3. Observations and reduction
IC 4406 was observed on March 1994, at the 1.52m ESO telescope of La Silla, Chile, equipped with a Boller & Chivens spectrograph. The detector used was a Ford FA2048L CCD. On the first night, the ESO grating No.25 was used giving a reciprocal dispersion of 0.28 nm pix-1 (spectral resolution 2 pix) and a useful spectral range from 360 nm to 840 nm (see below). The slit of the spectrograph was placed on the long axis of the nebula and passed through its central star. The position angle (P.A.= ) was chosen as to avoid a foreground star which lies almost exactly along the symmetry axis of IC 4406. A drawing of the slit, superimposed on an H +[NII]658.3 image of the nebula taken from Corradi & Schwarz (1993), is shown in Fig. 1. The slit width was , and the spatial scale along the slit pix-1. With this configuration, two spectra with exposure times of 45 and 60 min were obtained. The seeing was around .5 FWHM (full width at half maximum).
On the second night, red spectra of IC 4406 were obtained by using ESO grating No.14 and an order separator filter cutting all wavelengths below 680 nm. The useful spectral range is from 710 nm to 930 nm, with a reciprocal dispersion similar to that of the blue spectrum. Two exposures of 2 and 45 min were obtained. The slit width and position angle were the same as for the blue spectrum. Only in the central position the shorter exposure proved to be sufficiently exposed to allow to measure the line intensities and to compare them with the values given by the longer exposure. On each night, three spectrophotometric standard stars were also observed.
The data were reduced using the standard MIDAS procedure for long-slit spectroscopy. In the blue spectrum the instrumental response curves from individual standard stars agree within 10 of their mean value from 360 nm to 840 nm. This is the wavelength range that we have adopted for the subsequent analysis. Outside this range, the instrumental response curves from different standard stars deviate significantly ( 20 ). For similar reasons and because of uncertainties in the absolute fluxes tabulated for the selected standard stars above 950 nm, we have used a spectral range between 710 nm and 930 nm, although in the original spectrum nebular lines as red as the HeI 1083 nm were detected.
3.1. Flux measurements
Line emission from several ions was detected through the whole () optical extent of IC 4406. Before measuring line fluxes, one-dimensional spectra of a number of regions were extracted by spatial binning along the slit. We have decided to divide the section of IC 4406 covered by the slit into seven parts (a central region, plus three symmetrical ones on each side of the central star), in order to increase the S/N ratio in the faint lobes while keeping a good number of positions where to investigate for radial variations of chemical abundances. In the following the central region is referred to as c, while regions on the eastern lobe (the "positive" side of the nebula according to the oriented direction defined by P.A.= ) are named p1, p2, and p3 for increasing distances from the centre, respectively. Analogously, the three regions on the western ("negative") lobe are called n1, n2, and n3. The exact location of the middle point of each region and its extension are given in Table 1. The limits of the regions are also indicated in Fig. 1.
Table 1. The selected regions of IC 4406
Line fluxes in the average 1-D spectra were measured using the MIDAS/ALICE package: multi-Gaussian analysis was performed to separate partially blended components. Line fluxes measured in the two spectra with different exposure times in the same position were then averaged with a weight proportional to the exposure time, after normalisation within each spectrum to H . In the blue spectra, the errors on the average fluxes have been estimated by considering the deviations of the individual measurements from the average values. These deviations resulted larger than the errors as computed by the MIDAS/ALICE program from the r.m.s. of the continuum fit, and larger in the more external positions than in the positions closer to the central star. Accordingly, the errors on the average fluxes have been adopted as follows. In positions n2, n3, p2, p3, they are taken to be as large as 10 , 20 , 30 and 50 for the lines whose intensity is 0.2 H , between 0.05 and 0.2 H , between 0.01 and 0.05 H , and 0.01 H , respectively. It has to be mentioned than only in very few cases in the subsequent analysis we will use line whose intensity is less than 0.01 H . For the central positions (n1, c, p1), which are significantly brighter, the estimated uncertainty of the average fluxes resulted to be well represented by one half of the values quoted above. In the near-IR spectra, fluxes from the two differently exposed frames can be compared only in position c, and give deviations similar to those of blue spectra in the external positions. This spectral region, however, has larger uncertainties in the correction for the instrumental response curve, and we conservatively assign an error of 40 to the fluxes (note that only the [SIII]906.8 line is used in the following chemical analysis). Fluxes from the redder spectrum were then rescaled to the blue one using lines in the overlapping spectral region (710-840 nm). The average observed fluxes are listed in Table 2. Conditions during the observing run were photometric, and we also give in Table 2 the absolute flux of H . We estimate the uncertainty on these absolute fluxes to be of the order of .
Table 2. Observed line fluxes.
© European Southern Observatory (ESO) 1997
Online publication: June 5, 1998