Late-time near-infrared observations of SN 2005df

We present late-time near-infrared (NIR) spectral evolution, at 200–400 days, for the Type Ia supernova SN 2005df. The spectra show numerous strong emission features of [Co ii], [Co iii], and [Fe ii] throughout the 0.8–1.8 μm region. As the spectrum ages, the cobalt features fade as would be expected from the decay of ⁵⁶Co to ⁵⁶Fe. We show that the strong and isolated [Fe ii] emission line at $1.644\mu m$ provides a unique tool to analyze NIR spectra of SNe Ia. Normalization of spectra to this line allows the separation of features produced by stable versus unstable isotopes of iron group elements. We develop a new method of determining the initial central density, ${\rho }_c$, and the magnetic field, B, of the white dwarf (WD) using the width of the $1.644\mu m$ line. The line width (LW) is sensitive because of electron capture in the early stages of burning, which increases as a function of density. The sensitivity of the LW to B increases with time, and the effects of the magnetic field shift toward later times with decreasing ${\rho }_c$. Through comparison with spherical models, the initial central density for SN 2005df is measured as ${\rho }_c=0.9(\pm 0.2)\times {10}^{9}g{cm}^{-3}$, which corresponds to a WD close to the Chandrasekhar mass, with $M_{WD}=1.31(\pm 0.03)M_{\odot }$ and systematic error less than $0.04M_{\odot }$. This error estimate is based on spherical models. We discuss the potential uncertainties due to multi-dimensional effects, mixing, and rotation. The latter two effects would increase the estimate of the WD mass. Within $M_{Ch}$ explosions, however, the central density found for SN 2005df is very low for a H-accretor, possibly suggesting a helium star companion or a tidally disrupted WD companion. As an alternative, we suggest mixing of the central region. We find some support for high initial magnetic fields of strength ${10}^{6}G$ for SN 2005df, however, $0G$ cannot be ruled out because of noise in the spectra combined with low ${\rho }_c$. We discuss our findings in the context of mixing by Rayleigh–Taylor instabilities during deflagration burning and a wide variety of explosion scenarios. Observations strongly support a very limited amount of mixing during a deflagration phase and high central densities characteristic of a $M_{Ch}$ WD.