Fundamental Properties of Novae in M31

TL;DR

This study uses the homogeneous M31 nova sample of 177 events to infer fundamental binary parameters (notably $M_ ext{WD}$, $\dot M$, $V_\text{max}$, and $P_\text{rec}$) by mapping peak bolometric luminosities and decline times onto the Yaron2005 nova models. Through interpolation of a coarse model grid and a careful grid-inversion procedure, the authors derive individual nova properties and construct observed distributions, revealing a Gaussian-like $M_ ext{WD}$ distribution centered around $\sim1.15$ M$_\odot$ and a bimodal $\log \dot M$ distribution with a dominant low-$\dot M$ population and an edge-associated high-$\dot M$ tail. After correcting for recurrence-time bias, intrinsic means shift to $\langle M_\text{WD} \rangle_\text{int} \approx 1.01$ M$_\odot$ and $\langle \log \dot M \rangle_\text{int} \approx -11.1$, implying very low secular accretion rates and long recurrence times compatible with a hibernation scenario. Comparisons with the Galaxy indicate broadly similar WD-mass properties but systematically lower accretion rates in M31, with only weak evidence for stellar-population-driven differences; the findings support irradiation-driven hibernation and highlight the need for more complete data and refined models to fully reconcile population-level discrepancies.

Abstract

The peak luminosities and rates of decline for a large sample of novae recently published by Clark et al. have been analyzed using the Yaron et al. nova models to estimate fundamental properties of the M31 nova population. The apparent white dwarf (WD) mass distribution is approximately Gaussian with a mean $\langle M_\mathrm{WD} \rangle = 1.16\pm0.14~M_{\odot}$. When corrected for recurrence-time bias, the mean drops to $\langle M_\mathrm{WD} \rangle = 1.07~M_\odot$. The average WD mass of the M31 nova sample is found to be remarkably similar to that found by Shara et al. in their study of 82 Galactic novae, but $\sim0.15~M_\odot$ more massive than the mean recently determined by Schaefer in his comprehensive study of more than 300 systems. As expected, the average WD mass for the recurrent novae included in the M31 sample, $\langle M_\mathrm{WD} \rangle = 1.33\pm0.08~M_{\odot}$, is significantly higher than that for novae generally. Other parameters of interest, such as the accretion rate, velocity of the ejecta, and the predicted recurrence time, are characterized by skewed distributions with large spreads about means of $\langle \log \dot M ~(M_\odot~\mathrm{yr}^{-1}) \rangle \simeq -9.27$, $\langle V_\mathrm{max} \rangle \simeq 1690~\mathrm{km~s}^{-1}$, and $\langle \log P_\mathrm{rec}~\mathrm{(yr)} \rangle \simeq 4.39$, respectively. The role of hibernation in affecting the $\dot M$ and $P_\mathrm{rec}$ distributions is briefly discussed. Finally, the nova properties were studied as a function of apparent position (isophotal radius) in M31, with the preponderance of evidence failing to establish any clear dependence on stellar population.

Figures (11)

• Figure 1: The interpolated model grids for our key input parameters, the bolometric luminosity at the peak of the eruption, $L_4$, (left panel), and the mass-loss timescale, $t_\mathrm{ml}$ (Right Panel) as functions of $\log \dot M$ ($M_{\odot}$ yr$^{-1}$) and $M_\mathrm{WD}$ ($M_{\odot}$). The interpolation was performed by fitting second-order polynomials to the nova models of Yaron2005 as described in section 3.2.
• Figure 2: The same as in Fig. 1 except showing the model interpolations for the auxiliary parameters, the maximum expansion velocity of the nova ejecta, $V_\mathrm{max}$ (left panel), and the predicted recurrence time, $\log P_\mathrm{rec}$ (right panel).
• Figure 3: Distributions of the observed nova properties based on the model predictions from Tables \ref{['tab2']} - \ref{['tab5']}, Top Left: The observed WD mass, $M_\mathrm{WD}$, distribution (known RNe shown in red). Top Right: The mass accretion rate, $\log \dot M$, distribution. The dashed region in the brightest bin shows the contribution of all novae (RNe in light red), including those with uncertain estimates of $\log \dot M = -7.0$, while the grey and dark red regions omit these novae. Bottom Left: The maximum ejecta velocity, $V_\mathrm{max}$, distribution (known RNe shown in red). Bottom Right: The recurrence time, log $P_\mathrm{rec}$, distribution. For known RNe, observed values of the recurrence times are shown.
• Figure 4: The observed distribution of the WD mass for He/N and Fe$\;$b M31 novae compared with the overall WD mass distribution. The red distribution show all known and suspected He/N and Fe$\;$b novae, while the blue distribution shows firmly established He/N novae. The average WD mass, $\langle M_\mathrm{WD} \rangle = 1.15\pm0.14$ M$_{\odot}$, $1.26\pm0.12$ M$_{\odot}$, and $1.32\pm0.09$ M$_{\odot}$ for all novae, known and suspected He/N + Fe$\;$b novae, and firmly established He/N novae, respectively.
• Figure 5: The observed distribution of the RN recurrence times (top panel) compared with the predicted RN recurrence time distribution based on the model fits (bottom panel). The model fails to accurately predict the recurrence times for three RNe (M31N 1984-07a, 2006-11c, 2013-10c).