Astron. Astrophys. 351, 1075-1086 (1999)

3. Data reduction

The spectra were reduced to absolute intensity units using the IRAF reduction package, following standard procedures for the long-slit case. The overscan region in each frame and the bias frames were used to remove the electronic pedestal level, its variations from frame to frame and the residual structure in the bias level.

The pixel-to-pixel gain variations were removed by using the tungsten flat-fields taken each night for each grating position; the twilight exposures were used to correct for the deviations of these lamp flat-fields from the slit illumination function.

The wavelength calibration was performed at different positions along the spatial axis, thus correcting for the optical distortions noticeable in the curvature of the slit image. The flux calibration was performed separately for each exposure, using the standard stars fluxes and the mean extinction curve for La Palma.

The sky emission lines were removed by subtracting the sky spectra after scaling them by factors of 0.5-1.5 to obtain the best cancellation. The equivalent nebular spectra were then combined to remove cosmic rays and to improve the signal-to-noise ratio. Some of the shortest nebular exposures were also retained, since some lines saturate in the longest exposures.

Sky subtraction is a critical step for most of the near-infrared lines. In general, all lines with  Å are affected because of the uncertainty contributed by sky emission to the continuum around the lines, but the biggest problems can arise for lines whose wavelengths are close to sky-emission lines: [OI ] , [NiIII ] , [ArIII ] . Even relatively strong lines, like [OII ] , are affected by sky emission in some objects, like M16 and M20.

The stars appearing in most of the slit positions were used to estimate the possible spatial shifts of the slit image from one spectral range to another and across the same range: in all cases the misalignments remained to within one pixel. These stars were also used as references for extracting the spectra in several regions for each object, especially when the sum was performed over just a few pixels. Table 2 gives the positions and sizes of the selected regions, the nebular areas covered are indicated in Figs. 1 and 2 on reproductions of the seven HII regions.

Fig. 1. Broad band image of the core of M42 (from Münch & Wilson 1962). The image size is about ; north is at the top, west to the right. The Trapezium cluster - and  Ori C, the main ionizing star of the nebula - can be seen north of slit position A. The bright star south of the bar (slit position B) is  Ori A. The regions A-5 and A-6 are consecutive, cover 33" each and are located west of the region A-4, with A-5 starting 7" west of A-4. The region B-5 covers and starts 26" from the end of B-4

Fig. 2. images of M16, M8, M17, NGC 7635, M20 and M43, along with the slit positions selected for each object. The size of the original images is ; north is at the top, west to the right. The images were obtained at the IAC80 telescope (Observatorio del Teide, Canary Islands, Spain). The image size for M43 is about ; this image was kindly obtained by M. Manteiga

Table 2. Positions, sizes and extinction corrections.
Notes:
a) Distance from the centre of each area to the slit positions listed in Table 1. Positive numbers indicate shifts to the west (to the north for M8)
b) In erg cm^-2 s^-1
c) The selected region is wide, but the central , containing a stellar spectrum, have been excluded

Line intensities were measured by integrating between two given limits above a continuum fitted around each line. Blended lines, especially [ClII ] 8579 and HeI  8583, were measured by fitting two Gaussian profiles of equal width, using the wavelength differences of the two components to constrain the fit. All these measurements were made using the fitlines routine implemented in IRAF by J. Acosta.

The line intensities have been normalized to the brightest HI line appearing in the same spectral range: H for the blue range, H for the red range and Pa12 or Pa13 for the near-infrared range. These line ratios, corrected for reddening (see Sect. 4 below), are used to determine the ionic abundances presented here. The tables listing the observed and reddening-corrected relative intensities for all lines of interest and each region considered are available on request to the author.

© European Southern Observatory (ESO) 1999

Online publication: November 16, 1999
helpdesk.link@springer.de