|Annu. Rev. Astron. Astrophys. 1982. 20:
Copyright © 1982 by . All rights reserved
4.4. Correlations of Molecular Sources with Nuclear Activity
The first group of galaxies for which CO was detected tended to be "active" galaxies, in the sense that their central regions are strong emitters of infrared and radio continuum radiation and optical emission lines, and in some cases show evidence of noncircular motions or mass ejection (Rickard et al. 1977a, c). While these conditions have not proven sufficient to distinguish class 1 galaxies from other types, they are manifest in correlations among the class 1 galaxies themselves.
The best established correlation links integrated emission in CO with the far-infrared flux. As the characteristics of infrared continuum emission from the nuclei of spiral galaxies have been reviewed quite recently (Rieke & Lebofsky 1979), we confine ourselves to discussing the specific associations between the available far-infrared and molecular line data.
In an analysis of the first CO survey, Rickard et al. (1977a) suggested that the central CO emission was correlated with the 10µ flux detected over the same area. The original relationship was based on little data (including an incorrect 10µ flux for M51), and was subject to some inconsistencies (such as the absence of CO emission from the strong 10µ source NGC 5195). It was supported, however, by theoretical analyses (Jones et al. 1977). Telesco & Harper (1980) showed that the correlation improved at longer wavelengths.
Recently, Rickard & Harvey (in preparation) have collected a large enough sample of 100µ observations, with aperture sizes comparable to the CO beam sizes, to evaluate this statistically.
The relation between the 100µ flux and P(CO) for 14 galaxies in which both are measured is shown in Figure 5. The flux-flux correlation is quite good. It should be noted that there is some tendency for correlation between both CO and IR luminosity and the optical luminosity of the whole galaxy. However, this is principally due to the fact that NGC 1068 is quite luminous in all its characteristics. In any case, the CO vs. IR correlation persists (at the 90% confidence level) when both quantities are scaled to the optical luminosity.
Figure 5. The correlation of integrated 12CO intensity with (a) total galactic radio emission at 21-cm and (b) central far-infrared flux at 100µ data taken from Rickard et al. (in preparation).
The interpretation of this relationship is not entirely straightforward. The dust (which is responsible for the 100µ radiation) appears to be heated primarily by the luminous energy released in the formation of massive stars (Gatley et al. 1978, Becklin et al. 1980), whereas the H2 may be heated primarily by cosmic rays (Gusten et al. 1981, Morris et al., in preparation). Furthermore, it would be somewhat surprising if the gas/dust ratio were roughly constant in the central regions of all the galaxies surveyed; this consideration would be unimportant only in the unlikely case that both the CO and the 100µ emission are optically thick. The simplest interpretation of the CO-100µ correlation is that the emitting regions are coextensive (i.e. almost all the dust is in molecular clouds), that the ratio of 100µ and CO surface fluxes from an average cloud is approximately constant (for reasons not apparent), and that the intensity differences amongst the central regions of different galaxies essentially reflect different beam-filling factors for molecular clouds. This interpretation presumes that cloud temperatures play only a secondary role in determining the CO-to-100µ. intensity ratio, an assumption supported by the fact that the 50-160µ color temperatures do not correlate with the CO (Rickard & Harvey, in preparation).
The preferential detection of OH absorption toward galaxies with strong nonthermal sources is probably a selection effect. The CO emission should be free of this effect, however. As shown in Figure 5, there is an excellent correlation between the central CO flux and the total nonthermal radio flux, which persists if both are scaled to the optical luminosity. In addition, there is a tentative correlation (at the 90% confidence level) of the CO luminosity/optical luminosity ratio with the fractional concentration of non-thermal radio flux within the nuclear region (Rickard et al., in preparation).