The observations were done with the 15-m Swedish-ESO Submillimeterwave Telescope (SEST) at La Silla in Chile. The , and lines were observed in December 1995, the HCN, HNC and CS lines in July 1996, where we also obtained some additional observations of the and lines. A final observing session in August 1996 only involved the line.
For all the observations we used the 3-mm SIS mixer, giving a system temperature in the range 140-160 K at the observed frequencies (see Table 1). The low allowed us to obtain spectra with significantly better signal-to-noise than previous observations of the molecular absorption line system in Cen A. As backend we used the high resolution Acoustical Spectrometer (AOS), giving a resolution of 43 kHz or 0.14 km s-1. Due to a Lorentzian response of the AOS, the actual velocity resolution of the spectra are slightly worse than this. The effective resolution is 0.2 km s-1. The observations were done in a dual beamswitch mode, with a beamthrow of in azimuth at a switchfrequency of 6 Hz. Both the receiver system and the weather were good and stable during the observations. The pointing was checked on the continuum of Cen A itself and the rms variations were less than 3". The good pointing accuracy is evident as small rms variations of the continuum level in 10 minute averages of spectra obtained during 8 hours of observations.
Table 1. Observed molecules
The observed continuum levels at the different frequencies are given in Table 1. Some of the high resolution spectra show a distinct curvature due to the presence of emission. We subtracted a second order baseline from the average spectra of each molecular line. Redshifted absorption features in the velocity range 560-640 km s-1 can be seen in the , HCN, HNC and CS spectra. The region between 490-660 km s-1 was therefore omitted from the fitting procedure. The bandwidth of the spectrometer does not cover the full extent of the emission seen in low resolution data (cf. Israel et al. 1992). We derived the continuum level at the edges of the spectra in order to eliminate contribution from the emission. For and HCN, however, the presence of emission produces a flux 2% too high. This transforms into an error in the optical depth calculations which is 10% for the deepest absorption. For a line with a depth of 0.5 relative to the continuum, the error is 3%.
The 3-mm continuum flux decreased by 20% between December 1995 and July 1996. This is consistent with intrinsic variations of the millimetric flux of Cen A (cf. Tornikoski et al. 1996). The dispersion in the continuum flux determinations are caused by pointing and calibration errors. The latter is of the order 10%, whereas the former is likely to be less than this. The opacities of the absorption lines are, however, independent on the pointing 1, given only by the continuum to line ratio. In the following we have normalized all the continuum levels to unity using the fluxes given in Table 1. The continuum measured at 3mm does get a contribution from the far-infrared dust emission in the disk. This is, however, insignificant. With a dust temperature of 36 K, as derived from the IRAS 60 and 100 µm fluxes, and a flux of 320.6 Jy at 100 µm, the contribution to the continuum at 3mm is at most 2%. This is calculated assuming a dust emissivity . A more realistic dust emissivity () makes the dust emission negligble at 3mm.
The rest-frequencies of the observed molecules are given in Table 1. We used km s-1 as the systemic velocity for Cen A. All velocities are given as heliocentric with the relativistic definition of the frequency shift. This means that the center frequency of the spectrometer is given as
where v is the heliocentric velocity of the source (here 550 km s-1), c is the speed of light, the rest-frequency as given in Table 1 and the Lorentz factor . The corresponding radio and optical definitions are
The observations in December were in fact done with the LSR velocity system and using the radio definition of the redshifted frequency. These observations have been shifted to a heliocentric velocity system and using the relativistic definition. In December 1995 km s-1. The velocity difference between the radio and relativistic frequency definitions amounts to 0.504 km s-1, with the relativistic definition giving a higher sky-frequency, i.e. the frequency to which the receiver is in effect tuned to. In order to convert the spectra obtained with the radio definition to the relativistic definition, we have to add 0.504 km s-1. Hence, the data obtained in December 1995 has been shifted by a total of 2.712 km s-1. Before adding the and obtained in December with that obtained in 1996, we mapped the 1995 spectra into the same channels as the 1996 spectra.
© European Southern Observatory (ESO) 1997
Online publication: May 26, 1998