I also maintain profiles on Linkedin, Google Scholar and Researchgate.

Some of the Current Projects

Magnetar Powered Supernovae

with Ke-Jung Chen and Stan Woosley

Two dimensional CASTRO results for a magnetar powered supernova explosion of a bare CO core. The density profile is shown at t=0, 800, 1600 and 2400 seconds since the deposition started.

Here we explore in one and two dimensions, the properties of luminous supernovae whose light curves might be powered by embedded magnetars. Using the 1D implicit hydrodynamic code KEPLER we have calculated a large library of models based on both bare CO cores and full hydrogenic progenitors. A simple dipole formula is used for energy deposition in stars that, in most cases, have already been artificially exploded by a parametric piston. Good agreement with observations can be found in numerous cases, but a troubling feature for some of the models is the non-homologous expansion of the supernova when the magnetar energy is comparable to the explosion energy. A sample of such cases are explored in two dimensions using the compressible hydrodynamic code CASTRO, and we demonstrate the density spikes seen in one dimension result in the development of Rayleigh-Taylor and nonlinear deceleration instabilities, which drastically influence the morphology and mixing in the ejecta.

The 2D study just passed the peer-review and can be found here: Chen et al. (2016). The 1D paper will come out soon.

Overshooting In The Upper Main-Sequence

with Casey Meakin, Hassen Yesuf and Zeyana Musthafa

Figure 1 of Meakin et al. (2011) - overshoot parameter estimates from detached Eclipsing Binaries (diamonds) and asteroseismological data (triangles). The values adopted by Girardi et al. (2000) and Pietrinferni et al. (2004) for populations synthesis are labeled as G2000 and P2004. The mass uncertainties for the eclipsing binary data are negligible on the log(M) scale used. It is one big mess!

Our distant goal in this project is to tackle the problem of mixing at the boundaries in 1D calculations. Traditionally such situations are handled with the "step" or exponential overshooting prescriptions within the Mixing Length Theory (MLT) formalism. Since these phenomenological appoaches have little physical basis, the associated free-parameters are hardly constrained by the observations. In this work, we are applying the high precision data from detached eclipsing binary systems coupled with Bayesian statistics to constrain and compare various mixing models in 1D calculations.

Some of the Published/Submitted Works

The Most Luminous Supernovae

with Stan Woosley. Submitted to The Astrophysical Journal Letters.

Fig. 2 from the paper. Top: The luminous transient ASASSN-15lh (red bars) compared with a magnetar-powered model. Similar magnetars embedded in a smaller ejecta with lower opacity can, in principle, give even brighter light curves exceeding $\mathrm{10^{46}\ erg\ s^{-1}}$. Bottom: The same magnetar used in the top panel for fitting ASASSN-15lh is embedded in the ejecta of a massive red supergiant progenitor. The light curve is dimmer than the Type I case, but substantially brighter than the high energy prompt explosion (gray). The dashed part of curves mark the transition to nebular phase.

Inspired by the recent discovery of ASASSN-15lh (Dong et al. 2016), we investigate limiting cases from various scenarios - prompt explosions, collisions of shells, radioactivity and rotation. We argue rotation is the most powerful source of energy and any transient that is significantly brighter than $\mathrm{10^{44}\ ergs\ s^{-1}}$ in peak luminosity and has an emitted energy far greater than $\mathrm{10^{51}\ ergs}$, is either an extreme magnetar-powered case or it is unlikely to be a supernova.

This work got some press attention:
AAS Nova Highlight
Some French Article
Cosmos Magazine
I Fucking Love Science

Survey of CCSNe between $\mathrm{9-120\ M_{\odot}}$ based on Neutrino-Powered Explosions

with Thomas Ertl, Stan Woosley, Justin Brown and Thomas Janka. The Astrophysical Journal.

Fig. 13 from the paper. The explosion outcomes are shown for five different 'engines' all calibrated to SN1987A. Engines are listed in increasing strength. NS=neutron star; BH=black hole.

In this work, we explore the explosion outcomes of 200 solar metallicity progenitor models between $\mathrm{9-120\ M_{\odot}}$ through a novel approach. Instead of placing the piston in an arbitrary location and dial-in the final kinetic energy at infinity, we first create series of calibrated (to SN1987A and SN1054 observations) ''engines'' from the combination of one-zone analytic core-cooling model and approximate neutrino-transport. Then we place these engines inside each progenitor, and follow the evolution until the remnant and ejecta are clearly separated (neutrino-radiation hydrodynamics is calculated with Prometheus-HOTB as described in Ertl et al. 2015). As we give up the luxury of exploding any star in any way we want, now the outcomes are based on the neutrino-driven mechanism and therefore heavily dependent on the progenitor structure - i.e. each successful explosion for each engine has a different energy, remnant mass and nucleosynthesis. As predicted in our prior work on progenitors, most stars explode below $\mathrm{20\ M_{\odot}}$ with significant modulation, and above there is a sea of black hole formation with a distinct island of explosions near $\mathrm{27\ M_{\odot}}$. We map the explosion results back into KEPLER to post-process the resulting nucleosynthesis (excluding r-process but including Type-Ia contribution) with a large network, and follow the light curves through flux-limited diffusion.

All calculations from this work are publicly available and are being applied to problems ranging from explosion mechanism to galactic chemical evolution.

The Compactness of Presupernova Stellar Cores

with Stan Woosley. The Astrophysical Journal.

Fig. 1 of Sukhbold et al. (2016). The compactness parameter, $\mathrm{\xi_{2.5}}$ , enclosing inner $\mathrm{2.5\ M_{\odot}}$ at the time of presupernova is shown as a function of initial mass for 200 solar metallicity non-rotating progenitors.

Though the connection between the presupernova structure and the resulting explosion/implosion have been noted by several prior studies, little attention has been given to exactly why a given star ends up carving certain core structure before its death. In this work, we have studied the systematics of advanced stage evolution through dense grids of presupernova models (full hydrogenic and bare CO cores) calculated using both 1D implicit hydrodynamic code KEPLER and the open source stellar evolutionary code MESA, considering the uncertainties in the physics of stellar convection, mass loss and nuclear reaction rates. The presupernova core structure is found to be robustly non-monotonic as a function of initial mass, i.e. there is no single mass, where above stars are always harder to explode and below easier; rather most stars below $\mathrm{20\ M_{\odot}}$ are easier to explode, but above there are islands of explodability. We have found that the exact non-monotonic nature of the core structure is carved by the complicated interactions of carbon and oxygen burning shells with the carbon-depleted and oxygen-depleted cores. Whether the shell burning takes place within or outside the effective Chandrasekhar core, and whether the convective silicon burning shell includes or excludes the overlying carbon and oxygen burning shells, significantly modulates the final presupernova core structure. Due to these effects, in certain mass ranges, models differing by a mere $\mathrm{0.1\ M_{\odot}}$ in initial mass can produce vastly different presupernova cores. The uncertainties in the stellar convection is found to be the leading cause for large variations of models seen in the literature.