MOLECULAR GAS IN STARBURST GALAXIES - LINE-INTENSITIES AND PHYSICAL CONDITIONS

In a study of molecular gas in IR-bright galaxies, we have observed th e central position in (CO)-C-12 J=1-0, J=2-1; (CO)-C-13 J=1-0, J=2-1; (CO)-O-18 J=1-0 and HCN J=1-0 line emission. The sample (in total 32 g alaxies) consists of starburst galaxies, interacting galaxies and two quiescent systems. We find a mean (CO)-C-12 2-1/1-0 line ratio of 0.93 +/-0.22, and a mean (CO)-C-13 2-1/1-0 line ratio of 1.3+/-0.66. The me an (CO)-C-12/(CO)-C-13 1-0 ratio (=R((1-0)) is approximate to 13 +/- 6 and for (CO)-C-12/(CO)-C-13 2-1 it is approximate to 13 +/- 5. The me an (CO)-C-12/HCN 1-0 ratio is approximate to 16. The (CO)-C-12/(CO)-O- 18 1-0 intensity ratio ranges from 20 to 140 in 6 galaxies. These valu es of ratios refer to central positions corrected for effects of beam- size and source-size. A considerable fraction of the (CO)-C-13 emissio n may be saturated and originate in dense cores of smaller volume fill ing factor than the surrounding (CO)-C-12-emitting gas. High gas tempe ratures and turbulence work against large optical depths in the (CO)-C -12 1-0 line of the envelope gas. In contrast to what is usually assum ed for the (CO)-C-12 1-0 line, we infer moderate optical depths, tau a pproximate to 1, for the dominant (CO)-C-12-emitting structure. We hav e found that galaxies with large intensity ratios of [C II] 158 mu m t o (CO)-C-12 1-0 also have (CO)-C-12 2-1/1-0 ratios greater than or equ al to 0.8, implying that the (CO)-C-12-emitting gas may be both dense and hot enough to excite the [C II] 158 mu m line, a characteristic of warm (T-k greater than or equal to 100 K), photon-dominated regions. We have not found a correlation between the intensity ratio, I(CO)/I(H CN), and disturbed morphology, far-infrared emission, or measures of s tar-forming activity in our sample of galaxies, in contrast to the fin dings of Solomon et al. (1992). The (CO)-C-12/HCN intensity ratio is q uite uniform over a sample of 11 interacting galaxies and mergers. HCN is not detected in the one isolated spiral in the sample. Furthermore , we have observed R((1-0)) in off-centre positions in 6 galaxies: NGC 1808, NGC 3256, NGC 4038/39, NGC 5055, NGC 6221 and NGC 7552. Offset positions in the (CO)-C-12/(CO)-C-13 2-1 intensity ratio have been obs erved in 3 galaxies: NGC 660, NGC 2146 and Arp 299. We suggest that ga laxies that display significant variation in the ratio, such as the me rger NGC 3256, have at least two populations of molecular cloud ensemb les: high-pressure clouds in the centre and an extended disk-component of more quiescent clouds. The molecular gas in the centres of luminou s mergers with large gas surface densities (> 10(4) M. pc(-2)) will be highly turbulent. In general, we suggest that R((1-0)) is a measure o f the cloud environment: the extreme values R(1-0) > 20 originate in t urbulent, high-pressure gas in the centres of luminous mergers; interm ediate ratios 10 less than or similar to R((1-0)) less than or similar to 15 originate in the inner kpc of more normal starburst galaxies; s mall ratios R((1-0)) approximate to 6 are a signature of a disk popula tion of clouds. We address the notion that abundance anomalies cause t he elevated values of R((1-0)) in luminous merging galaxies. To some e xtent, this notion rests on the assumption of tau >> 1 in the (CO)-C-1 2 1-0 line. We demonstrate how this assumption may be flawed and concl ude that in order to measure abundances in the molecular medium, one m ust carefully model the physical properties of the molecular gas. Alth ough abundances may be unusual in extreme mergers, they are only one a spect of the properties of the molecular gas in these regions.