Astron. Astrophys. 363, 671-674 (2000)

4. Discussion: which nebular geometry?

The present data confirm the result of P90 that RX Pup possesses an ionized nebulosa which is mainly extended toward the North (P.A.^o) out to 3:002 from the central stars (5800 a.u. for the adopted distance of 1.8 kpc). Our spectra detect the extended nebula only in the [NII ] light, and thus do not support the estimate of P90 of a large H/[NII ] line ratio. Also the interpretation of P90 that the extended nebula might represent a jet is not supported by our radial velocity measurements. The decrease in radial velocity with radius is in fact quite peculiar, although it might be due to projection effects of a curved jet.

Our spectra also reveal the existence of a marginally resolved [NII ] nebulosity along the EW direction, with its eastern component blueshifted by 80 km s^-1, and the western one redshifted by the same amount. This nebulosity is contained within a distance from the centre of less than 1:005 (2700 u.a.), but definitely much larger than the separation of the central stars (a few tens a.u.). As for HM Sge and V1016 Cyg (Corradi et al. 1999b), we find that the [NII ] feature in the spectrum of symbiotic Miras trace the existence of low density, low ionization material located in a very extended (few 100 times the binary separation) circumbinary region. The observed velocity splitting of the EW nebulosity of RX Pup appears to be too high to indicate a rotational pattern possibly linked to the binary motion or to the presence of rotating circumbinary discs. It is more likely that instead it represents an outflow from the system, in the form of an expanding ring (such as the one in the other symbiotic Mira He 2-147, Corradi et al. 1999b) or, - more likely given the behaviour of the velocity field with P.A. - a collimated or bipolar outflow extending EW. If so, its kinematical age would be  yr, where D is the distance in kpc, is the apparent radius of the outflow in arcsec, and i its inclination to the line of sight. Considering the uncertainties in D and (see previous section), and depending on the (unknown) inclination of the outflow, the EW nebula might well be the ejecta of the 1894 nova-like explosion or even, for moderate to high inclinations, the result of the energetic stellar wind started after 1975 during the present outburst (see Sect. 1).

In any case, RX Pup seems to show two main `preferred' directions for mass ejection which are roughly pependicular to each other: the EW innermost component, and the more extended structure at about P.A.=15^o. The same also appears in the radio data (Hollis et al. 1989); at 2 cm, emission is resolved in three components along the EW direction and located within the innermost 0:004, while the 6 cm emission has a northern protrusion. With the present data, it is not possible to get further insight into this complex nebular geometry. In some respect, the situation resembles that of HM Sge (Corradi et al. 1999b), whose outflow shows a complex geometry, with different symmetry axes appearing at different observing wavelengths. At this stage, we cannot even exclude, for both systems, the naive possiblity that the direction of mass ejection is drastically changing from one outburst to the next. Clearly, a better understanding of the complex, multiple ionised nebulae of RX Pup and HM Sge requires [NII ] imaging with HST resolution. Considering also that polarimetric observations give us strong indications about the apparent orientation of the central binary stars (Schmid et al. 2000; Mikolajewska et al. in preparation), such HST imaging would be especially important to determine the actual direction of the outflows with respect to the orbital planes, which would in turn provide important insights into the mass loss mechanisms from symbiotic binaries and related systems.

© European Southern Observatory (ESO) 2000

Online publication: December 11, 2000
helpdesk.link@springer.de