CNR-GNA-Osservatorio Astronomico, Via Tiepolo 11, Trieste, Italy
Institut d'Astrophysique, Paris, France
CNR-Istituto di Astrofisica Spaziale, Frascati, Italy
Keywords: old nova, winds
V603 Aql is an old nova, whose outburst was seen in 1918. The outburst properties were those of a ``fast'' classical nova with t 8, and ejection velocities of approximately -1800 km s for the principal shell. Its orbital elements (Ritter 1990, Warner 1976) suggest component masses near to 0.8 and 0.3 for the white dwarf and the cool companion, respectively, and a system inclination close to 17. The present observations with the HST-GHRS had two aims: one was to try to detect any absorption lines due to the material ejected during the nova outburst, the other was to obtain high quality line profiles in order to investigate the properties of the wind emanating from the inner disk region in CVs accreting at high rates. Either study could be undertaken by observing the spectral regions of the resonance lines of CIV, SiIV, and NV.
The spectra were taken on February 19 and 20, 1994, using the GHRS-G160M first-order grating (resolution of 2.5 10) in the accumulation mode with the FP-SPLIT procedure (Soderblom et al. 1995), in which each exposure is broken up into a number of equal sub-exposures (default = 4) at slightly different grating positions in the dispersion direction. After proper realignment in wavelength the sub-exposures were co-added, leading to a significant reduction in fixed pattern noise. A separate examination of the individual sub-exposures has allowed a time resolved study of the various spectral features, at the price of a non-optimum signal-to-noise ratio.
In view of the probable non-deceleration of the nebula (Duerbeck 1994), the absorption due to the ejected envelope is expected at a velocity close to that of the ``principal absorption system'' observed during outburst, which contains most of the ejected mass (Friedjung 1987). Therefore, we have searched for sharp absorption components in the resonance lines at a velocity near to -1800 km s, but no such lines were seen in the co-added spectra. Assuming half the local noise range as an upper limit for the absorption line central depth, we have derived upper limits on the order of 0.05 for the optical thickness, and on the order of 1.1-- cm for the column densities. From these values, taking the nova abundance and ionization fraction determinations by Ändrea et al. (1994) and assuming spherical symmetry in the ejecta, we have obtained an indicative value of as upper limit for the envelope mass. This value should be substantially modified, however, if the ejected shell is not spherically symmetric and/or has a clumpy structure.
The CIV, SiIV and NV resonance profiles are substantially different from each other. The merged spectrum of the CIV doublet (see Figure 1)
Figure: The merged spectrum of the CIV 1550Å doublet. The sharp feature near 1526.5Å is the interstellar SiII(2) 1526.70Å line.
shows a P Cyg profile, in which the strong emission blend (FWHM=750 km s, HWZI=1000 km s, after subtraction of the doublet separation) is almost symmetrical around 1549.2Å, while the flat and weak absorption blend has edge velocity close to -2500 km s.
The merged spectrum of SiIV shows two symmetric emissions having FWHM=660 km s, HWZI=750 km s, and the emission center violet shifted by about 100 km s. In the case of the NV 1240Å doublet, the merged spectrum shows a blue-shifted absorption profile for each doublet component with edge velocity close to -1800 km s.
Certainly the most striking property of the spectra is their rapid variations, particularly for the absorption components. For instance, the NV absorption is very strong on the first sub-exposure and rather weak on the second, while sub-exposure 4 shows an almost featureless continuum (see Figure 2).
Figure: The individual subexposures of the NV 1240Å doublet. For clarity, the spectra have been offset in the y-direction by 1.510 erg cm sÅ each.
Similarly, the CIV absorption is barely detectable on the first sub-exposure and still very weak on the second before becoming strong. These remarkable variations take place with a timescale of 7--10 min. The changes (particularly for NV) seem to affect the whole profile all at once and are probably caused by variations in the flux of the ionizing radiation. Similar changes also occur in the emission and continuum fluxes, with maximum amplitude near to 30. Unfortunately, GHRS observations in different spectral lines are not simultaneous and this has prevented us studying possible correlations in the observed variations.
We recall that high resolution IUE observations of cataclysmic variables have generally shown that emission profiles are present only in systems seen at fairly high orbital inclination. In low-inclination high mass transfer systems the resonance lines of NV and Si IV exhibit strong blue-shifted absorption features, the typical signature of outflow, while the CIV doublet generally shows a P Cyg profile with a rather weak emission component.
Winds of cataclysmic variables seem to be preferentially accelerated perpendicularly to the plane of the orbit, that is to any accretion disk which is present (Cordova & Mason 1992) but it is still unclear what is the ionization structure in the wind and whether the velocity structure is that of a slowly accelerating outflow.
Theoretical calculations by Knigge et al. (1995) of the CIV profile formed in a wind coming from an accretion disk, suggest that no CIV emission should be seen at an inclination as low as 17. This strongly favors a non-wind origin for the strong CIV emission observed in V603 Aql.
We suggest, therefore, that the emission profiles in V603 Aql come from a chromosphere-corona-like region which surrounds the accretion disk and rotates with it. This explanation is consistent with the fact that the CIV emission is wider than that of SiIV, since the inner edge of formation for CIV is nearer to the disk center than that of SiIV. Simple Keplerian considerations give 8.0 cm and 1.3 cm, respectively, for the inner radii of the CIV and SiIV emission regions. These values indicate that the two regions are rather close to the surface of the 0.8 white dwarf (R 7.0 10 cm) and, therefore, that the emission regions could be photoionized by radiation from the boundary layer.
The blue-shifted absorption components are interpreted as being formed in a conical wind coming from the disk. The more ionized lower NV region has a lower velocity than that of C IV, compatible with outward acceleration of a photoionized wind.
Benefiting from the high sensitivity of the HST-GHRS it has been possible to obtain high-resolution time-resolved spectra for the resonance lines of the ex-nova V603 Aql. Unlike in other low-inclination, high mass transfer rate systems, strong and symmetric emission features are present in the profiles of the CIV and SiIV doublets. Remarkable variations with timescales on the order of 10 minutes are present in the emission and absorption components of the NV, SiIV and CIV resonance lines, and in the adjacent continua. We suggest that the emission and the absorption lines have origin from two separate physical regions, i.e., 1) a chromosphere-corona which surrounds the accretion disk and co-rotates with it and is responsible for most or all the flux in the symmetric and rotationally broadened emission lines, and 2) a conical-shaped wind region, with axis nearly perpendicular to the accretion disk, where the absorptions are formed. The wind is accelerating outward and is photoionized by the energetic radiation coming from the innermost disk/boundary layer.
We are grateful to P. Bonifacio for expert assistance in data reduction.
Andreä, J., Drechsel, H., & Starrfield, S. 1994, A&A, 291, 869
Duerbeck, H.W. 1994, personal communication
Cordova, A.F. & Mason, K.O. 1982, ApJ, 260, 716
Friedjung, M. 1987, A&A, 180, 155
Knigge, C., Woods, J.A., & Drew, J.E. 1995, MNRAS, 273, 225
Ritter, H. 1990, A&AS, 85, 1179
Soderblom, D.R. et al. 1995 in Instrument Handbook for the Goddard High Resolution Spectrograph (GHRS) Version 6.0, Space Telescope Science Institute
Warner, B. 1976 in IAU Symp. 73, ed. P. Eggleton, S. Mitton, J. Whelan, p.85