarity to normal luminosity hypernovae, which are sometimes associated with long gamma-ray bursts. However, no energetic gamma-ray burst was associated with SN 2015bn. Together these results imply an engine that deposits energy over a relatively long timescale, such that the energy input has to diffuse out of the expanding remnant rather than driving a GRB.
We also undertook the largest statistical study of superluminous supernovae, developing a brand new flexible, Bayesian light curve fitter (now publicly available!) to try to firmly pin down the ejecta and engine properties. Comparing our posterior distributions for key parameters to SLSN rates, we suggested that SLSNe mostly come from the top 10% of fast rotating, low-metallicity stars, greater than 20 solar masses. We then showed that the same framework can apply to the closest ever SLSN in a surprisingly metal-rich galaxy.
Having become interested in a wide range of astrophysical transients, one important question we have started to address is whether recently discovered fast radio bursts could be connected to any known classes of energetic explosions. We calculated rates for FRBs, SLSNe, normal supernovae and gamma-ray bursts, and determined that FRBs could come from the rare classes of SLSNe and GRBs in the decades after they explode. We also showed how statistics about their host galaxies hold the key to confirm (or ‘disconfirm‘) that hypothesis.