1 Introduction

The polar molecule HCN (dipole moment 2.98 debye) is commonly used as a tracer of dense molecular gas, i.e. gas at $n({\rm H}_2) \geq 10^4$cm^-3. In particular in distant luminous ( $L_{\rm IR} > 10^{11}~L_{\odot}$, LIRGs) and ultraluminous ( $L_{\rm IR}> 10^{12}~L_{\odot}$, ULIRGs) systems the HCN 1-0 line is the prototypical tracer of dense gas content (e.g., Solomon et al. 1992; Helfer & Blitz 1993; Curran et al. 2000 (CAB)). Solomon et al. (1992) find a tighter correlation between FIR and HCN luminosity than the one found between FIR and CO. They suggest that, in general, the IR luminosities originate from star formation rather than AGN activity in FIR luminous galaxies. The HCN to CO intensity ratio, however, varies substantially ( ${1 \over 3}$- ${1 \over 40}$) among luminous galaxies, and it is unclear whether this difference can simply be interpreted as variations in the dense gas content or is also due to abundance and/or excitation effects. Apart from being collisionally excited, HCN may become excited via electron collisions (at $X(e) \approx 10^{-5}$) or be pumped by 14$\mu$m continuum radiation through vibrational transitions in its degenerate bending mode. It is also difficult to know if the gas is really engaged in star formation, or if it is simply dense in response to being near the central potential of the galaxy (e.g., Helfer & Blitz 1993; Aalto et al. 1995) where other mechanisms (AGN, turbulence etc.) may heat the gas and dust.

In order to understand the activities in the centers of luminous galaxies it is essential to also understand the prevailing conditions of the dense gas. Apart from observing higher transitions of HCN it is important also to study the emission from other high density tracers. One such tracer is the HNC molecule, the isomer of (and chemically linked to) HCN. For example, at high temperatures HNC can be transferred into HCN via the reaction ${\rm HNC} + {\rm H} \to {\rm HCN} + {\rm H}$. It is predicted, e.g. in (maybe oversimplified) chemical steady state models, but also by shock models, that the ${{\rm HCN} \over {\rm HNC}}$ ratio increases with increasing temperature and gas density (e.g., Schilke et al. 1992 (S92)). This is supported by the fact that the measured ${{\rm HCN} \over {\rm HNC}}$ abundance ratio is especially high in the vicinity of the hot core of Orion KL. Most of the temperature dependence is between 10 and 50 K, after which there is a considerable flattening (S92).

Compared to these results, the ${{\rm HCN} \over {\rm HNC}}$ intensity ratios found (so far) in nearby starburst galaxies are rather low (ranging from 1-5) closer to dark clouds than to hot cores (Hüttemeister et al. 1995 (H95)). This result is in apparent contradiction with the idea that the gas is warm (  K) in the centers of starburst galaxies (e.g., Wild et al. 1992; Wall et al. 1993). However, Aalto et al. (1995) suggest that the dense cores of the molecular clouds of the starburst NGC 1808 are cold (10 K) and thus these cores could be responsible for the HNC emission in NGC 1808, but possibly also in other galaxies.

The radical CN is another tracer of dense gas, with a somewhat lower (by a factor of 5) critical density than HCN. Observations of the CN emission towards the Orion A molecular complex (Rodriguez-Franco et al. 1998) show that the morphology of the CN emission is dominated by the ionization fronts of the HII regions. The authors conclude that this molecule is an excellent tracer of regions affected by UV radiation. Thus, the emission from the CN molecule should serve as a measure of the relative importance of gas in Photon Dominated Regions (PDRs).

We have searched for HNC and CN emission in a sample of LIRG and ULIRG galaxies with warm ( ) FIR colours. We were interested to see whether the HNC emission would be relatively fainter compared to the cooler, nearby objects studied by H95. Is HNC a reliable cold gas tracer, or would we find evidence for the contrary? We furthermore wanted to assess the relative importance of dense PDRs in these objects through comparing the CN line brightness with that of HCN. If indeed the HNC emission is a tracer of the amount of cold, dense gas, then perhaps an anti-correlation between the CN and HNC emission is to be expected. Many of the galaxies in the survey are powered by prodigious rates of star formation and thus a bright CN line relative to HCN is to be expected. Some of the galaxies are dominated by an AGN where the CN brightness may also be high (e.g., Krolik & Kallman 1983).

In Sect. 2, we present the observations and in Sect. 3 the results in terms of line intensities and line ratios. In Sect. 4.1 we discuss the interpretation of the HNC results and in Sect. 4.2 we discuss CN. In Sect. 4.3 possible connections to starburst evolution and scenarios of the dominating gas components are discussed.