Supernovae are the violent deaths of stars, and my group uses multi-wavelength observations to decode the signatures they leave behind — revealing progenitor systems, explosion mechanisms, and the environments in which they occur.
As Co-PI (with Prof. Ryan Foley, UCSC) of the Young Supernova Experiment (YSE), we catch supernovae within days of explosion. YSE Data Release 1 delivered processed multi-color light curves for 1,975 supernovae observed with Pan-STARRS1 and ZTF between 2019 and 2020, together with photometric classifications using the ParSNIP classifier at 82% accuracy across Types Ia, II, and Ib/c — a public resource for algorithm development and preparation for Rubin/LSST (Aleo et al. 2023). High-cadence early-time photometry from Kepler/K2 tightly constrains progenitor systems: our analysis of the smooth power-law rise of the Type Ia SN 2018agk rules out non-degenerate Roche-lobe overflow companions at viewing angles below 45° (Wang et al. 2021). For the Type Ic SN 2020oi, we detected an early optical and ultraviolet excess inconsistent with standard shock-cooling, pointing instead to a ~9.5 M☉ binary progenitor interacting with chemically distinct circumstellar material near the explosion site (Gagliano et al. 2021).
Mass loss in the years before explosion leaves dramatic imprints on supernova spectra and light curves. SN 2021foa is a remarkable “flip-flop” transient that switched spectroscopic appearance between Type IIn and Type Ibn three times within 50 days of peak — a signature of helium-rich ejecta sweeping through a complex, multi-shell hydrogen-rich circumstellar medium expelled at ~2 M☉/yr (Farias et al. 2024). More broadly, our analysis of the largest sample of Type Ibn supernovae to date (61 events, 24 from YSE) reveals these are predominantly low-energy explosions from binary systems with compact helium-stripped envelopes surrounded by small CSM shells (Farias et al. 2025).
At the extremes of stellar mass, SN 2020acct is a rare double-peaked stripped-envelope supernova whose two luminous peaks — separated by 58 days with a factor-of-20 drop in flux between them — are best explained by pulsational pair-instability from a ~72 M☉ helium core, making it one of the most massive stellar explosions ever characterized (Angus et al. 2024).
Gravitationally lensed Type Ia supernovae open a new window on cosmology and on the accuracy of lens models. We used lensed SN H0pe to test seven independent lens modeling approaches, finding that predicted magnifications are systematically overestimated by more than 1 magnitude — a bias that, if uncorrected, propagates directly into errors on H₀ (Agrawal et al. 2025). With Rubin/LSST, Roman, and Euclid set to discover hundreds of lensed supernovae, developing accurate lens models is now a critical priority for precision cosmology.
