Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 364, L75-L79 (2000)

Previous Section Next Section Title Page Table of Contents

3. Discussion and conclusions

We have shown that the abnormally high luminosity of the short-period recurrent nova T Pyx can be explained if the system belongs to the class of wind-driven SSSs. T Pyx's accretion rate is then much higher than that of an ordinary CV, because the system's evolution is dominated by a radiation-induced wind from the secondary star. The wind-driven evolution is self-sustaining, because the wind-induced accretion rate is sufficient to power the luminosity that excites the wind.

Our model predicts that T Pyx is an intrinsically bright soft x-ray source. T Pyx was observed serendipitously as part of the ROSAT all-sky survey, but not detected (Verbunt et al. 1997). This is not yet in conflict with a SSS model for T Pyx, because of the limited sensitivity of these observations and the relatively high column density towards the system ([FORMULA] cm-2; Dickey & Lockman 1990 1). However, a sensitive, pointed observation with Chandra or XMM could yield a detection and provide a direct estimate of the WD luminosity and temperature. For example, assuming a distance of 3 kpc (e.g. Patterson et al. 1998), the flux produced by a blackbody with [FORMULA] ergs/s and [FORMULA] K is below the ROSAT detection limit, but would provide around 1000 counts in a 30 ksec observation with XMM.

If T Pyx is, in fact, a wind-driven SSS, its current evolutionary timescale is [FORMULA] yrs. Clearly, this is an extremely short-lived evolutionary state. So how will T Pyx's evolution actually end? We can think of at least three possible outcomes. First, the mass ratio of the system may be driven so low that irradiation and wind-driving cease to be effective. Second, the wind-driven evolution may proceed until the secondary completely evaporates, leaving T Pyx as an isolated WD. Third, if the total mass of the system is somewhat higher than we have assumed, the WD may reach the Chandrasekhar limit and explode in a Type Ia supernova.

The likelihood of the last outcome depends on whether the amount of material that is ejected in each of T Pyx's recurrent nova outbursts is significantly less than that which is accreted during each quiescent interval. Theoretically, this is plausible (e.g. Prialnik & Kovetz 1995; Livio & Truran 1992). The question might also be addressed observationally: the total mass accreted during quiescence is [FORMULA]. The mass ejected during an outburst can be inferred by comparing the pre- and post-outburst orbital periods, although this is sensitive to the details of the ejection process (Livio 1991; Livio et al. 1991). Such a comparison could also distinguish between wind-driven and nova-driven evolution scenarios. As noted in Sect. 2.2, if the specific angular momentum of the mass ejected during T Pyx's nova eruptions is very high ([FORMULA]), the angular momentum loss associated with nova eruptions could, in principle, itself drive T Pyx's high mass transfer rate. In the context of this model, the difference between pre- and post-outburst orbital periods must be relatively large. By contrast, the wind-driving model makes no such prediction.

Given the strong possibility that the system will destroy itself on a short timescale, T Pyx represents an evolutionary channel by which short-period CVs may be removed from the general CV population. The existence of such a channel is desirable from a theoretical point of view. As most recently pointed out by Patterson (1998), the observed CV population shows a significant dearth of short-period systems relative to the predictions of standard evolutionary scenarios. Whether T Pyx's "assisted stellar suicide" represents a sufficiently wide channel in this context is an important question: do most CVs eventually enter a wind-driven phase or is T Pyx really a unique system? The answer clearly depends on how wind-driving is actually triggered, an issue we have not addressed in the present work. One possibility is that triggering occurs during a period of residual nuclear burning in the aftermath of a nova eruption, though this would have to be limited to systems satisfying other constraints (e.g. low q and/or high [FORMULA]).

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001