Novae a theoretical and observational study

In this thesis, we present studies relating to novae that include both theoretical and ob- servational aspects. Being hosted by accreting white dwarfs (WDs), they have drawn attention in the context of the supernova Ia (SN Ia) progenitor problem. In the case of the nova explosion, the WD host is not disrupted. Instead, it continues to supply energy, even after the optical outbust, via stable nuclear burning of the remnant hydrogen envelope that survived the outburst. Accordingly, nova emission progresses toward the harder part of the electromagnetic spectrum, where it lasts longer than in the optical regime. As a consequence, novae are found to constitute the majority of the observed supersoft X-ray sources (SSSs). This is particularly well established for the galaxy M31.

For high mass accretion rates in the unstable nuclear burning regime (or nova regime), there is evidence that significant mass accumulation by the WD is possible. This paved the way for SN Ia progenitor models in the single degenerate (SD) scenario involving novae. Based on the statistics of novae in M31, which is the most frequently used target for nova surveys, we investigate the role that novae may play in producing SNe Ia. Using multicycle nova evolution models and the observationally inferred nova rate in M31, we estimate the maximal SN Ia rate that novae can produce, assuming that all of the involved WDs reach the Chandrasekhar mass. Comparing this rate to the observationally inferred SN Ia rate for M31 constrains the contribution of the nova channel to the SN Ia rate to 2-7%. Additionally, we demonstrate that a more powerful diagnostic can be obtained from statistics of fast novae, which are characterized by decline times t2 10 days. Most novae resulting from a typical SD SN Ia progenitor accreting in the nova regime are fast. Specifically, as the WD in the nova grows in mass, it produces novae more frequently and with decreasing decline times. We therefore investigate how efficiently fast novae can be detected in nova surveys of M31 by a PTF (Palomar Transient Factory) class telescope, with the aim to formulate requirements for future high-cadence surveys directed toward fast novae. We find that a survey with a limiting magnitude of mR ≈ 22, detects ≈ 90% of fast novae expected in the SD scenario, with observations at least every second night. If f is the fraction of SNe Ia accreting in the nova regime for accumulation of the final tenth of a solar mass preceding the SN Ia explosion, then such surveys should be detecting fast novae in M31 on the order of 1000 × f per year.

The population of fast novae, whose importance has been highlighted above, is to date unconstrained for any galaxy. Most recent nova rate measurement in M31 is also devoid of them. However, for a first approximation, we make use of an older survey of M31 by Arp (1956), to probe the high-mass end of the WD population involved in novae. In the context of the SD scenario, it is possible for a SN Ia progenitor to pass through different nuclear burning regimes (stable, unstable, and wind regimes). Comparing the observed WD mass distribution in novae with that predicted for the SD scenario, we obtain constraints on the fraction of mass that can be accumulated in the nova regime by the SN Ia progenitor. For low WD masses ( 1.30 M⊙), significant mass accumulation, at the level of 10-60%, is possible in the nova regime. For the more massive WDs, the constraints tighten to ∼ 2% for mass accumulation in this regime.

Given the longer-lasting electromagnetic signature of the nova explosion in the soft X-ray regime, we venture into exploring the population of post-nova SSSs in M31. To this end, we make use of theoretical post-nova evolution models. Depending on the assumed WD mass distribution in novae, we obtain at any instant 250-600 post-nova SSSs with a luminosity of Lx ≥ 10^36 erg/s at photon energies of 0.2-1.0 keV, ignoring absorption. Their combined luminostiy is of the order of 10^39 erg/s. We also derive the distributions of their luminosities and effective temperatures. The former exhibits a significant steepening at log(Lx) ∼ 37.7-38 with a cut-off at ≈ 2×10^38 erg/s. The latter distribution is roughly a power law with logarithmic differential slope 4-6 up to the maximum effective temperature of ≈ 1.5 × 10^6 K. We compare our predictions with the observational results of the XMM- Newton monitoring of the central region of M31, obtaining a good agreement.

To characterize the population of fast novae in M31, we undertake an analysis of the M31 data from the intermediate Palomar Transient Factory (iPTF) Survey. In this pursuit, we develop a new method for efficiently detecting candidates for transients and variable sources from the outputs of time-domain data pipelines, significantly reducing the contamination by artifacts. The method is particularly relevant for handling regions that are crowded by a large number of false detections from the data pipelines, such as the bulges of galaxies and the Galactic plane.

To illustrate the methodology, we use iPTF M31 observations, with a baseline of five months, covering the bulge. The iPTF pipeline produces ∼ 10^5 detections from the dif- ference images of these observations, including multiple detections from the same source. Projecting these raw detections onto a blank image of the analyzed field, we obtain an analogue of an X-ray image. In this image, the candidates for variable sources appear as clusters of raw detections on top of a background of artifacts. We then make use of a wavelet-based tool called WAVDETECT, developed for the analysis of Chandra data, to identify the unique candidates for variable sources. We apply our method to conduct a systematic search for novae in M31. To this end, we construct lightcurves of all candidates and apply a lightcurve selection algorithm based on the expected properties of novae. We find eight nova candidates, which we have verified to be confirmed novae in all cases.

The implementation of our method, including the lightcurve selection, is automated, which makes it possible to quantify the sample completeness. This will be done in the future, following the analysis of a ∼ 10-fold larger iPTF data set. Our method will thus facilitate the search for novae, in particular fast ones, including a completeness estimate. The derived rates will enable us to put more stringent constraints on the role novae play in the context of SN Ia progenitors, and to obtain a more reliable WD mass distribution

in novae at the massive end.