2. IUE observations of post-novae

Our sample consists of classical novae which have been observed with IUE at least one year after the outburst. From this work we exclude GQ Mus and V1974 Cyg, which have extensively studied by other authors, and whose turn-off times have been derived from ROSAT PSPC data. In the following context we refer to the UV flux or luminosity in the IUE "short'' wavelength range, i.e. the integrated flux between 1230 and 1950 Å. The lower wavelength limit has been chosen to avoid the strong geocoronal Lyman emission. Table 1 presents some of the most relevant characteristics of the objects included in our sample. In Table 2 we list the IUE images used in this work, the exposure time, the date of observation and the corresponding day after outburst. We also list in this table the UV luminosities (or the corresponding upper limits) measured in each spectrum, computed with the values of distance and reddening listed in Table 1. The spectra of all galactic novae have been dereddened with a standard UV extinction curve (Savage & Mathis 1979). For the LMC novae we have taken a distance to the LMC of 50 kpc (Panagia et al. 1991), and as interstellar reddening, we have assumed in all cases two components, one Galactic with E(B-V)=0.05, and a LMC internal reddening of E(B-V)=0.10 with a non-30 Doradus extinction law (Fitzpatrick 1986). We include both the total luminosity in the 1230- 1950 Å range, and the continuum luminosity, obtained subtracting the luminosity in the emission lines from the total luminosity. Figures 1 and 2 show the spectra used in this work.

Since most of these novae were only observed sporadically in X-rays or in the UV range, we cannot systematically monitor their entire decline phases. We therefore have made an operational definition of the UV turn-off time as the time it takes for the remnant to drop below erg s^-1, and for the X-rays turn-off the time when the blackbody temperature of the stellar remnant declines to less than 20 eV (Orio et al. 1997). These definitions are consistent with the lowest luminosity and black-body temperature predicted in the models for a hydrogen burning post-nova white dwarf, and with both the simulations (Starrfield, Prialnik, private communications) and the X-ray observations of nova turn-offs. In addition, the UV definition agrees with the luminosities of in old novae where nuclear burning is known to be absent (e.g. Selvelli et al. 1989). Lower luminosities imply either that the nova has turned off or that the rapid cooling due to exhaustion of hydrogen is likely to be occurring.

 
Figure 1: Observed IUE spectra of the objects in the sample

 
Figure 2: Observed IUE spectra of the objects in the sample (continued)

 

Object Date of visual E(B-V) distance t₃  

   maximum (kpc)   (days)

V1668 Cyg 1978 14/09/78 [1] 0.40 [1] 3.6 [1] 30 [1]

PW Vul 1984 04/08/84 [2] 0.55 [2] 1.3 [2] 147 [2]

QU Vul 1984 24/12/84 [3] 0.61 [3] 3.5 [3] 31 [4]

V842 Cen 1986 24/11/86 [5] 0.55 [5] .92 [5] 48 [5]

OS And 1986 07/12/86 [6] 0.25 [6] 5.1 [6] 20 [6]

QV Vul 1987 15/11/87 [7] 0.32 [7] 4.5 [7] 60 [7]

V433 Sct 1989 13/09/89 [8] 0.41 [9] 8.0 [9] 46 [9]

V838 Her 1991 23/03/91 [10] 0.53 [10] 3.4 [10] 3 [10]

V351 Pup 1991 27/12/91 [11] 0.30 [11] 4.7 [11] 40 [11]

V705 Cas 1993 07/12/93 [12] 0.50 [12] 3.2 [12] 100:[12]

LMC 1988 No. 1 23/03/88 [13] 0.05+0.10 50 [14] 33 [13]

LMC 1988 No. 2 14/12/88 [13] 0.05+0.10 50 [14] 10 [13]

LMC 1991 25/04/91 [15] 0.05+0.10 50 [14] 6 [15]

References:
(1) Stickland et al. 1981, (2) Andreae et al. 1991, (3) Saizar et al. 1992, (4) Rosino et al. 1992, (5) Sekiguchi et al. 1989, (6) Schwarz et al. 1997, (7) Gehrz et al. 1992, (8) Rosino et al. 1991, (9) Anuapama et al. 1992, (10) Lynch et al. 1992, (11) Orio et al. 1996, (12) Hauschildt et al. 1994, (13) Cappacioli et al. 1990, (14) Panagia et al. 1991, (15) Della Valle 1991.

 

Object Image Date of Texp Day after Comments

observation (min) outburst (10³⁴ erg s^-1)

V1668 Cyg SWP07621 09/01/80 240 482 7.1 2.8 Weak continuum. Lines

SWP09065 23/05/80 395 615 5.4 3.3 Weak continuum. Lines

SWP10886 24/12/80 360 840 < 1.6 No continuum. Marginal lines

PW Vul SWP28068 31/03/86 70 604 9.5 6.1 Continnum and lines

SWP28461 09/06/86 70 628 5.6 3.0 Continuum and lines

SWP33803 23/06/88 423 1428 1.7 1.7 Continuum. No lines. Wrong star?

QU Vul SWP33794 21/06/88 303 1277 23 15 Weak continuum. Lines

SWP36933 03/09/89 960 1727 9.5 6.6 Weak continuum. Lines

V842 Cen SWP38684 27/04/90 95 1249 5.8 5.2 Continuum and lines

SWP41667 20/05/91 230 1639 3.8 3.5 Continuum and lines

SWP42122 24/07/91 790 1705 3.0 2.8 Continuum and lines

OS And SWP32336 16/11/87 105 344 14 5.2 Continuum and lines

SWP38031 15/01/90 290 1138 2.2 2.2 Weak continuum. No lines

SWP42292 19/08/91 372 1719 2.6 2.6 Weak continuum. No lines

QV Vul SWP41672 21/05/91 304 1284 < 1.1 No detection

V443 Sct SWP39495 19/08/90 300 344 11 7.4 Weak continuum. Lines

SWP44219 24/03/92 360 927 <4.6 No detection

V838 Her SWP42119 24/07/91 395 123 11 9.2 Continuum and lines

SWP48029 02/07/93 428 837 < 1.8 No detection

V351 Pup SWP49921 31/01/94 470 765 14 12 Continnum and lines

SWP52876 22/11/94 405 1062 4.8 4.2 Continuum and lines

V705 Cas SWP55975 21/09/95 400 653 14 6.0 Continuum and lines

LMC 88 No. 1 SWP36218 07/05/89 865 413 53 46 Weak continuum. Lines

LMC 88 No. 2 SWP36615 05/07/89 410 266 390 385 Continuum and lines

SWP40135 18/11/90 830 768 < 4.8 No detection. Other star in aperture

LMC 91 SWP44230 25/03/92 410 343 145 35 Weak continuum. Lines

2.1. The Galactic sample

2.1.1. V1668 Cyg (Nova Cyg 1978)

In January and May 1980 a weak continuum and emission lines of NIII], NIV], CIII] and CIV were unambiguously present in the IUE spectrum. The UV continuum luminosity at that time was . The nova was not detected in the last IUE observation, in December 1980. Some very weak emission lines could be present (NV 1240 and CIV 1550), but they could well be camera artifacts (Crenshaw et al. 1990). The turn-off time is therefore < 1.3 years.

2.1.2. PW Vul (Nova Vul 1984 No. 1)

The IUE spectra obtained in March and June 1986 showed both UV lines and continuum clearly present. Two years later there was a faint continuum, with no emission lines, and the luminosity in the 1230-1950 Å band was . The possibility that this spectrum corresponds to a nearby star cannot be totally excluded, since there are a few faint stars close to the nova (see e.g. Ringwald & Naylor 1996). Assuming that the observed spectrum corresponds to the nova, we see the signature of the turn-off that started already in June 1986, so the turn-off time is less than 2 years. In agreement with this, the ROSAT upper limits for this nova in 1991 and 1992 imply .

2.1.3. QU Vul (Nova Vul 1984 No. 2)

This nova was detected by ROSAT 6.5 years after the outburst, with a flux much lower than immediately after the outburst (Orio 1993). The X-ray flux did not seem to be due to the hot central source. The September 1989 IUE spectrum showed a very faint continuum with some emission lines, some of which could be just camera artifacts due to the very long exposure. The only lines which seem real are HeII 1640 Å and NeIV] 1602 Å. The UV continuum luminosity was still high, . Given the upper limit provided by the X-ray data, we conclude that the turn-off was about 5.5 ( 1) years after the outburst.

2.1.4. V842 Cen (Nova Cen 1986 No. 2)

At the time of the last IUE observations (May and July 1991) this nova still showed strong emission lines of CII, SiIV, NIV], CIV, HeII and CIII]. The UV continuum luminosity in July 1991 was , and between the May and July 1991 observations the UV energy distribution became flatter, consistently with cooling of the central source. This nova was not detected during a very short exposure in the ROSAT survey (Orio et al. 1992a), implying an upper limit to the soft X-ray luminosity of approximately . This upper limit does not exclude continued nuclear burning, but since the UV luminosity was low in 1991, we conclude that the turn-off time was approximately 3.5 years.

2.1.5. OS And (Nova And 1986)

The last IUE observation of this object was taken in August 1991. A flat faint continuum was detected without any evidence of lines. Emissions were already absent in January 1990. The August 1991 UV continuum luminosity was . OS And was not detected in a ROSAT pointed observation about a year later (Orio et al. 1992a), indicating a black body temperature of the central source below 20 eV (Orio 1993; Ögelman & Orio 1995). This source appears to have shut down at the time of the last IUE observation, or even before, since the UV continuum luminosity in November 1987 was only . We find therefore an upper limit of 1 year for the turn-off time.

2.1.6. QV Vul (Nova Vul 1987)

This nova was poorly observed by IUE during the outburst. It was not detected 3.5 years later in a five hours exposure. The upper limit to the luminosity in the SWP band was . Since also ROSAT did not detect the nova after less than 4 years after the outburst (Orio 1993; Ögelman & Orio 1995), we find that the turn-off occurred in less than 3.5 years.

2.1.7. V443 Sct (Nova Sct 1989)

V443 Sct was sparsely observed by IUE during the first stages of the outburst. In August 1990 there was no continuum, but only marginal emission lines of NIV], CIV, HeII and CIII]. In March 1992 it was not detected in a six hours exposure, implying an upper limit to the UV luminosity of . The temperature of the ionizing source was estimated to be (Rosino et al. 1991; Anupama et al. 1992) at day 250. If the source had maintained this blackbody temperature it should have been detected by IUE, while ROSAT did not detect it at approximately the same time of the last IUE observation (Orio et al. 1997). A turn-off time of less than 2.5 years is a realistic estimate.

2.1.8. V838 Her (Nova Her 1991)

This was a very fast nova with t₃=3-5 days (Lynch et al. 1992). It showed X-rays emission only a week after the outburst, possibly due to shocked material in the shell (Lloyd et al. 1992). Leibowitz et al. (1992) and Leibowitz (1993) claim that the disk was not disrupted during the outburst. It was not detected by ROSAT a year after the outburst (Ögelman & Orio 1995), but 7 months later weak hard X-ray emission was detected in long exposure (Szkody & Hoard 1994). This hard spectrum X-ray emission could be due to an accretion disk or to ongoing shocks in the ejected shell. This nova was poorly observed by IUE in the first stages of the outburst. It still showed a rather strong continuum with some broad emission lines in July 1991. It was not detected in July 1993, implying an upper limit to the UV luminosity of . Therefore, we estimate a turn-off time of less than two years.

2.1.9. V351 Pup (Nova Pup 1991)

X-ray emission from this nova was detected 16 months after the outburst. This emission was too hard to be associated with an X-ray source, and it was rather due either to shocks or to accretion (Orio et al. 1996). The comparison of optical spectra at different epochs led to an estimate of two years for the turn-off time. This is confirmed by our measurements of the UV continuum luminosity, which shows that the central source had already turned off at the time of the last IUE observations, 3 years after the outburst.

2.1.10. V705 Cas (Nova Cas 1993)

This nova has been observed by IUE during the two years following the outburst. The last spectrum still shows prominent emission lines and a weak continuum. The derived UV continuum luminosity shows that it was close to turn-off at the time of the last IUE observation, 1.8 years after the outburst.

2.2. The LMC sample

Novae in the LMC have been monitored by IUE only since 1988. This restricted sample is very meaningful because the different objects can be compared without uncertainties in distances and reddening. As stated above, we have used the same distance and reddening correction for all the objects. We include here only the novae which have been observed at least for approximately one year after the outburst i.e., LMC 1988 No. 1, LMC 1988 No. 2 and LMC 1991. Nova LMC 1990 No. 1 and Nova LMC 1992 were observed only for a few months, and Nova LMC 1990 No. 2 was a recurrent nova (Shore et al. 1991).

2.2.1. Nova LMC 1988 No. 1

The spectrum taken in May 1989 showed only a weak continuum with evidence of emissions of NIV] 1486 Å, CIV 1550 Å and CIII] 1910 Å. Emissions near 1663 and
1750 Å are most likely camera artifacts. The UV continuum luminosity in May 1989 was , and by that time the nova had clearly not turned off. However, it was not detected in ROSAT observations in 1992 and 1994, implying a turnoff time of less than  years (Ögelman & Orio 1995).

2.2.2. Nova LMC 1988 No. 2

The IUE spectrum taken in November 1990 showed a rather strong flat continuum without any emission line. The only spectral features are a few absorption lines of interstellar origin. A careful check of the spacecraft pointing has shown that this spectrum, which is clearly off-centered in the aperture does not correspond to the nova, but to a closeby star. The nova itself should have been well centered in the aperture, and it is not detected. The previous spectrum (July 1989) also shows two stars in the aperture, one of them with a very clear emission at
1240 Å (NV), so we conclude that it was the nova itself. For the values of distance and reddening quoted above, the continuum luminosity at the time of the July 1989 observation was and the turn off time is more than 1 year. The upper limit for the UV continuum luminosity derived from the November 1990 observation is , and therefore the nova had switched off by then. The ROSAT observations between 1991 and 1994 imply a turn-off time < 5 years and most likely even < 3 years (Ögelman & Orio 1995).

2.2.3. Nova LMC 1991

The IUE spectrum taken in March 1992 showed a weak continuum and very broad emission lines of NV, NIV], CIV, CIII] and possibly NIII]. The continuum luminosity in the IUE short wavelength range at the time of the last observation was , and therefore no turn off has been observed.