Astron. Astrophys. 347, 258-265 (1999)

2. Observations

2.1. 2.1 and 3.5mm continuum measurements

We observed during 1997 December and 1998 February at the NRAO 12m antenna on Kitt Peak using the usual dual-polarization SiS mixers tuned double-sideband to GHz and GHz; the IF frequency was 1.5 GHz and the instantaneous bandwidth in either sideband was reputed to be 600 MHz. The measurements are not exactly monochromatic but we simply could not afford to sacrifice half the possible incoming signal. In the discussion we take the frequencies of observing to be 86.0 and 140.0 GHz and ignore some slight non-linearities, such as the slope of the Cas A spectrum, , because it was not possible to determine the relative sideband gains.

The quality of the sky was excellent; the monitored opacity at 220 GHz was a few percent or less, and we have ignored opacity corrections to the measured fluxes since both Cas A and our calibration sources were measured at comparably high elevations. The system temperatures were also excellent, typically 150-170 K (SSB) at both 86 and 140 GHz, but we were chagrined to discover, after the fact, that the noise in our continuum maps is 3-5 times the theoretically-expected level. This is apparent in the raw beam-switched data samples and is in some measure caused by excess noise in the broadband IF electronics (J. Mangum, private communication).

The continuum maps were made by beam-switching at 4 Hz over a 4´ ´) distance in azimuth while scanning in azimuth over a raster covering the face of the Cas A remnant: a region just larger than one full throw must be observed on either side of the source as well. This is basically the EKH method of Emerson et al. (1979) which is now accomodated within the AIPS software package due to the recent efforts of E. Greisen. We mapped Cas A several times at both frequencies, and mapped Jupiter and Saturn in the same way as calibrators at the lower and higher frequencies, respectively. To determine the flux scale we simply integrated over the maps of Cas A and those of the planets, taking the ratio and using the fluxes of Ulich (1981) for the latter. To calibrate the temperature scale, we use the canonical values of the gain (Jy/K) for the 12m antenna, which are 30.6 Jy/K and 34.5 Jy/K at the two frequencies. The default calibration values gave empirical gains of 29.4 and 27.5 Jy/K and their difference from the canonical values must partly reflect slightly unequal gains in the two receiver sidebands. Note that the temperature scale of the 12m is somewhat elevated, which explains the perhaps unexpectedly high gain for such a small dish.

At 3.5mm, the error in the measured flux arises from several contributions; the point-to-point noise of 0.005K; a systematic error which we estimate at 2% by integrating over variously-sized regions encompassing Cas A; and the inherent error in the flux scale, quoted as 2.6% by Ulich (1981), compounded slightly by the non-monochromatic nature of our measurement and the need to interpolate his widely-spaced measurements to the mean frequencies of our observations. The planets are very nearly perfect radiators over the range 90-150 GHz, with .

The point-to-point noise is quite small compared to the peak 3.5mm signal in Cas A, 0.35 K, and it contributes negligibly to the total flux measurement. The quadrature sum of all the sources of error at 86 GHz is 3.3%. At the higher frequency, the point-to-point noise is higher, 0.008K compared to a peak of 0.11 K, but the dominant source of error in the total flux measurement is the 8% error quoted by Ulich (1981) for the 2mm flux. We quote a 9% error, the quadrature sum of 8% and 4%.

The HPBW of the 12m was calculated using standard quantities to be 71.9" and 44.17" at 86 and 140 GHz, respectively. These agree perfectly with the maps of Jupiter and Saturn, where we did gaussian fits to the images taking into account the apparent semi-diameters of the planets at the time of observation.

2.2. CO on-the-fly spectral line mapping

We made maps of ¹²CO J=2-1 (HPBW = 28") and ¹³CO J=1-0 (HPBW = 60") emission in the usual on-the-fly (OTF) mode, but used a reference position only 4´ N of the center of the nebula to minimize switching times and to ensure correct tracking of the local oscillator. That is, the local oscillator could be set for either the map center or the off-position and still produce correct results. At the time of observation there was reason to doubt that this would be true for either very distant off-positions or very large maps. Emission north of Cas A is quite weak, but we took a high signal/noise spectrum of the off position relative to a more distant reference 1^o north of the center of the nebula following Wilson et al. (1993) and added this to the datacube at all positions.

Only the J=2-1 ¹²CO dataset is discussed here in detail. It was made by feeding the hybrid correlator configured as two 786-channel devices. The correlator took data in channels separated by 97.7 kHz, but these data are always Hanning smoothed before presentation to the user to remove a sin(x)/x response. We decimated the spectra by discarding every other channel so that the channel separation is 195.4 kHz or 0.254 km s^-1, which is also the effective resolution. The final datacube was produced on a grid having 8" pixels, compared to the HPBW of 28". The channel-to-channel noise is 0.35-0.4K. The ¹³CO J=1-0 map was made in the same way but with channels half as wide in frequency (0.133 km s^-1 in the raw data) and on a 16" grid (compared to the 57" HPBW). The noise was 0.20 K except for a narrow strip near the top, where it was 0.4 K.

The CO J=2-1 datacube is represented by a series of moment maps (line profile integrals over 2 km s^-1 intervals) in Fig. 4, and in profile form via a series of declination-velocity diagrams in Fig. 5. Some ¹³CO spectra are shown in Figs. 3, 6 and 7.

2.3. Other spectra

We also took frequency-switched spectra of the J=1-0 lines of HCO⁺ and C¹⁸O with 97.7 kHz resolution at three positions crossing the strongest molecular emission feature in a north-south direction at the eastern edge of the nebula (see Fig. 6 and the caption to Fig. 4), and at two positions near the western rim (see Fig. 7). These were chosen to coincide with peaks in the ¹³CO map and appear to be displaced from corresponding ¹²CO J=2-1 features in a minor way. Interpolated to the 89.2 GHz frequency of the HCO⁺ line, the nominal beam-averaged continuum signal at the uppermost of the three easterly positions is 0.22 K, and at the more easterly of the other two spectra, = 0.21 K. The actual continuum fluxes at the positions of the spectra are somewhat uncertain owing to possible telescope pointing errors and errors of registration between the continuum and spectral line datasets.

© European Southern Observatory (ESO) 1999

Online publication: June 18, 1999
helpdesk.link@springer.de