2. The observations
Spectral line and continuum observations of an area covering the three main fingers at millimetre and sub-millimetre wavelengths, were made with the 15 metre James Clerk Maxwell Telescope (JCMT) in Hawaii during April and September 1997. These were supplemented by millimetre wavelength spectroscopic observations using the Onsala Space Observatory (OSO) 20 metre telescope during March 1998. The telescope efficiencies and beamsizes at the various frequencies are summarised in Table 1 (observations below 116 GHz were made at OSO, those above 200 GHz at the JCMT):
Table 1. JCMT telescope parameters - April 1996
The spectral data were collected during periods of good atmospheric transparency, when both the JCMT and OSO's pointing and tracking accuracy were better than . The JCMT spectra were collected and processed using standard JCMT receivers with a digital autocorrelation spectrometer backend. The receivers had single sideband system noise temperatures ranging from 300 K in the CO J = 2-1 and 3-2 transitions, 400 K for HCO+ J = 4-3 to 3000 K for the CI measurements, and the data were taken with velocity resolutions of 0.2 km s-1. The OSO spectra were observed using a frequency switching with an SIS receiver which was operated in a single-sideband mode. This was used with a digital autocorrelation receiver, which provided a spectral resolution 0.06 km s-1. The data were calibrated in units of main beam brightness temperature using a standard chopper wheel calibration technique with an additional correction for the relative sideband gains, and an value measured from observations of the planets. The scale describes the brightness of a source which completely fills the main beam, but is not seen by the side-lobes. The maps presented later indicate that this is generally the case in the J = 3-2 observations.
Several `Standard' calibration line sources were observed over a range of elevations at both OSO and the JCMT, and their intensities found to be in good agreement with historical values. In particular the Eagle Nebula was observed at low elevations from Onsala (maximum elevation 18 degrees). Frequent observations of compact standard calibration sources observed at low elevations agreed to within 10 percent of values obtained at higher elevations. At the JCMT, each of the map grids were sampled every - except the CI map, which was sampled every . All of the JCMT spectral line data were position switched against a clean off-position 0.8 degrees NW. At the distance of M16, corresponds to a linear size of 0.1 pc. All observations reported in this paper are labelled using 1950 coordinates, or, where appropriate, using relative offsets.
The submm continuum data were obtained with the JCMT's SCUBA array. This is a submillimetre bolometric camera containing two detector arrays cooled to 0.1 K. These simultaneously view the same patch of sky, about in diameter, via a beam-splitter. The long-wavelength (LW) array has 37 elements optimised for operation at 850 µm, and the short-wavelength (SW) array has 91 elements for use at 450 µm. The individual bolometer beams are spaced about one beamwidth apart. Three longer wavelength bolometers surround the 850 µm array, offset on the sky from each other by about , which are used for point-source photometry. The 450 µm array can also be used at 350 µm, and the 850 µm array at 750 µm and 600 µm, with somewhat reduced sensitivity. The data were obtained in the `Jiggle Map' mode, where the chopping secondary mirror is moved so as to steer each detector beam over a pattern on the sky. This allows fully-sampled maps to be made with points spaced and apart for the SW and LW arrays respectively, after on-line re-gridding to an RA / Dec frame. For the observations reported in this paper, the restored half-power beam-widths at 2000, 1300, 850, 750, 450 and 350 µm were , , , , and respectively.
Radio observations at 8.69 GHz were obtained using the NRAO Very Large Array (VLA) in the D-configuration, which provided a synthesised beam of HPBW over an area defined by the primary beam of . Two short integration (`snapshot') images separated by about 2 hours in time were taken with integration times of 5 minutes across a bandwidth of 100 MHz. The data were edited and calibrated using the AIPS software package and mapped onto a grid with a pixel size of . The final images were produced using the routine MX, a version of the CLEAN algorithm, and contained 1000 CLEAN components with a loop gain of 0.1. The maps were then restored with a circular gaussian beam of FWHM, giving 4 pixels across the synthesised beam. Further details of the observing set-up are described in Balser et al. (1995).
ISOCAM observations were taken from measurements of the original dataset reported by Pilbratt et al. (1998). Two broad bands were observed; a region at a central wavelength of 6.7 µm (LW2), with pixels and an effective resolution of , and a region at a central wavelength of 14.3 µm (LW3), with pixels and an effective resolution of . Based on current accumulated ISOCAM experience, the calibration is believed to be accurate to within 10%.
© European Southern Observatory (ESO) 1999
Online publication: December 22, 1998