1999A&A...348..600P

Result. of the J=1-0 HCN observations toward 34 bright far-infrared sources selected from the IRAS Point Source Catalog are reported. Together with 17 sources observed in this line earlier (Pirogov et al., 1996, A&A Transactions 11, 287) they form a complete sample of the sources with flux densities S(100µm)>500 Jy and δ> 0° in the outer Galaxy. The HCN data are compared with the HCO⁺, NH₃, CS and CO data taken from literature. Prominent correlations with nearly similar slopes of ∼1 are revealed between line integrated intensities of the molecules known to be high density tracers (HCN, HCO⁺, NH₃ and CS). The correlations become higher after adding the data for dark clouds, small globules and cirrus cores implying similar excitation and formation mechanisms of the considered molecules. Collisional excitation in regions with different densities as well as different molecular abundances and velocity dispersions in different types of cores seem to be important in producing these correlations. The following relations hold on the average over ∼3 orders of magnitude of integrated intensities: I(HCN)>I(HCO⁺)∼I(CS)>I(NH₃) where ammonia integrated intensities are several times lower than HCN ones. Correlations are also found between HCN and CO integrated intensities for the sample sources as well as between HCN line widths and those of other species. The HCN lines have the same widths as the HCO⁺ ones and are larger than CS and especially NH₃ line widths. Weak correlations are found between HCN line widths and luminosities of IRAS sources as well as between HCN integrated intensities, IRAS flux densities at 100µm and luminosities of IRAS sources divided by distance squared. The sources with most intense HCN lines have associated water masers and molecular outflows while the lack of associated maser and outflow implies weak or no HCN emission. In order to reproduce the anomalies of the J=1-0 HCN hyperfine structure (R₁₂<0.6) detected in several sources together with suprathermal line widths the model calculations are performed. Two models have been considered: a microturbulent one with a smooth density distribution and a clumpy model. It is found that in the parameter space explored a clumpy model is preferable in comparison with a microturbulent one due to the absence of self-reversals on calculated profiles and the possibility of reproducing HCN profile anomalies and suprathermal line widths. Detailed clumpy model simulations have been performed to fit the observed HCN and H¹³CN spectra toward S140 IRS1. The best fit model parameters are calculated, including radial dependencies of clump density and volume filling factor.