Astron. Astrophys. 338, 262-272 (1998)

2. Observations

The measurements of the ²PP fine structure transition at 157.7 µm of singly ionized carbon were performed during two flights in October 1991 with the MPE/UCB Far Infrared Fabry-Perot Imaging Spectrometer (FIFI, Poglitsch et al. 1991) on board the Kuiper Airborne Observatory (KAO) from NASA Ames Moffet Field, California, USA.

With FIFI's efficient mapping capabilities due to its 55 focal plane array with detectors spaced by 40", we observed 425 points in 17 array settings along two cuts at constant Galactic latitude. As indicated in Fig. 1, the observations covered the HII region/molecular cloud interface as well as the remote part of the molecular cloud. For each detector, the beamsize is 55" (FWHM) with an approximate Gaussian beamshape (68" equivalent disk). Due to the large spatial extent of the RMC, we observed in the frequency-switching mode of FIFI. The Fabry-Perot filter was switched at 10 Hz to a reference wavelength while sweeping the observed center wavelength of the filter across the [CII] 158 µm line. The Fabry-Perot was operated in 120th order, resulting in a velocity resolution of about 52 km s^-1 (v₅₂) so that the spectral lines are not resolved.

Fig. 1. An overview of the Rosette region. Overlaid on the digitized Palomar Observatory Sky Survey plate is a KOSMA 12CO J=21 map (contour lines range from 6.7 K km s-1 (3) to 87.4 K km s-1 in steps of 9), outlining the most massive part of the Rosette molecular cloud. This map shows the close association with the HII region NGC 2237/2246, which is illuminated by the OB cluster NGC 2244 located in the central cavity of the nebula. The regions covered by [CII] observations are indicated by black lines.

The data were calibrated by observing an internal blackbody source. The absolute pointing uncertainty is below 15". The velocity integrated [CII] line intensity is given by the amplitude of a fitted Lorentzian profile to the spectra. Due to instrumental effects such as walk-off for the non-central pixels, the width and position of the instrumental profile varies across the array. The observed variation is close to the one expected theoretically (Geis 1991) so that we choose the theoretical value for the Lorentzian center (relative to the central pixel) and width in each pixel and fixed them for the fit. Fig. 2 shows as an example one FIFI array setting with the observed data and the thus defined Lorentzian fit.

Fig. 2. The original [CII] data (spectral line intensity per resolution element) of the 55 pixel Array no. 12 (see Fig. 3 for the position of array setting) are shown as histograms. The lines are not spectrally resolved. The line integrated intensity is obtained by simultaneously fitting Lorentzian shape profiles to the 25 pixels, with the offsets and widths in each pixel fixed to the theoretically calculated instrumental profile. The fit profiles are drawn as continuos lines.

An instability of the electronic detector readout of FIFI during frequency-switching observations on the flight series in fall 1991 led to a nearly periodic disturbance which is hardly visible in the individual pixels but is clearly apparent in the average spectrum across the 25 pixel of each array. From the averaged data of each array, we derived the period and amplitude which always was close to 2.90.2 channels and (0.40.2)10^-4 erg s^{- 1} cm^-2 sr^-1. We then subtracted a sine function with the respective parameters from the averaged data of each array and performed the fit of the Lorentzian profile. Table 1 lists the values of all arrays for the fitted C⁺ intensities and their standard deviation, and the ratio between intensity and rms noise, indicating the significance of the detection. The average standard deviation of the fitted intensity is (0.350.06)10^-4 erg s^{- 1} cm^-2 sr^-1 so that the rms noise for one pixel is accordingly = which is (1.750.3)10^-4 erg s^{- 1} cm^-2 sr^-1.

Table 1. (1) No. of array, (2) Offset in arcsec of the central pixel from the position , (3) line integrated C⁺ intensity, averaged over one array, and obtained by a Lorentz fit (4) standard deviation of the fitted intensity (5) ratio between the fitted intensity and the respective standard deviation, i.e. the S/N ratio of the data.

© European Southern Observatory (ESO) 1998

Online publication: September 8, 1998
helpdesk.link@springer.de