5 Summary

We have undertaken a SEST and OSO survey of CN and HNC line emission in a sample of 13 luminous IR galaxies, plus one more "normal'' starburst (NGC 1808). This survey is the first in its kind for IR luminous galaxies. The main conclusions we draw from this survey are as follows:

1.

We have detected HNC 1-0 in 9 of the galaxies, while CN 1-0 is detected in 10 galaxies. The CN 2-1 lines (J=5/2-3/2, F=7/2-5/2; J=3/2-1/2 F=5/2-3/2) (which could only be measured at SEST) are detected in 7 galaxies - in one (NGC 34) the CN 1-0 line was not detected, while the 2-1 line was.

2.

The line intensitites vary significantly from galaxy to galaxy and the line ratios relative to the "standard'' high density tracer, HCN, are far from constant. The ${{\rm HCN} \over {\rm HNC}}$ ratios vary from 1 to 6 and the ${{\rm HCN} \over {\rm CN}}$ ratios range from 0.5 to 6. From this we can learn that the actual properties of the dense gas vary significantly from galaxy to galaxy, even if their HCN luminosities are similar.

3.

We report the detection of HC₃N 10-9 emission in Arp 220 at 40% of the HNC 1-0 intensity. In NGC 1808 the line is tentatively detected.

4.

We find that the IR luminous galaxies are at least as HNC luminous as more nearby (IR-fainter) galaxies. We conclude that the HNC emission is not a reliable tracer of cold (10 K) gas in the center of luminous IR galaxies, the way it may be in clouds in the disk of the Milky Way. Instead, we list several alternative (testable) explanations. Provided that the situation in the centers of starburst galaxies can be approximated with a steady state chemistry, the observed ${{\rm HCN} \over {\rm HNC}}$ line ratios can be explained by the emission originating in gas of densities where the chemistry is dominated by ion-neutral, instead of neutral-neutral, reactions. The temperature dependence of the HNC production in the ion-neutral reactions is strongly suppressed.

5.

We were surprised to find that the CN emission of the more luminous galaxies is often faint. This can be consistent with the relative faintness of other PDR tracers such as the 158 $\mu$m [CII] line. Both these phenomena can be understood in terms of a two-component molecular ISM consisting of low density molecular gas surrounding dense clouds. However, the relative brightness of CN in Arp 220 may be difficult to reconcile with its relative [CII] faintness.

6.

Within Arp 299 (the merging pair IC 694 and NGC 3690) there is a line ratio difference between the two nuclei. NGC 3690 has a CN luminosity twice that of HCN and its ISM is thus strongly affected by UV radiation while CN is not detected towards IC 694. This reflects an evolutionary gradient in the burst - or that the star formation activity takes place in very different environments in the two galaxies. For Arp 220 there may also be an evolutionary gradient between the two nuclei because of differences in emission velocities for CN and HC₃N.

7.

The measured line ratios can in general be used to identify stages in the starburst evolution, but we conclude that the dichotomy between scenarios is a complication. For example, faint HNC emission is expected both in a shock dominated ISM as well as for a cloud ensemble dominated by dense warm gas in the very early stages of a starburst.

Acknowledgements

Many thanks to the OSO and SEST staff for their help. We are grateful to M. Walmsley and P. Schilke for discussions on the HNC chemistry. We thank the referee, F. Herpin, for many useful comments and suggestions.