Imaging and spectrophotometric observations were carried out during 5 consecutive nights in August 1996 using the near-infrared camera/spectrograph CAMILA (Cruz-González et al. 1994) on the 2.1m telescope of the Observatorio Astronómico Nacional (at San Pedro Mártir, Baja California, México). The camera/spectrograph has two cameras with f/4.5 and f/13.5 (hereafter "low resolution" and "high resolution" cameras, respectively), equipped with a , 24µm pixel NICMOS3 array. The low and high resolution modes yield /pixel and /pixel scales, respectively.
We obtained several images and spectra of the IRAS 20126+4104 region through H2, cK and -band filters. Both images and spectra were reduced using standard sky subtraction and bad pixel removal techniques with IRAF-based programs. The images were processed by subtracting a median-filtered image of nearby sky frames taken with the same integration time and with an offset of about from the IRAS source position. The images were flattened by a combination of low and high illumination sky flats obtained at sunset. For each filter, mosaics were made with frames containing the IRAS region, which were then aligned using several field stars. Finally, we estimated the plate scale and the orientation of the images using 7 optically visible stars in the Digitized Sky Survey, obtaining a positional uncertainty of .
In Table 1 we present the width and central wavelength in microns of the filters used, as well as the total integration times for the final mosaic images. The total integration times correspond to the overlapping areas toward the center of our narrow-band mosaics. Fig. 1 shows our low resolution H2 1-0 S(1) + continuum image, in grey scale, which is discussed below.
Table 1. Filters and exposure times used in direct image observations
For the near-infrared -band spectra (1.95-2.30 µm), we employed the f/4.5 camera and a slit of width and length, giving a spectral resolution of R= =500 in . Initially, the slit was oriented E-W in order to intersect the brightest object located NW of the IRAS 20126+4104 position (labeled "1" in figure 2a). The total exposure time for the final spectrum in this position was 10 minutes. We also obtained several spectra for two slit positions separated by in declination (labeled "2a" and "2b" in figure 2a) with a PA position angle, measured counter clockwise, covering most of the nebula located to the SE of the IRAS position. Each spectrum has a 2 minute integration time. For each setting, explicit sky observations were obtained. We constructed a spectrum with 12 minutes of total integration time by adding the spectra obtained through slits 2a and 2b.
The wavelength calibration was performed using exposures of an Argon lamp. The absolute flux calibration was achieved through observations of SA94-242 (UKIRT faint standard list) and HD 201941 (Elias et al. 1982), both early type standard stars. For the correction by atmospheric absorption we used the early type star HR 7628. The conversion factor from counts to flux was determined as a correspondence between the counts and flux per unit wavelength in the central wavelength of the K filter, using the standard spectra corrected for atmospheric absorption. Errors in the absolute flux calibration are around of 10% for the brighter emission lines detected. Since the lines are not resolved in our spectra, velocity shifts along the slits are not detected. The narrow and broad band mosaics were calibrated in flux using the K spectrum for the brightest object (slit 1). In each mosaic, we integrated the "flux" in counts/sec inside the region covered by the spectrograph slit. We then computed a conversion factor between this "flux" and the integrated flux density in the wavelength range that corresponds to the bandpass of each filter (see Table 1).
© European Southern Observatory (ESO) 1998
Online publication: March 30, 1998