ROSAT Basic Properties

A breakthrough came when ROSAT started its observation in 1990. With the position sensitive proportional counter (PSPC), which had an energy range 0.1-2.4 keV, X-ray data of much higher sensitivity could be obtained than from any comparable X-ray telescope/detector before. Even if the energy resolution of the PSPC was poor, some spectral information could be obtained and many basic X-ray properties of novae could be detected. Several observational programs mark the process in our understanding of X-ray properties of novae.

The most fundamental results were obtained from the observations of V1974 Cyg (1992), a moderately fast (t3 ^ 35d) ONeMg nova, which was discovered in outburst on February 20. With a maximum brightness Vmax ^ 4.4 mag it was the brightest nova since V1500 Cyg (1975). Observations with the PSPC onboard of ROSAT started on April 20, 63 days after outburst, as soon as the nova had entered the ROSAT observing window (Krautter et al. [11]). Subsequently, V1974 Cyg was observed by Krautter et al. [12] with ROSAT in the 0.1-2.4 keV energy range over a period of nearly two years on a total of 18 occasions.

Three phases can be distinguished in the lightcurve presented in Fig. 13.1: From days 63 to 147 during an initial rise phase the count rates are low, however, the count rate is strongly increasing from 0.03 to 0.37 counts s-1. Three months later,

i in

11.0

0.01

Fig. 13.1 X-ray lightcurve log counts s 1 vs. days after outburst of V1974 Cyg [12]

i in

11.0

1

i

i

0

1 1 :

-

%

r

i

-

1

-

# :

-

$

$

i

J

i i

i

i

Fig. 13.1 X-ray lightcurve log counts s 1 vs. days after outburst of V1974 Cyg [12]

200 400 600

200 400 600

on day 255, the beginning of the second phase, the "plateau phase," a big surprise came when the count rate had enormously increased to 11.3 counts s^1. The count rate increased further to a peak of about 76 counts s^1 on day 434 and remained essentially the same on the next observation on day 511. The last phase, the decline phase, starts between days 511 and 612 and is characterized by a strong decline of the count rate down to 0.2 counts s^1 on day 653. Because of the large gaps of about 3 months between the observing windows, there is an uncertainty in the beginning and the end of the individual phases.

The spectral properties are reflected in the three phases of V1974 Cyg. SED characteristics for each of the three phases is presented in Fig. 13.2 [2]. During the rise phase, the SEDs are hard with essentially no photons below 0.7 keV with the hardest SED found on day 63. Subsequently, the SED got softer, the first indications

0. 001

1 ,

1

Energy/keY

;+ +

1+ i

1

Energy/keV

Energy/keY

Energy/keY

Energy/Key

3 OJ

Energy/Key

Energy/keY

Energy/KeU

Fig. 13.2 Spectral energy distributions of V1974 Cyg in the ROSAT PSPC band on three different epochs. The actual PSPC data are indicated with crosses. The data are fitted with O-Ne-enhanced WD atmosphere models (soft component) and Raymond-Smith plasma emission models (hard components). The lower figures show the residuals between the data and the models in standard deviations [2]

for a soft component showed up on day 97. A dramatic change occurred between days 147 and 255, since on the latter date a strong soft component had appeared. The SED exhibited now, like for all other spectra obtained during the plateau and the decline phase, the characteristics of a super-soft source (SSS) (e.g., [9]). With a maximum count rate of 76.5 counts s"1 on day 511 V1974 Cyg was by far the strongest SSS ever observed by ROSAT. The hard component that had shown up in the early spectra was present during the whole plateau phase. During the decline phase, the soft component decreased strongly and the hard component became more prominent again in relation to the soft component.

The origin and temporal behavior of the soft component can be best interpreted in terms of the TNR model of the nova outburst. As Krautter et al. [12] had shown, fits with blackbody energy distributions did not give any reasonable results. This result can be generalized to all other SSS [9]. Balman et al. [2] analyzed the ROSAT data with LTE white dwarf model atmospheres with O-Ne overabundances by MacDonald and Vennes [17]. They could show that, as predicted by the TNR model, during the plateau phase Teff increased at constant luminosity from ~4x 105 K up to to a peak temperature of ~6x 105 K, while the white dwarf radius decreased by about a factor of 2. During the decline phase, both L and Teff decreased indicating that hydrogen burning on top of the white dwarf had turned off. Using a cooling timescale of about 6 months Krautter et al. [12] estimated a mass of ~10"5 M0 for the hydrogen-exhausted, remnant envelope of the white dwarf.

Balman, Krautter and Ogelman also carried out a detailed analysis of the hard component applying Raymond-Smith thermal plasma emission models. Although the plasma temperature decreased from an initial ^5-10 keV to a more or less constant level around keV, the flux increased to a maximum around day 150 with a subsequent decline. The peak unabsorbed flux corresponded to a luminosity of (0.82.0) x 1034 erg s"1 at a distance of 2-3 kpc. The temporal evolution and the plasma temperature suggest a shock origin of the hard X-ray flux. Balman, Krautter, and (Ogelman favor a model in which a fast wind from the nova collides with either a preexisting gaseous shell or material ejected at the initial phase of the nova.

The reason for the absence of any soft X-ray emission during the early phases is the initial high hydrogen density NH which decreased by about a factor of 10, until it reached a constant level around day 255. Lloyd et al. [16] found a similar result for the fast nova V838 Her (1991), which was observed 5 days after visual maximum. Like V1974 Cyg only a hard component with no photons below 0.7 keV was found at a count rate of 0.16 counts s"1.

The second milestone was GQ Mus, the first nova of which X-rays were discovered in outburst in 1984. GQ Mus was also detected by ROSAT, 6 years after its last EXOSAT observation and 8.5 years after maximum (day 3118) at a count rate of 0.143±0.006 counts s"1 in the ROSAT All-sky Survey (RASS). At subsequent observations carried out in the pointed mode the count rate dropped from 0.127 counts s"1 on day 3322 [25] via 0.007 counts s"1 on day 3871 to <0.003 (3a) counts s"1 [31]. Balman and Krautter [1] showed that the hydrogen burning in GQ Mus had most likely ceased already on day 3322. With a duration between 8.5

and 9.1 years GQ Mus exhibited an extraordinarily long phase of hydrogen burning, the longest known so far. The much longer turn-off timescale of GQ Mus, when compared with V1974 Cyg, indicates that the mass of the white dwarf in GQ Mus is lower than that one in V1974 Cyg. GQ Mus had also some unsusual properties in the optical spectral range [10]. For many years its emission line spectrum exhibited a very high ionization with the coronal lines [FeXI] A7892 as strong and [FeX] A6374 about 2.5 as strong as Ha, respectively, which has not been observed - at least to the knowledge of the author - so far in any other astronomical object. As Krautter and Williams showed, the high ionization lines were due to photoionization by a hot (several 105 K) underlying source. In a spectrum taken on March 10, 1993, after the turn-off of the hydrogen burning, the coronal lines had disappeared and the ionization was much lower. This shows that the ionization stage of the optical spectrum could be used generally as a qualitative indicator for the turn-off of the hydrogen burning.

To get more information on the duration of the phase of constant luninosity, Orio et al. [26] did a systematic search for supersoft-sources in the ROSAT archive. They analyzed 350 pointed and serendipitous observations of 108 different classical and recurrent novae in outburst and in quiescence. They found only three novae with a super-soft spectrum, the already mentioned V1974 Cyg and GQ Mus and, in addition, N LMC 1995. For a time of up to 10 years after explosion, 30 galactic and 9 LMC novae were in the ROSAT sample. In the post-ROSAT phase a few more novae with soft X-ray emission were detected by recent X-ray satellites (see later), but it is clear that the vast majority of all postnovae were not observed as soft X-ray sources. One has to ask whether this missing soft X-ray emission is real, i.e., novae switch off hydrogen burning after a relatively short time, or whether it is due to a selection effect. A plausible selection effect could be that the interstellar hydrogen column density is so high that the soft X-ray radiation gets absorbed. However, as Nickel [21] showed, 21 novae, which were observed in the ROSAT All-sky Survey within 10 years after the outburst and for which extinction data were available, have an average extinction E#-y ~ 0.5 mag. With this value, only for a few novae the hydrogen column densities should be so high that all the soft X-ray radiation gets absorbed. Even more striking is that with one exception no LMC nova was found where the average extinction EB-y is below 0.2 mag. Circumstellar extinction should not play any role, since most novae were observed months or even years after maximum when the expanding envelope should have become transparent for soft X-ray radiation. On the basis of these data, a selection effect can be safely excluded. One can conclude that the majority of all novae switch off after a relatively short time. Of course, a precise number cannot be given, but it seems safe to conclude that for most novae hydrogen burning turns off after less than 2 years.

The turn-off time scale as function of the white dwarf mass was calculated by Starrfield et al. [35] who assumed that the accreted envelope is ejected by a radiation driven wind. They find a strong inverse dependence of the turn-off time scale from the white dwarf mass. For instance, a nova with a 1.00 M© white dwarf should turn-off after about 100 years whereas a 1.25 M© white dwarf has a much shorter turn-off time of about two years. The short turn-off timescales found from the X-ray observations is a clear evidence that most novae have white dwarfs with masses well above —1.0-1.1 M0. Such a result had been predicted by the TNR model, since according to this model the critical mass that is needed to ignite the thermonuclear reactions and to start the runaway decreases strongly with increasing white dwarf mass [33]. Statistically, novae with high mass white dwarfs should be found much more frequently than novae with white dwarfs with lower masses.

Was this article helpful?

0 0

Post a comment