The M31 Near-Infrared Period-Luminosity Relation and its non-linearity for $\delta$ Cep Variables with $0.5 \leq \log(P) \leq 1.7$

TL;DR

This study presents the largest near-infrared Cepheid sample in M31, revealing that the Period-Luminosity relations are better described by a broken slope at ten days, impacting Hubble constant estimates.

Contribution

Introduces a new MAD outlier rejection method for Cepheid classification and demonstrates the non-linearity of the PLR in M31 with implications for cosmology.

Findings

Small dispersion in PLRs (0.155 mag) despite random phased observations.

PLRs are better fit by a broken slope at ten days rather than a linear relation.

Using this sample as an anchor increases the Hubble constant estimate by 3.2%.

Abstract

We present the largest M31 near-infrared (F110W (close to J band), F160W (H band)) Cepheid sample so far. The sample consists of 371 Cepheids with photometry obtained from the HST PHAT program. The sample of 319 fundamental mode Cepheids, 16 first overtone Cepheids and 36 type II Cepheids, was identified using the median absolute deviation (MAD) outlier rejection method we develop here. This method does not rely on priors and allows us to obtain this clean Cepheid sample without rejecting a large fraction of Cepheids. The obtained Period-Luminosity relations (PLRs) have a very small dispersion, i.e. 0.155 mag in F160W, despite using random phased observations. This remarkably small dispersion allows us to determine that the PLRs are significantly better described by a broken slope at ten days than a linear slope. The use of our sample as an anchor to determine the Hubble constant gives a $3.2\%$ larger Hubble constant compared to the Riess et al. (2012) sample.