Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 337, 253-260 (1998)

Previous Section Next Section Title Page Table of Contents

3. Distances

3.1. HIPPARCOS distances

Distances to planetary nebulae are very difficult to obtain with HIPPARCOS due to their large distances (the nearest PN are located at about 100 pc). Table 3 (col.2) gives the distances deduced from the HIPPARCOS results; we give the mean value, and between brackets the minimal and maximal values derived from the parallaxes uncertainty. Good parallaxes have been measured for 4 objects with the usual first order approximation [FORMULA] valid to within about 40%: A 35 ([FORMULA] = 0.2), NGC 1514 and PHL 932 ([FORMULA] = 0.3), SaSt 2-12 ([FORMULA] = 0.4). Furthermore, for most of these objects, the large proper motion supports the large parallax. Positive values are given for A 36, He 1-5 and Hu 2-1 ([FORMULA] = 0.6), NGC 1360 and SwSt 1 ([FORMULA] = 0.7). There is some indication of a positive measurement for LoTr 5 ([FORMULA] = 1.4), NGC 246 ([FORMULA] = 2.), NGC 2392 and He 2-138 ([FORMULA] = 5.), and M 2-54 ([FORMULA] = 8.).


Table 3. Hipparcos distances DH compared to other distances (a= Rao (1987), SB = spectroscopic binaries, G = spectroscopic and atmospheric models, ext = extinction, exp = expansion, kin = kinematic. CKS92 = Cahn et al., 1992; Z95 = Zhang, 1995; VZ = van de Steene and Zijlstra, 1994/1995). PN are ordered by decreasing accuracy of Hipparcos distances (for the objects identified by *, see 3.1). The small PN (diam [FORMULA] 5 arcsec) are listed separately. The mean ratio DH/Dother, weighted according to the errors on DH, is shown for each distance scale. The value D of the adopted distance is given in col. 11 (see Sect. 3.3).

Brown et al. (1997) discuss systematic errors in the HIPPARCOS catalogue, and about the interpretation of the measured parallaxes in terms of distances and luminosities of stars. Following their analysis, the absolute magnitudes computed from HIPPARCOS distances are unbiased for small relative errors, but are on the average 0.6 mag too faint for a 200% relative error. Therefore we increase the values of DH by a factor of 1.32 for the PN with [FORMULA] [FORMULA] 2, which is the case for NGC 2392 ([FORMULA]=1.15) and He 2-138 ([FORMULA]=1.6) (marqued * in Table 3).

A set of 5 PN shows a negative trig. parallax (see Table 1). In addition, very uncertain values of the proper motion components are shown for some of them. Therefore we don't use the HIPPARCOS astrometric data for NGC 2346, for He 2-36 (binary systems) and for the two late [WC] stars. For the central star of PN He 3-1333 ([FORMULA]=-35[FORMULA]30 mas), the large proper motion may indicate that the distance is not too large, but no precise value can be derived from the present measurements.

For the very small PN , there is an additional problem not encountered in other Hipparcos measurements: the presence of a nebula surrounding the central star. If the nebula is not symmetric and if it is as bright as the central star at the same time, spurious measurements may be made. It may be possible to take this effect into account since the nebula brightness is accurately known at a sub-arcsecond resolution in the Hipparcos passband. Then, the individual HIPPARCOS scans could be better interpreted. This has not been done in the present paper because we do not have the necessary photometric images. The same problem may occur for SwSt 1 and Hu 2-1, because these candidates are the smallest nebulae with the highest surface brightness (H and H* differ from 1-2 magnitude, see Table 2). For He 3-1333, the very large negative parallax leads to suspect that that the nebular itself is influencing the parallax measurement.

The error associated with the parallax of these objects is much higher than expected for stars of this magnitude, indicating that the problems that we just quoted, may play an important role. We just want to recommend further investigations.

3.2. Comparison with other distances

Comparison with distances from the literature allows to test the reliability of the methods used. A recent discussion about distances is given in Pottasch (1996, 1997) and Terzian (1997).

Individual distances are based on (1) the spectroscopic parallax for classical MK-type spectra, for the central star itself in a few cases, or for the cool companion if the nucleus is a spectroscopic binary; (2) the analysis of absorption lines in the spectrum compared to model atmosphere in terms of the surface gravity (Mendez et al., 1988, 1992; Pottasch, 1996; Maciel and Cazetta, 1997); (3) the interstellar extinction (see Pottasch 1983, 1996); (4) the tangential versus radial expansion velocities (Hajian et al., 1993, 1995; Kawamura and Masson, 1996); (5) the kinematic distances from interstellar lines intensities (Maciel, 1995, 1997).

Statistical distances are based on average values of the ionized mass or on relations adopted or calculated for some physical parameters (the nebular mass for the Shklovsky-distances, recalibrated by Cahn et al, 1992; radio-continuum brightness temperature versus radius relation by Van de Steene and Zijlstra, 1994, 1995; mass versus radius relation by Zhang, 1995). For He 3-1333, we give the distance adopted by Rao (1987) who used a mean relation between the linear radius and the dust temperature.

Table 3 is separated in 3 parts: the first part indicates the 9 PN with fairly large diameters; then the 4 PN with a diameter smaller than 5 arcsec are listed; finally we show variable and binary stars, 4 of them presenting a negative trig. parallax.

For the different distance methods, we calculate the ratio [FORMULA]; each ratio is weighted according to the errors on HIPPARCOS distances: for [FORMULA] [FORMULA] 0.4, we attribute a weight of 3 (the 4 first objects in Table 3); for 0.5 [FORMULA] [FORMULA] [FORMULA] 0.7, a weight of 2 is given, and a weight of 1 for all other PN with positive parallax measurements.

It appears that all distance scales are overestimated. For our sample, the individual distances which are in closest accordance with HIPPARCOS trigonometric parallaxes are the spectroscopic estimates for binary systems. Concerning the statistical scales, the nearest values to DH are those derived in the scale of Cahn et al., 1992, using the ratio R/r, where r is the angular radius and R the linear radius determined from the ionized mass and an optical thickness factor. The angular radius is very uncertain and can differ by a factor 2 or 3 from one observer to the other. Especially in the case of very small PN, a stellar appearance leads to an underestimation of r. But for the small nebulae, the uncertainty in the HIPPARCOS measurements, as discussed previously, may play a role, leading to unreliable distances, largely too small compared to all statistical distance scales (see Table 3).

3.3. Adopted distances

The adopted distances (shown in Table 3, col. 12) are those deduced from HIPPARCOS parallaxes for the large PN. For the others, the distance (between brackets) is the mean value of all available distances (excepting He 2-36 kinematic distance), weighted by the ratios calculated for each distance scale (Table 3). For LoTr5 with very uncertain parallax, we select a value of the distance inside the HIPPARCOS distance-range, in agreement with the best literature value (based on spectroscopic classification). We give the adopted distances with a number of significant decimals corresponding to the uncertainty.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: August 6, 1998