New EoR Power Spectrum Limits From MWA Phase II Using the Delay Spectrum Method and Novel Systematic Rejection

Abstract

We present an analysis of Epoch of Reionization data from Phase II of the Murchison Widefield Array using the \texttt{simpleDS} delay spectrum pipeline. Prior work analyzed the same observations using the FHD/$\varepsilon$ppsilon imaging pipeline, and so the present analysis represents the first time that both principal types of 21 cm cosmology power spectrum estimation approaches have been applied to the same data set. Our limits on the 21 cm power spectrum amplitude span a range in $k$ space of $|k| < 1~h_{100}{\rm Mpc}^{-1}$ with a lowest measurement of $\Delta^2(k) \leq$ $4.58\times10^3$ mK$^2$ at $k = 0.190 h_{100}\rm{Mpc}^{-1}$ and $z = 7.14$. In order to achieve these limits, we need to mitigate a previously unidentified common mode systematic in the data set. If not accounted for, this systematic introduces an overall \emph{negative} bias that can make foreground contaminated measurements appear as stringent, noise-limited constraints on the 21 cm signal amplitude. The identification of this systematic highlights the risk in modeling systematics as positive-definite contributions to the power spectrum and in ``conservatively'' interpreting all measurements as upper limits.