2 The observations

The observations were carried out with CCD imagers on three separate telescopes.The majority of the observations were made with the focal reducing imager at the Nasmyth B f/18 focus of the ATT 2.3 metre telescope operated by the Australian National University at the Australian Siding Spring Observatory (ANU - SSO) during two campaigns (July 1995 - July 1997). This instrument provides an f/3.5 focal ratio at the detector. NGC 6853 (the Dumbbell nebula) was observed with the OHP 120-cm telescope at St Michel l'Observatoire. The image scale and field coverage of the images are very similar for the three instruments: the Nasmyth imager SSO-Tektronix 1024 $\times$ 1024 CCD has an image scale of 0.60$^{\prime \prime}$/24 $\mu$m-pixel, the f/7.8 1-m McLellan telescope at MJUO with its 380 $\times$ 576 CCD offers an image-scale of 0.62$^{\prime \prime}$/23 $\mu$m-pixel giving a much smaller field of view (4$^{\prime}$$\times$ 6$^{\prime}$), and the f/6 120-cm telescope in Haute Provence with its Tektronix 1024 $\times$ 1024 CCD is characterized by a plate scale of 0.686$^{\prime \prime}$/ 24 $\mu$m-pixel and a larger field of view of $\sim\!12'$.

Typically, three exposures of 600s (SSO's data) were made in each filter, which allowed the identification and removal of cosmic ray events. At MJUO and OHP observations were done with 1 hour exposures. Images were bias-subtracted and flat-fielded in the usual way. The flat fields were generally obtained in twighlight to ensure that the illumination across the detector was the same as the sky illumination in the observations. The separate images were offset relative to each other by $\sim\!1$ arcsec, and combined using the IRAF task imcombine which allowed to remove detector artifacts such as dead or noisy pixels.

2.1 The interference filters

Most of the published "narrow band'' PNe images have been actually obtained with filters having bandpasses of order $\Delta$$\lambda$ $\approx$ 50 Å. Mostly these consist of [OIII]$\lambda$5007 Å and H$\alpha$  images. As a result of the broad bandpass, the nitrogen lines, particularly the [NII]$\lambda$6583 Å emission, contaminate the H$\alpha$ flux, especially for Type I PNe, in which the [NII]$\lambda$6583 Å line may in some cases be 4-6 times stronger than the H$\alpha$ line. Problems may also arise in the case of evolved hydrogen deficient objects such as A 79 (see Jewitt et al. 1987; Schwarz et al. 1992). The interference filters used in the present study have bandwidths $\leq$  10 Å, and are centred at specific wavelengths suitably selected for investigating nebular ionization structure and/or the PN abundance distribution. These filters are generally mounted directly in front of the detector. Provided that the f-ratios are not too small, f/6 being about the limit with such filters, the bandpass broadening due to the off-axis rays in the beam is negligible. In the case of the Nasmyth B imager, the filters were placed at a 50 mm pupil image in the focal reducer, so that every portion of the field passes through the same region of the filter. This removes the effects of inhomogeneities in the multi-layers, but induces a field-dependent shift in the bandpass. However, this should also be negligible for the filters we used in this study. The image quality was typically $1\hbox{$.\!\!^{\prime\prime}$}5 - 2\hbox{$.\!\!^{\prime\prime}$}0$ during the observations.

2.2 The photometric calibration

The CCD responses were calibrated by observing the standard star $\tau$ Sco (HD 149438, Tüg 1975), and the "compact'' planetary nebula flux standards previously calibrated by Dopita & (1997) using the Nasmyth A spectrograph of the ATT 2.3 metre telescope. The cleaned standard star frames then were used for the ADU counts/absolute flux conversion after airmass correction, sky subtraction and allowance for the filter bandwidths. Absolute fluxes are given in erg cm^-2 s^-1 units. Nebular and standard star frames were processed and calibrated using both the IRAF and MIDAS software packages.

2.3 The selected planetary nebulae

The PNe we have chosen to observe were selected from the Acker et al. (1993) Catalogue (hereafter referred to as Acker93). Table 1 lists these objects whose main characteristics were extracted from the Cahn et al. (1992) (CKS92) report where available. Otherwise we have used data from Acker93. We generally have concentrated on the PNe with the largest angular extent, and which are classified as high excitation ([OIII]/F(H$\beta$)> 7) and are characterized by strong [NII] lines (intensity ratio H$\alpha$/[NII] $\leq$ 1; see Table 2). According to Kaler 1978, such high-excitation class PNe would have low IRE and central star effective temperatures larger than 60 000 K (see Fig. 11, p. 78 in Zijlstra's thesis). Note the large discrepancies between optical and radio extinction values (Table 2) for the PNe of our sample.

These above selection criteria ensure that we have observed relatively nearby PNe, and which are mostly Peimbert Type I PNe. Such PNe presumably have relatively massive progenitors which have been through a third dredge-up stage with "hot-bottom burning'' on the AGB, producing a relative overabundance of nitrogen. These more massive PNe lose a greater fraction of their mass while on the AGB, and have more and more frequent helium shell flashes. They are therefore better candidates to discover remnants of material ejected in previous helium shell-flash episodes. They are also characterized by more rapid evolution and more energetic stellar winds during the PN phase.

 

Table 1: List of observed PNe  

S(6 cm) in mJy. Distance values marked by (*) were taken from Maciel 1984. Size's values (Col. 11) and some other parameters are mainly taken from Acker93.

 

Table 2: Parameters of observed PNe  

He/H and extinction values are from CKS92. Flux density values are computed with the radio flux S(6 cm) in mJy from Table 1 and using the formula S(6 cm) $\times$($d_{\rm kpc}$/7.8)², and F(H$\beta$) $\times$($d_{\rm kpc}$/7.8)². ND = not detected. The dimensions (for inner areas) in three emission lines are in arcseconds as measured in this report. Column 11 gives $\alpha$'s values to be multiplied by appropriate $\epsilon$ and $N\rm _e$ to obtain estimates of ionized mass in solar masses for the overall nebula (see text).