Astron. Astrophys. 347, L43-L46 (1999)

3. Discussion

3.1. Supersoft X-ray emission

U Sco belongs now to those novae for which SSS X-ray emission has been discovered (cf. Orio & Greiner 1999).

The TR theory predicts two processes which can generate soft X-rays: Shock acceleration in the nova ejecta and steady nuclear burning. For Nova Cyg 1992 an optically thin component due to shocks has been detected. In addition an optically thick SSS X-ray component has been observed 60 days after the outburst and for 600 days (Krautter et al. 1996; Balman et al. 1998).

According to the calculations of Kato (1996) performed for U Sco, and assuming a massive (=) WD a SSS component is predicted to be observed from 10 days after the outburst. In the case of the H-rich model (He/H=0.1) the supersoft component is expected to rise till 50 days after the outburst to a maximum luminosity of . In the He-rich model (He/H=2), a maximum luminosity for the SSS component of is reached 20 days after the optical outburst. Using the He enriched fit with the N/C ratio enhanced (Table 1), the observed bolometric luminosity 19-20 days after the optical outburst is . Assuming a distance 14 kpc (see Introduction) a bolometric luminosity of is derived. The luminosity is in agreement with the bolometric luminosity of predicted for novae (Mc Donald et al. 1985). The temperature of and the luminosity of derived from the X-ray spectral fit requires a very massive WD consistent with an almost CH mass WD (e.g. Kato 1997).

Table 1. Best-fit values derived from spectral fits to the BeppoSAX LECS and MECS spectrum of U Sco using (a) a blackbody model with absorption edges and (b) an optically thick non-LTE WD atmosphere model (with He enriched and the N/C ratio enhanced) and an optically thin Raymond and Smith component (assuming He enriched and the ratio N/C enhanced). 90% confidence parameter ranges are given. For the edges the absorption depth at the given energies are listed

3.2. Spectrally hard component

In addition to the optically thick SSS X-ray model spectrum, the spectral fits require a spectrally hard component. A similar component in addition to a SSS component was used by Balman et al. (1998) for X-ray spectral fits to the classical nova Cyg 1992. Using an optically thin thermal model we derive a temperature of , an emission measure if He is enriched and the ratio N/C enhanced.

If we assume a terminal wind velocity of the wind mass loss rate for a He/H=2 mixture can be estimated from . Here r is the typical radius of the emitting region. We assume , the radius of the Roche-lobe, and use the result of the spectral fit assuming He is enriched. We then obtain a wind mass loss rate of . For a distance to U Sco of 14 kpc, we derive a wind mass loss rate of . A near-CH mass WD at experiences an envelope mass loss of due to both steady nuclear burning and a wind from the WD. The steady nuclear burning mass loss can be estimated to be (Hachisu et al. 1999). The mass loss due to the wind is . This value is in agreement with the range we derived above. For a duration of the steady nuclear burning phase plus wind mass loss phase of 0.1 year (Kato 1996) we derive a mass loss from the WD envelope of of which is due to the wind. In addition the predicted post-outburst envelope mass is (Hachisu et al. 1999). This would mean that 70% of the envelope mass has remained on the WD allowing it to increase in mass. Williams et al. (1981) derive from the UV lines (for 14 kpc and ) a wind mass loss rate which differs significantly from our value, although it is subject to many uncertainties, and differences between outbursts cannot be accounted for.

If the helium fraction is indeed large () only part of the accreted He/H envelope might have been ejected and steady nuclear burning proceeded for at least one month. This result is consistent with the analytical model of Kahabka (1995). Assuming an X-ray on-time of 0.1 years and a recurrence period of 10 years, we constrain the WD mass to . U Sco and RN in general are therefore probably SN Ia progenitors (cf. Li & van den Heuvel 1997; for a recent review on SN Ia, see Livio 1999).

© European Southern Observatory (ESO) 1999

Online publication: June 6, 1999
helpdesk.link@springer.de