Astron. Astrophys. 334, 646-658 (1998)

5. Discussion

5.1. Cloud properties as a tracer of large scale dynamics

We will now explore how the relative SiO abundances may be related to the position of the cloud in the large scale Galactic center environment.

In Fig. 4 we plot the the beam averaged SiO abundance, , on a longitude-velocity diagram of the large scale distribution of . While any estimates of line-of-sight locations in the Galactic center are rather uncertain, this plot gives a measure of location.

Fig. 4. The beam averaged SiO abundance plotted on a longitude- velocity diagram of the large scale distribution of , adapted from Paper I (spatial resolution ).

There is a clear trend in the large-scale distribution of SiO abundances. For , between the location of the Sgr C and Sgr B2 regions, we find a pronounced lack of very large X (SiO). Most clouds with very high abundances, including the exceptional source M+1.31-0.13, are located at . The `Clump 2' Region is not included in this plot, since, at , this is not really part of the continuous molecular bulge or, in the notation of Morris & Serabyn (1996), the Central Molecular Zone (CMZ), although it shares many characteristics of the Galactic center gas. in this region ranges from 1 10<sup>-9</sup> to 2 10<sup>-9</sup>, typical `Galactic center' values that are high compared to the disk but not as extreme as what is found at .

The detection of an `SiO hole' at is confirmed by an extension of the survey of Martín-Pintado et al. (1997) (unpublished data).

In recent years, the large scale dynamics of the gas in the CMZ, characterized by large non-circular motions and a distinct `parallelogram' shape of lv -diagrams based on ¹² CO and ¹³ CO (e.g. Bally et al. 1988), have been explained in terms of a model involving a rotating (stellar) bar with corotation at 2.4 kpc and oriented at an angle of with the line of sight to the Galactic center (see Morris & Serabyn 1996 and references therein). Within such a potential, there exist two categories of closed elliptical orbits, called and -type orbits. Inside a cusped orbit, the -orbits, which are elongated along the bar axis, become self-intersecting. Clouds on these orbits encounter a shock and within a dynamical time plunge into -orbits that lie considerably deeper within the potential and mimic circular orbits (Mulder & Liem 1986, Athanassoula 1988, 1992). This encounter breaks the flow along the cusped orbit into a spray that spreads out into the interior of the orbit (Binney et al. 1991, Athanassoula 1992, Jenkins & Binney 1994).

According to this model, Sgr B2 and Sgr C can be interpreted to be at the locations of the intersections of the - and the -orbits. The dense, virialized clouds on the -orbits, inside of Sgr B2 and Sgr C, are more likely to form stars, while the less dense clouds on the orbits are too disrupted by shocks for efficient star formation. Generally, this is supported by the fact that stars of young and intermediate age are restricted to the area inside of Sgr B2 and Sgr C, suggesting sustained star formation in this region (Serabyn & Morris 1996). In Sgr B2, the ongoing star formation, the large amount of molecular gas, the existence of a very hot molecular component (Hüttemeister et al. 1993a, 1995, Flower et al. 1995) and the high optical depth of the ¹² CO emission (Paper II) also provide ample evidence for this picture. The Sgr C complex is a much more quiescent region, but it contains a large H II region and GMCs.

Because the sprayed gas crashes into material that is still on -orbits, collision regions are expected along the acceleration part of the cusped orbits. Due to perspective, these main collision areas should be at higher positive longitudes than Sgr B2 and between Sgr A and Sgr C. While the situation at negative l is confused (the gas on orbits and the collision regions are expected at similar Galactic longitudes), at positive l, especially the ` -complex' (Bally et al. 1988) shows all characteristics expected from a collision region (see the discussion in Paper II).

Inspecting Fig. 4 with this picture in mind, we find remarkable support from our data. The exceptional cloud M+1.31-0.13 and three other sources with high SiO abundances are located within the 1.5 -complex, exactly where strong shocks are expected. Generally, the high SiO abundances at can be identified with a region in which the probability of gas encountering shocks is high. Thus, for the first time, we can tie the physical and chemical parameters of Galactic center molecular peaks to the large scale dynamics of the region.

Since CS traces all dense gas, not just the part that has been subjected to shocks, we do not expect a similar `zone of avoidance'. Indeed, a plot similar to Fig 4comparing the CS abundance in these clouds (data presented in Hüttemeister 1993), to the total H₂ column density does not show such an effect.

Of course, the correspondence is not absolute. Not every cloud at must experience strong shocks. Note, however, that the clouds with very low SiO abundances in this range of l have negative velocities and might not be located in the collision region. Also, there are certainly strong shocks occuring in the Galactic center that are unrelated to large scale dynamics. Martín-Pintado et al. (1997) explain the `SiO clouds' they find by shocks of a variety of origins: Interaction with SNRs close to Sgr A, interaction with non-thermal filaments in the radio arc and cloud-cloud collisions (as expected from large scale dynamics) or expanding bubbles in the vicinity of Sgr B2. Their map extends from to . Therefore, they have missed the systematic signature of large scale effects present at higher positive longitudes.

Another mechanism explaining the preferential occurence of shocks in the -complex might be fossil superbubbles, remnants of a phase of Sgr B2 type star formation activity. The expansion of superbubbles would offer a natural explanation of the large extent of the CO emission to positive Galactic latitudes perpendicular to the Glactic plane seen in this region. However, while the bar model as outlined above is two-dimensional and thus does not naturally produce vertical structure, sprayed gas colliding with gas on -orbits might give rise to turbulence also pushing gas out of the plane. It is noteworthy that a ¹² CO map (see, e.g., Paper II and Bitran et al. 1997) shows the molecular gas extending toward negative b at the southern edge of the CMZ. Thus, it seems possible that the CMZ is warped or susceptible to instabilities close to its edges (Morris & Serabyn 1996 and references therein).

The `bar model' is not the only possibility to explain the large scale dynamics of the Galactic center region. von Linden et al. (1993) suggest that an accretion disk can also reproduce the basic structure seen in position-velocity plots. While this is certainly true for the inner region ( -orbits are almost indistinguishable from circular orbits) it is not clear whether such a scenario can also explain the gas distribution and the different chemical properties in the entire CMZ. Based on the data presented in Papers I and II and this work, we will examine the questions related to large scale structure and dynamics closely in a forthcoming paper (von Linden et al. 1998).

5.2. The origin of SiO

Even though the SiO emission we observe is mostly associated with cool gas, we consider it likely that it forms under conditions of high kinetic temperature and grain erosion by shocks (Martín-Pintado et al. 1992). Both the distribution discussed in the previous section and the existence of the hot cloud core M+1.31-0.13 can be taken as evidence for this. From the fact that only one cloud in our sample, namely M+1.31-0.13, shows hot gas emitting in SiO lines, we can conclude that shocked gas forms dense, cooler cores fairly fast. Extrapolating from the analysis in Hollenbach (1988), we find a very short cooling time scale ( yr) for gas at a density of 10⁴ . A competing process is the condensation of gas phase SiO onto grain mantles. We estimate this `freeze out' time scale to be of order yr or slightly less (Rohlfs & Wilson 1996). Thus, gas with a high gas phase relative abundance of SiO has been cool for most of its lifetime.

After more than 10⁶ years, the gas slowly loses the chemical memory of having been shocked. From the typical radial velocities of the clouds, we estimate that, during this time, these can move pc. If the very SiO-rich clouds are preferentially located on -orbits along the bar, these will remain sufficiently close in longitude to the region where they encountered the shock to be recognised as a distinct population.

Since we find some SiO in all cloud cores, additional formation of gas phase SiO is required, probably by local turbulence and/or cloud-cloud collisions causing some shocks. This is a general characteristic of the cores of Galactic center GMCs. Ongoing local shock activity is likely to be also necessary as the main heating mechanism for the hot gas component seen in all clouds (see Flower et al. 1995 and the discussion in Hüttemeister et al. 1993).

It is known that gas in post-shock regions, away from chemical equilibrium, is characterized by abundant complex molecules (e.g. Brown et al. 1988). This is exactly what can be observed toward the positions within the CMZ that are strong in SiO. We show just one example in Fig. 5. In the Galactic disk, such complex spectra are typical for very confined regions, while in the CMZ they seem to be ubiquitous. This chemical complexity can only be maintained for years and thus requires frequent shocks. These may also be typical for the starburst environment (see Henkel et al. 1987, Mauersberger & Henkel (1991), Mauersberger et al. 1991, Hüttemeister et al. 1997 for the case of NGC 253). On larger timescales, an equilibrium state is reached. This consists of a gas phase component of mostly diatomic molecules (Herbst & Leung 1989).

Fig. 5. The chemical complexity of sources strong in SiO is demonstrated by our spectrum of M+0.83-0.18, covering GHz and including both 29 SiO and 30 SiO. The spectrum shown is a composite of two single SEST low frequency resolution spectra.

Most time-dependent chemical models have been calculated for either cold, quiescent clouds (e.g. Bergin et al. 1995) or 'classical' collapsing, star-forming hot cores (e.g. Brown et al. 1988, Caselli et al. 1993). Thus, detailed, model-supported recommendations of the molecules that should be used in future work to trace the chemical evolution of the the Galactic center cores cannot be given. CH₃ OH and SO₂ appaer, however, to be promising species. The former is a high temperature, high density tracer requiring 70 K to evaporate from dust grains. Its wealth of emission lines allows density and temperature determinations from multi-level studies. Observations in external galaxies show that its abundance can differ widely even in the central regions of galaxies (Hüttemeister et al. 1997). SO₂ is associated with shock chemistry and grain destruction. The abundance of the two molecules need not be correlated, as is shown by an evolutionary study of cores in the W3 region (Helmich et al. 1994). Both molecules have lines that are strong enough to allow large scale mapping in the Galactic center region, and differences in their distribution would give further insight into the predominant processes operating on the Galactic center cores.

© European Southern Observatory (ESO) 1998

Online publication: May 15, 1998

helpdesk.link@springer.de