Astron. Astrophys. 322, 962-974 (1997)

3. Large-scale emission patterns of CH, OH, and CO

The large-scale molecular emission pattern immediately around Oph has heretofore been studied only in ¹² CO (Liszt 1992, 1993; Kopp et al. 1996). There is a peculiar behaviour by which one or other of the kinematic components seen toward the star at -0.7 km s^-1 and + 1 km s^-1 brightens considerably within (projected distance 1 pc) in the North-South direction while less structure is evident East-West. Elements of this pattern are repeated in OH and CH, although with some striking variations. The pattern is also evident in HCO emission, as discussed in Sect. 4.

3.1. CH emission and excitation

The standard formula relating the CH column density to observable parameters at 3335MHz is (Mattila 1986)

where (K km s^-1)^-1. However, the most accurate value of N(CH) is that determined optically toward the star (Lien 1984), log N(CH) = 13.36. It is currently believed possible to relate the CH integrated intensity directly to extinction (Mattila 1986; Magnani and Onello 1995). The mean relation of Mattila (1986) determined toward two dark clouds, N(CH) = ( -0.3) (for = (4/3.1) ), fits the current data quite well; the threshold of 0.3 mag for CH formation corresponds to the long-known sudden increase in N() as it becomes self-shielding. In this case, the small variation in across the face of the Oph gas distribution (Table 1) is evidence that the wide swings in CO and HCO intensity are excitation and chemical effects within a nearly fixed total gas column. CH peaks in the higher-velocity line to the South of the star, as does CO, but emits most strongly to the North, like OH (Figs. 1-3).

Lien (1984) attempted to derive the CH excitation temperature directly by comparing column densities in absorption; he also compared the optical and radio-determined column densities to derive the excitation temperature. Unfortunately his results were somewhat equivocal, in part owing to his use of a smaller integrated line intensity than found here and perhaps because he attempted to derive the beam efficiency in a rather circuitous manner. For K km s^-1 toward Oph (Table 1) and using equation (1) it is possible to relate the unknown excitation temperature and beam efficiency. The locus for CH runs between ( (CH) = -2.92 K) and ( (CH) = -16.5 K). Thus the -doublet in CH appears to be inverted, and it, like OH and CO, cannot be employed as a thermometer. However, this result is unexpected because the usual understanding of negative excitation temperatures in CH invokes neutral-particle collisions (Bertojo, Cheung, and Townes 1976), while the excitation of CH toward Oph should be dominated by electrons, as discussed for OH in the next section.

3.2. OH emission and excitation

The resolution OH spectra are shown in Figs. 1-3. Surprisingly, we were unable to detect a 4.5 km s^-1 component toward the star. Given the appearance of this feature in the one earlier work where it was seen (Crutcher 1979), we are inclined to regard it as spurious. We know of nothing in our observing which might have precluded its detection (it was not corrupted by our frequency-switching interval). It seems clear from Crutcher's spectra that strongly OH-emitting gas at +4.5 km s^-1 exists along lines of sight which are well removed from the star, but, apparently, not near it; there is a modest atomic absorption component toward Oph (Hobbs 1969) at this velocity.

As noted by Black (1995, private communication) the supposed OH column density in this material was large enough (N(OH) ) that the absence of a stronger 4.5 km s^-1 component in optical absorption spectra was something of a mystery. The 4.5 km s^-1 emission was ascribed by Crutcher (1979) to gas which had been shocked by a stellar outflow, giving impetus to the study of CH formation mechanisms driven by interstellar shocks. Such models are no longer believed capable of providing the values N(CH ) which are commonly observed (Allen 1994; Barlow et al. 1995) and the absence of the putative pre-shock gas probably does not represent a real impediment to our understanding. The most recent models of CH formation invoke energy dissipation which occurs in turbulent (diffuse) clouds having moderate density and N(C ) N(CO) (Falgarone, Pineau des Forêts, & Roueff 1995; Hogerheijde et al. 1995; Federman et al. 1996).

Another feature which is not well-represented in the OH emission spectra is the strong, higher-velocity component of the CO and HCO distribution, especially near its peak to the South of Oph. Although the lower-velocity line shows about the same behaviour in CO and OH, its counterpart is manifested in OH only near Oph. OH in the higher-velocity feature seems limb-brightened, suggesting that it exists or is excited only on the periphery of the host gas. A more extended map of the OH might detect OH emission at the southern edge of the gas as well.

The column density of OH is related to other quantities as in equation 1, substituting (K km s^-1)^-1 from Dickey, Crovisier, and Kazès (1981). For observations toward the star, reconciliation of the observed 18 cm OH emission intensity (0.103 K km s^-1 / ) and the optical absorption-line column density () requires a combination of beam efficiency, excitation temperature, and optical depth which lies on a locus whose limits are ( (OH) = 16.1 K, (OH) = 0.007) and ( (OH) = 5.3 K, (OH) = 0.02); unlike CO there is no strong central minimum in the OH emission distribution, so that seems unlikely. The lower of these excitation temperatures is comparable to for CO. Roueff (1996) derived (OH) K by comparing OH lines in absorption as Lien (1984) did for CH. She attributed this result to pumping by IR radiation, since it is so much lower than any possible kinetic temperature (see just below).

The electron density (Savage, Cardelli, and Sofia 1992) derived from the ionization equilibrium of Mg and Fe is n(e) = 0.046 and the ratio of electron and radiative de-excitation rates across the ground-state - doublet is (Bouloy and Omont 1978). If these electrons are mainly contributed by carbon, and the analogous ratio for neutral particle excitation is of order unity. According to this line of argument OH should be a good thermometer, indicating 5 K 16 K in the neutral-bearing region toward the star. This represents a paradox because a gas with its carbon fully ionized is unlikely to be this cold and densities are far too small to excite CO at such very low temperatures. CO requires an n()- product which must at least be of order  K , (see Sect. 5).

There are now many cases where an expected high degree of electron excitation is simply not manifested in OH: its interstellar excitation temperature is measured to be within 0.5-1.0 K of the cosmic background in diffuse and translucent regions where electron and even neutral particle excitation should be substantial (Dickey, Kazès, and Crovisier 1981; Liszt and Lucas 1996). The anomalously low OH excitation temperatures seem to occur in the regime of the C CO transition (ibid). The factor in equation (1) can be very small for such weak excitation and, although it is possible that the OH abundance is really very low at the CO and HCO peak to the South, this does not necessarily follow from the OH emission spectra. For column densities comparable to those seen around Oph, N(OH) and N(HCO ) measured in absorption at radiofrequencies are very tightly related, with N(OH)/N(HCO ) (Lucas and Liszt 1996; Liszt and Lucas 1996); it is only OH emission that is weak, not N(OH) that is low.

The model of Kopp et al. (1996) predicts an undetectably low abundance N(OH) in the southerly gas, seemingly in accord with our data. Yet this situation is intimately related to their model's prediction N(HCO ) , which they note is nearly a factor 1000 too low (see Liszt and Lucas 1994 and Sect. 4.1 here); HCO should be made via the reaction of C + OH (Black and van Dishoeck 1986). It remains to be seen why the abundance of OH would be 100 times larger to the North, where its emission is easily seen, while the CO and HCO are only slightly weaker there. But a problem remains either with the chemistry of OH or with its excitation.

3.3. Comparison of OH, CH and CO emission

3.3.1. CH and CO

Fig. 4 shows line profile integrals for these species taken separately over positive and negative velocities; upward pointing symbols are used to denote those positions at or to the North of Oph. In the left-most panel, the CO line brightness is seen to increase greatly over narrow ranges of the CH profile integral. If the CH emission is a good surrogate for the column density N(H) (which CO most certainly is not), this behaviour can probably be understood as the rapid onset of self-shielding (van Dishoeck and Black 1986, 1988; Kopp et al. 1996) which accompanies the C CO transition. It seems to occur in two branches for the Northern and Southern gas, perhaps related to differing positions with respect to the the star (or the ambient uv flux in general), and there is an element of anti-correlation as well; the branch with stronger CH emission is somewhat weaker in CO. A particularly vivid example of the rapid rise expected of the CO intensity with changing N(H) or is given by Kopp et al. (1996), whose Fig. 9 greatly resembles Fig. 4 here.

Similar behaviour is apparent in comparisons of CO with HCO seen in absorption toward extragalactic continuum sources (Lucas and Liszt 1996), at N(HCO ) 1-2 . Here the analogous behaviour occurs at K km s^-1. If we scale from the observations toward Oph, in which K km s^-1 corresponds to N(CH) , the CO turn-on occurs at N(CH) , with N(CH)/N(HCO ) 8. Federman et al. (1994) show that higher values of N(CO) may occur at N(CH) in absorption spectra.

3.3.2. CO and OH

This is shown in the middle panel of the Fig. 4 triptych. Those points corresponding to the weaker branches of the CO emission ( v 0 km s^-1 to the South and v 0 km s^-1 to the North) are shown greyed. For them, and for most of the data, the range of OH profile integrals is scarcely greater than the expected noise envelope while the CO integral varies by nearly a factor of 20. The remainder of the points, those corresponding to the dominant branch of the CO emission, appear to exhibit the inverse relationship which hides the strongest CO lines in OH. As discussed next, distinguishing between the weaker and stronger CO branches is necessary to understanding the relationships between CH and OH, which may be correlated or anticorrelated.

3.3.3. CH and OH

Overall, there may seem to be little apparent relationship between the line profile integrals of these species: with the possible exception of 1 outlying point, the CH-OH comparison at the right in Fig. 4 at first shows only scatter. This is due to the superposition of data showing two opposite kinds of behaviour. The points corresponding to the weaker CO component at each position, shown as the greyed, upward-filled and downward-open triangles at intermediate and smaller , show a decline in OH as CH strengthens. Those points belonging to the dominant CO emission branches ( v 0 km s^-1 to the South and v 0 km s^-1 to the North) show a rapid increase in the OH brightness with increasing , but in such a way as to preserve the inverse relationship between OH and CO. For that subset of points showing this rapid growth of OH, the strongest CO lines are those with weaker CH (see 3.3.1).

In 3.2 we noted that N(OH)/N(HCO ) as seen in absorption in the radio regime while N(CH)/N(HCO ) is inferred from the existence of a rapid increase in N(CO) with either N(CH) or N(HCO ). The ratio which results from the comparisons with HCO , N(OH)/N(CH) , is about what is seen toward Oph optically, N(OH)/N(CH) .

© European Southern Observatory (ESO) 1997

Online publication: June 5, 1998

helpdesk.link@springer.de