The cooling of massive white dwarfs from Gaia EDR3

Abstract

We determine the distribution of cooling ages of massive Gaia EDR3 white dwarfs identified with over 90 per cent probability within 200 pc and with mass in the range 0.95 − 1.25 M⊙. Using three sets of publicly available models, we consider sub-samples of these white dwarfs sorted into three equally spaced mass bins. Under the assumption of a constant white dwarf formation rate, we find an excess of white dwarfs both along the Q branch and below it, corresponding respectively to stars that are in the process of freezing and those that are completely frozen. We compare the cooling age distributions for each of these bins to the recently determined time-varying star formation rate of Gaia DR2 main sequence stars. For white dwarfs in the two lightest mass bins, spanning the mass range 0.95 − 1.15 M⊙, we find that the cumulative cooling age distribution is statistically consistent with the expectation from the star formation rate. For white dwarfs in the heaviest mass bin, 1.15 − 1.25 M⊙, we find that their cumulative distribution is inconsistent with the star formation rate for all of the models considered; instead, we find that their cooling age distribution is well fit by a linear combination of the distribution expected for single stellar evolution products and the distribution expected for double white dwarf merger products when approximately $40-50{{\ \rm per\ cent}}$ of the 1.15 − 1.25 M⊙ white dwarfs that formed over the past 4 Gyr are produced through double white dwarf mergers.

The Origin of Ultramassive White Dwarfs: Hints from Gaia EDR3

“…In this work, we use only the thick Montreal white dwarf cooling models (Bédard et al 2020). These models gave the best-fitting cooling age distribution for the 1.05−1.15 𝑀 and 1.15−1.25 𝑀 mass bins in Fleury et al (2022). For the 0.95 − 1.05 𝑀 mass bin, the Plata models (Camisassa et al 2019(Camisassa et al , 2017Renedo et al 2010) and thick Montreal models were both determined to produce cooling age distributions consistent with the star formation history and yielded similar p-values, so for consistency we use the thick Montreal models for all three mass bins.…”

“…We use the data and white dwarf cooling models described in Fleury et al (2022). The data consist of Gaia EDR3 white dwarfs identified in the Gentile Fusillo et al ( 2021) catalogue with at least 90% probability, as determined by the Pwd parameter of the catalogue, and located within 200 pc of the Sun.…”

“…Using the publicly available WD_models package provided by Cheng (2020) 1 , we use white dwarf cooling models to determine the photometric age and mass of each white dwarf from its G-band magnitude, 𝑀 𝐺 , and colour, 𝐺 BP −𝐺 RP . In our previous work (Fleury et al 2022), we assessed the consistency of the photometric cooling age distribution determined using various models available through WD_models with the star formation history observed for Gaia DR2 main sequence stars (Mor et al 2019). In this work, we use only the thick Montreal white dwarf cooling models (Bédard et al 2020).…”

“…In our recent work (Fleury et al 2022), we used the improved Gaia EDR3 data to re-investigate this cooling anomaly by analysing the photometric cooling age distribution of massive (0.95 − 1.25 𝑀 ) white dwarfs identified in the white dwarf catalogue of Gentile Fusillo et al (2021). We considered a variety of publicly available white dwarf cooling models of different compositions with standard core crystallization, and found cooling age distributions that were consistent with the expectation from the star formation rate observed for Gaia main sequence stars (Mor et al 2019).…”

“…In this work, we follow up the work of Fleury et al (2022) with a kinematic analysis of the transverse motions of the same sample of 0.95 − 1.25 𝑀 white dwarfs in the solar neighbourhood that was considered in Fleury et al (2022), sub-divided into the same mass bins of 0.95 − 1.05 𝑀 , 1.05 − 1.15 𝑀 , and 1.15 − 1.25 𝑀 and with masses and cooling ages determined using white dwarf cooling models found in Fleury et al (2022) to produce photometric cooling age distributions consistent with the expectation from the star formation history. For each of these mass bins, we estimate the transverse velocity dispersion as a function of age using different estimators and analyse the distributions of the separate transverse velocity components for several age ranges.…”