DIFFUSE MOLECULAR CLOUDS AND THE MOLECULAR INTERSTELLAR-MEDIUM FROM (CO)-C-13 OBSERVATIONS OF M33

We have obtained (CO)-C-12 and (CO)-C-13 J = 1-0 observations of the n earby spiral galaxy M33 to try to resolve the long-standing discrepanc y between (CO)-C-12/(CO)-C-13 line ratios measured in Galactic giant m olecular clouds and external galaxies. Interferometer maps of the mole cular cloud MC 20 give a (CO)-C-12/(CO)-C-13 line ratio of 7.5 +/- 2.1 , which agrees reasonably well with the line ratio measured in Galacti c giant molecular clouds. In contrast, the (CO)-C-12/(CO)-C-13 line ra tio obtained from single-dish data is 10.0 +/- 0.9, significantly high er than Galactic values but in good agreement with line ratios measure d in other galaxies. The interferometer map of MC 20 reveals that the cloud has similar spatial and velocity extents in the two lines, and t hus the high single-dish line ratio cannot be due to different filling factors in the two lines. In addition, the single-dish data show no e vidence for significant variations in the line ratio with metallicity, which eliminates abundance changes as the explanation for the high si ngle-dish line ratio. We conclude that the high (CO)-C-12/(CO)-C-13 li ne ratios observed in M33, and in the disks of spiral galaxies in gene ral, are due to the presence of a population of diffuse molecular clou ds. We obtain a lower limit to the (CO)-C-12/(CO)-C-13 line ratio in t he diffuse molecular clouds of 13 +/- 5. The lower limit to the fracti on of the total (CO)-C-12 emission from M33 that originates in the dif fuse clouds is 30 +/- 30%, while the upper limit is -60%. We have comb ined our line ratios with published measurements of the (CO)-C-12 J = 2-1 to J = 1-0 line ratio to determine the (CO)-C-12 and (CO)-C-13 col umn densities to within an order of magnitude, but the density and the temperature of the gas are not well constrained by these measurements . We estimate that differences in the physical conditions in diffuse a nd giant molecular clouds are unlikely to cause the overall molecular gas mass calculated for the central kiloparsec of M33 to be in error b y more than a factor of 2.