 |
Center for Nonlinear Studies |
Synthesizing the spectra of kilonovae, the radioactively powered optical/IR counterpart of neutron star mergers, involves the simulation of radiative transfer in matter with opacity dominated by lines (transitions of electrons between bound atomic states). In general the computation is costly, given the high number of lines, the high optical depths, and the highly non-linear coupling between the photons and matter. In the supernova/kilonova radiative transfer code SuperNu, Monte Carlo transport is optimized with Discrete Diffusion Monte Carlo (DDMC), which replaces many effective absorption-reemission steps with single diffusion steps, in a manner similar to random walk. As a trade-off, when Monte Carlo particles are propagated with DDMC, their resolved wavelength is lost. But photons can interact with lines by redshifting, a process that depends on having a fully resolved initial photon wavelength. How can the loss of wavelength in DDMC be reconciled with continuous redshift in the transport? In this talk we review some existing methods for treating Doppler shift in transport with diffusion optimization. We then present a method that is motivated by self-consistency of the comoving transport equation in the asymptotic diffusion limit, which to our knowledge is novel. Finally, we present some basic verifications, discuss some pathologies of the method, and present spectral results for a kilonova model.
Host: Timothy Waters