Annu. Rev. Astron. Astrophys. 1991. 29: 581-625 Copyright © 1991 by Annual Reviews. All rights reserved

4. SPIRAL STRUCTURE IN M51

Ever since the earliest observations of molecular clouds in galaxies, there has been great interest in determining the extent to which the molecular clouds and their properties are correlated with spiral arms. Density waves, believed to be responsible for the spiral pattern in some galaxies, might provide a mechanism for molecular cloud growth either through gravitational instabilities or by increasing the incidence of cloud-cloud collisions and coalescence. Numerous studies have shown, however, that the molecular gas is widespread and not confined solely to the spiral arms. Although the early CO observations employing 45 to 66" beams (Rickard & Palmer 1981, Young & Scoville 1982a, Scoville & Young 1983) had resolution only marginally sufficient to separate the arm and interarm regions, recent higher resolution studies have confirmed that although significant spiral arm concentrations do exist, most of the molecular emission originates from interarm regions.

Perhaps the best example of a galaxy in which to investigate the extent to which the molecules are confined to spiral arms is M51. In this grand design spiral the exceptionally strong arms probably result from both density waves (in the inner disk) and the tidal interaction with NGC 5195 (in the outer disk). At 33" resolution, i.e. 1.6 kpc, the CO enhancement is 20% in the arms and the underlying axisymmetric, exponential disk contributes ~ 75% of the total emission (Rydbeck et al 1985).

Aperture synthesis mapping of the CO in M51 at 7" resolution has been reported by Lo et al (1987a), Vogel et al (1988), and Rand & Kulkarni (1990). The interferometric data (see Figure 3) show striking concentrations of CO emission along the dust arms, displaced on average by approximately 7" (300 pc) to the inside (upstream direction) of the arms seen in H. The masses of these molecular cloud associations are 10⁷ - 6 x 10⁷ M. The CO emission maxima coincide precisely with regions of enhanced dust obscuration in optical continuum images. Approximately 25% of the total CO emission is contained in the arm-like structures shown in Figure 3, with an arm-interarm contrast of approximately 3:1 averaged over scales of 500 pc (Vogel et al 1988, Guelin et al 1988). Discrete emission concentrations are also seen in the areas between the arms, although they are not so well organized into coherent structures (Rand & Kulkarni 1990). It is noteworthy that the spiral arms seen in nonthermal radio continuum are aligned most closely with the CO ridges shown in Figure 3, whereas the 21 cm H I emission concentrations are displaced downstream near the H arm (Rand & Tilanus 1990). Presumably, this HI arises from dissociation of molecular gas by young stars formed in the arms. Even in the downstream areas, the single dish data indicate that the surface density of molecular gas still exceeds that of atomic gas. Thus, the molecular gas is not significantly dissociated by the OB star formation, but is instead distributed into smaller structures (e.g. individual GMCs) that are less easily detected by the interferometer.

Figure 3. Mosaic map of the CO emission at 7" (350 pc) resolution in M51 superposed on gray scale images of the red continuum (upper left), H (upper right), and 21 cm H I emission (lower left) (Rand & Kulkarni (1990). On the lower right, the H I emission contours (Rots et al 1990) are superposed on a gray scale image of H. The spiral arm concentrations of CO emission are closely aligned with areas of increased dust obscuration seen in the red continuum images (upper left), but displaced approximately 300 pc upstream of the spiral arm H II regions (upper right) and H I gas (lower left). The CO interferometry recovers approximately 30% of the total CO line flux: the remaining molecular emission is more uniformly distributed on scales 25".

The CO studies in M51 suggest a picture in which pre-existing GMCs come into the arm from the back side, concentrate in large cloud complexes as a result of orbit crowding or gravitational instabilities in the potential well of the spiral arm, and form high mass stars slightly downstream of the potential minimum. The OB stars then dissociate some of the molecular gas to produce the H and H I ridges downstream from the molecular peak. The disappearance of the molecular cloud complexes on the downstream side of the spiral arms could result from expected divergence of the cloud orbits as they come out of the spiral potential or the disruptive effects of high mass star formation. Inasmuch as the quantity of atomic and ionized gas in the downstream direction is significantly less than the abundance of molecular gas at the spiral ridge, the former may be the primary cause.

The extent to which the M51 results may be applied to other spiral galaxies, in particular to those with lower gas densities or those with smaller spiral arm amplitudes, is not at all clear. All of the processes discussed above (orbit crowding, gravitational instability, and cloud disruption as a result of OB star formation) are likely to be most significant in a galaxy like M51 with a strong spiral pattern, a high interstellar gas density, and a high rate of OB star formation. Significant downstream offsets of the H I spiral arms from the dust lanes have been noted by Allen et al (1986) for the barred spiral galaxy M83; on the other hand, Lada et al (1988) found a close correlation between the atomic and molecular ridges in M31 in both velocity and space (see Section 5.1). We therefore caution that the results in either M51 or M31 should not be taken as a general rule; instead, these two situations may represent gas-rich and gas-poor galaxies, respectively.