Comprehensive Listing of 208 Nova White Dwarf Masses As the Primary Determinant of Spectral-Class and Light-Curve-Class

Abstract

For Galactic novae, I calculate and collect a comprehensive catalog of 208 measures of white dwarf (WD) masses ($M_{\rm WD}$) and 232 measures of average $V$ magnitudes in quiescence ($V_q$). These are collected into a comprehensive catalog of most fundamental properties of all 402 known Galactic novae. The nova light curve and spectral classes are determined primarily by $M_{\rm WD}$. With an apparently clean cutoff, nova with light curve shapes in the S, P, O, and C classes have $>$0.95 $M_{\odot}$, while the J, D, and F class novae have $<$0.95 $M_{\odot}$. The speed class of the light curves is $t_3$=$10^{(-1.73M_{\rm WD})}$$\times$1900 days. The spectral class of novae is Fe II below 1.15 $M_{\odot}$, is He/N above 1.15 $M_{\odot}$, and the Hybrid novae are spread around this division. Neon novae have WD masses ranging from 0.53--1.37 $M_{\odot}$, with 76\% being measured to be below their minimum formation mass of 1.2 $M_{\odot}$, demonstrating that most are losing mass over each eruption cycle. The FWHM velocity of the Balmer line profiles is close to 0.23 times the WD escape velocity, or roughly $10^{(M_{\rm WD}/2)}$$\times$500 km s$^{-1}$ for $<$1.3 $M_{\odot}$. And all the known Galactic recurrent novae are $>$1.2 $M_{\odot}$. For issues involving the late expansion of the ejecta, I find that the visibility of shells is strongly biased towards novae with orbital periods $<$0.33 days, and that the visibility of $γ$-rays from the shells are strongly biased towards novae with fast declines, with $t_3$ a proxy for the $γ$-ray luminosity.

Figures (11)

• Figure 1: Shara method for measuring WD mass, with contours corrected for bias. The method of Shara et al. (2018) is to theoretically model nova light curves as a function of the WD mass ($M_{\rm WD}$) and the accretion rate, then calculating the decline rate ($t_2$), and the nova amplitude in the $V$-band ($A$). For any choice of $t_2$ and $A$, a unique mass is returned. The figure plot the position of each of 80 nova analyzed in the Shara paper, with each point having a calculated mass. I have drawn contours of equal-mass as a family of green curves, one curve for each labelled mass. So for other novae not in the original Shara sample, a user need only plot the nova based on the observed $t_2$ and $A$, then read off $M_{\rm WD}$ from the contour lines. This allows for the extension of the Shara model calculations to many additional novae. A bias has been found for the smaller Shara masses, with this corrected by Equation 1, as represented by the contours in this plot.
• Figure 2: The differences in $M_{\rm WD}$ between the original Shara masses and masses from other sources. This is a comparison test, where the original Shara masses vary systematically with $M_{\rm WD}$, following a linear trend (the green line). The original Shara masses are systematically too large for low mass novae. This is true for the three independent sets of WD masses, from the RV measures in the Ritter & Kolb (2003) catalog, the model WD masses from Hachisu & Kato, and the mixture of methods collected as 'Other'. The significant trend is the same for the three independent data sets, proving that the original Shara masses have a systematic bias. This bias can be corrected with Equation 1.
• Figure 3: Histogram of WD masses for SPOC and JDF light curve groups. The JDF light curve classes (the red shaded histogram) have a distinctly low-mass distribution, with an average of 0.81 $M_{\odot}$ and a 1-sigma range extending up to 0.95 $M_{\odot}$. The SPOC light curve classes (the blue shaded histogram) have a distinctly high-mass distribution, with an average of 1.10 $M_{\odot}$ and a 1-sigma range extending down to 0.95 $M_{\odot}$. The distribution is like that expected for a sharp cutoff at 0.97 $M_{\odot}$, where the known measurement errors of $\pm$0.15 $M_{\odot}$ create smearing that mixes up the two distributions around the cutoff.
• Figure 4: Decline rate versus WD mass. From Table 1, the decline rate ($t_3$) is plotted versus $M_{\rm WD}$ for 74 JDF novae (red diamonds) and for 117 SPOC novae (blue diamonds). We see a highly significant straight line, on this log-linear plot, with the best fit shown with the green line, with $t_3$ equaling $10^{(-1.73M_{\rm WD})}$$\times$1900 days. The typical measurement errors ($\pm$0.15 $M_{\odot}$ in mass and $\pm$40% in $t_3$) are displayed for two points. The plot has substantial scatter about the green line, but all the scatter is consistent with ordinary measurement errors, so the real underlying relation apparently has only small intrinsic scatter.
• Figure 5: The FWHM is a function of $M_{\rm WD}$. The 127 novae follow a highly significant trend. The green curve is for 0.23 times the escape velocity of the WD. Through most of the range, it is adequate to quantify the model as $10^{(M_{\rm WD}/2)}$$\times$500 km s$^{-1}$. The escape velocity has an increasing deviation as the Chandrasekhar mass is approached, with this being seen both in the escape-velocity model and in the nova data. Indeed, for masses $>$1.3 $M_{\odot}$, the observed upward deviations are substantially larger than predicted by the escape velocity alone. The scatter is large, but is consistent with the $\pm$0.15 $M_{\odot}$ for the real uncertainty in the masses and with the $\pm$0.22 for the real uncertainty in the logarithm of FWHM. Two typical error bars are presented for two of the points. That is, the observations look to closely follow a single function, while the known measurement uncertainties produce the observed scatter in the plot. With this, the real intrinsic relation is consistent with Equation 4 with no variation.