About Me

Research

CV

Contact

Current Research

Time-Dependent Numerical Simulations of Type-I X-ray Burst Winds

Advised by Shane Davis


In my primary research work, I calculate time-dependent numerical solutions for winds driven from the surfaces of neutron stars (NS) undergoing photospheric radius expansion (PRE) during Type I X-ray bursts, supervised by Dr. Shane Davis in collaboration with Dr. Nevin Weinberg.


This study is motivated by the potential to measure NS radii using burst spectra to constrain the Eddington flux—a method that has seen success but highlights a need for a better understanding of the wind dynamics and composition in order to reduce systematic spectral modeling uncertainties (Ozel & Freire, 2016). Our work couples the 1D time-dependent MESA models of Yu & Weinberg (2018), which use the diffusion approximation for radiation, to an Athena++ simulation with realistic radiation transport across the optically-thick to optically-thin transition region of the wind. This work aims to reveal how optically-thin radiative transfer and the presence of heavy elements in the wind can impact burst spectra.

  1. Implicit Radiative Transfer I utilize the implicit Radiation Hydrodynamics (RHD) algorithm presented in Jiang (2021) to allow for larger global timesteps while still resolving the fine spatial scales near the NS surface. In the course of developing the Athena++ problem generator, I have also designed a suite of diagnostic outputs for the implicit solver that enables close examination of every component of the radiation field at each sub-step of the iteration. This has been an asset in debugging the sensitive inner radiation boundary condition.
  1. Coupling MESA and Athena++ at a Fixed Radius I have built infrastructure that allows the inner boundary of our Athena++ simulation to follow the evolution of the radiative flux and energy density as calculated by MESA, interpolated to a chosen radius near the NS surface. Through extensive trial-and-error, I have finally designed an adaptive boundary condition with enough flexibility to allow self-correction of destabilizing artefacts produced by differences in discretization between the two codes.
  1. A New MESA Data Backend for Athena++ To facilitate communication between these two powerful codes, I have written a performant, stand-alone MESA backend in C++ that is natively compatible with Athena++ and can seamlessly import data profiles at runtime.

My ongoing research on this project will form the central component of my thesis. This will include extensions of the problem to include general relativity and 2D axisymmetric geometry, exploring possible deviations from the commonly-presumed spherical symmetry of most existing PRE burst models. In conducting this research I have become comfortable making substantial modifications to the Athena++ source code in the exploration of new physics problems.

Exoplanetary Outflows Driven by Extreme Ultraviolet Heating and Lyman Alpha Radiation Forces

Advised by Phil Arras and Shane Davis


This work, supervised by Dr. Phil Arras, is motivated by Hubble observations of a handful of transiting exoplanets that show signatures of significant mass loss in their upper atmospheres where UV irradiation is known to play an important role in heating and accelerating the gas (Bourrier, Ehrenreich & Lecavelier des Etangs, 2015). In some cases, these transits exhibit blue-shifted absorption of Lyα with velocities of up to 100 km/s (Ehrenreich et al., 2015). The primary aims of the project are to (1) better understand and represent the underlying physics of resonance line radiative transfer in the context of exoplanet atmospheres, and (2) leverage this improved understanding to design a numerical simulation that may explain how the escaping atoms are accelerated to such large velocities.

  1. Semi-Analytic Solution for Resonance Line Radiation I accomplished aim (1) by finding a semi-analytic solution to resonant scattering in a uniform sphere of large optical depth, resolving a long-standing discrepancy in previous analytic solutions of this problem (c.f. Harrington 1973; Dijkstra, Haiman & Spaans 2006). In follow-up work, I designed and implemented a new resonance scattering method in the Athena++ Monte Carlo (MC) module that enabled direct numerical simulation of this problem in full 3D spherical polar coordinates and demonstrated close agreement with our semi-analytic solution in the limit of large Lyα line-center optical depth. These findings are published in McClellan, Davis & Arras (2022).
  1. Development of a Monte Carlo Acceleration Method Even at low physical densities, the large cross-section of resonance lines means that Lyα photons may undergo hundreds of millions of scatterings per hydrodynamic timestep. To circumvent this steep computational cost, I designed a MC acceleration scheme in which photons ”trapped” at large optical depths are evolved probabilistically rather than with a random walk. This algorithm integrates seamlessly with Athena++’s existing parallelization and load-balancing infrastructure and has been successfully deployed on NASA’s Pleiades High-Performance Computing Cluster. Synthesis of spectra from accelerated MC outputs demonstrates the new method’s convergence with both a standard random walk and our semi-analytic solution. This method was presented at the 2023 Athena++ Developer Meeting and is integral to my ongoing work on this problem.
  1. Extension of Athena++ Monte Carlo for Multiple Photon Sources In order to model the radiation incident on the planet’s upper atmosphere, I have extended the functionality of the Athena++ MC module to handle simultaneous evolution of multiple populations of photons with unique behavior depending on whether they are continuum photons or if they interact with the gas via resonant scattering. This enables direct simulation of photoionization in the turbulent upper atmosphere where ionization rate equilibrium prescriptions would fail to capture the correct dynamics.


In conducting this research I have gained a strong background in resonance line transfer, Monte Carlo methods, and understanding the 3D kinematics of winds in the context of atmospheric outflows. My contributions to this project, including advancements in the relevant theory as well as in the development of new physics modules for Athena++, have enabled the first time-dependent calculation of an irradiated exoplanet atmosphere in full 3D geometry including radiation forces from Lyα. A production run pipeline is currently in development and will be deployed on Pleiades in the next month. In follow-up analysis I will report the morphology and kinematics of the outflow alongside synthetic transit spectra, which will make for useful comparison with observed transit signatures (McClellan, Funkhowser, Davis & Arras, 2024, in prep).

References

Bourrier V, Ehrenreich D, Lecavelier des Etangs A. 2015. A&A 582:A65

Dijkstra M, Haiman Z, Spaans M. 2006. ApJ 649:14–36

Ehrenreich D, Bourrier V, Wheatley PJ, Lecavelier des Etangs A, Hebrard G, et al. 2015. Nature 522:459–461

Harrington JP. 1973. MNRAS 162:43

Jiang YF. 2021. ApJS 253:49

McClellan BC, Davis SW, Arras P. 2022. ApJ 934:37

Ozel F, Freire P. 2016. ARA&A 54:401–440

Yu H, Weinberg NN. 2018. ApJ 863:53