3.1.1. Structure and kinematics
In Fig. 1 we present the interferometric maps of HCO+(1-0) for the different LSR velocity channels where emission has been detected. In Fig. 2, a map of the HCO+ velocity integrated intensity is shown. The HCO+ emission arises from two opposite, narrow lobes oriented at PA 21o, the symmetry axis of the optical nebula (Sect. 1). Near the center, these lobes are hollow (Fig. 1), with a characteristic radius of ( 5.5 1016 cm). In the central part of the nebula, the HCO+ emission seems to be slightly more extended in the equatorial direction than that of 12CO (see Fig. 1 and the CO velocity maps in Paper III) and other molecules (see e.g. SO data in Sect. 3.2). In the direction of the axis, the HCO+ emitting region has a total (deprojected) extent of 7 1017 cm and shows a remarkable asymmetry with respect to the equator, the north lobe being almost twice as extended as the southern one. The equatorial asymmetry is also present in the 12CO nebula, but in this case, the south lobe is the largest of the two (with a total, deprojected length of 1.5 1018 cm; see Fig. 2). However, we note that the extent of the detected HCO+ emission is limited to the interferometer field of view, that in the case of the present observations is 60", in contrast to 100" for the 12CO mosaic (Paper III). Therefore, at the southern tip of the 12CO nebula (at 30" from the center) the HCO+ emission would be underestimated by a factor 2. On the other hand, the HCO+ intensity in the north clumps at 15" from the center is underestimated by 15. The axial asymmetry of the HCO+ emission is also noticeable. In fact, most of the emission arises from the west side of the nebula.
Our high resolution observations reveal the remarkable clumpiness of the HCO+ emission. Note, for instance, the intense northern clump flowing at an LSR velocity 50 km s-1, coincident with the outermost northern CO clump, and the southern clumps at 70 and 105 km s-1. Note also that the maximum HCO+ emission (Fig. 2) arises from a relatively compact clump that is displaced from the nebula center.
The HCO+ emission shows a clear axial velocity gradient that appears in both the intensity maps per velocity (Fig. 1) and in the position-velocity (p-v) diagram along the nebula axis (direction "a" in Fig. 2) represented in Fig. 3. In particular, the emission from the southern (northern) clumps is red- (blue-) shifted with respect to the systemic velocity ( 33 km s-1). This velocity gradient is similar to that found for 12CO and other molecules (see below and Sánchez Contreras et al. 1997), indicating that HCO+ shares the overall axial expansion of the molecular envelope. In Fig. 3, p-v cuts along other selected directions ("b", "c", and "d" in Fig. 2) are given. The p-v diagram along the nebula equator (direction "b") reveals some expansion (at velocity 30 km s-1) in the direction perpendicular to the major axis of the nebula. We note that the figure given above for the expansion velocity can be affected by the mixing within our beam of regions axially expanding. Some expansion (in the direction perpendicular to the nebular axis) is also present at higher latitudes, as the p-v cut along the direction "c" indicates. The expansion velocity derived from this diagram is 15 km s-1. The axial asymmetry of the north lobe can be also appreciated from the previous diagram. In fact, only HCO+ emission from the east side is found, suggesting that at these latitudes the lobe appears as a half-shell. In Fig. 3 (bottom-left panel) the p-v cut along the line joining the two central maxima of the HCO+ emission (Fig. 2, direction "d") is plotted. This cut reveals a large velocity dispersion ( 80 km s-1) in the relatively compact region from which the maximum HCO+ arises.
3.1.2. Excitation and abundance variations of HCO+
Probably the most remarkable aspect of the HCO+ emission is the presence of an intensity dip at low velocities (between 10-55 km s-1) near the nebula center, where the rest of the observed molecular species reach their maximum intensity (for SO data see Figs. 5 and 6, and for CO see Paper III). In fact, the most intense emission of this molecule arises from regions moving at high velocities with respect to the central star (Fig. 3). In order to state the origin of the HCO+ intensity contrast between the slow, central component and the faster lobes, we will discuss the different factors that could affect the line intensity.
The central intensity dip of the HCO+ line could be due to the presence of self absorption in the line. The importance of opacity effects has been evaluated through the H12CO+/H13CO+ (J = 1-0) intensity ratio, that in the nebula center is found to be 8 (see HCO+ data in Sánchez Contreras et al., 1997, and in Fig. 4). This value agrees with the 12C/13C isotopic ratio estimated from CO data in regions in which the emission is known to be optically thin (Sanchez Contreras et al., 1997), and from the H12CN/H13CN (J = 1-0) intensity ratio (see Sect. 3.4). We thus conclude that the HCO+ lines are optically thin or only moderately opaque.
The regions with higher HCO+ intensity could also correspond to regions of enhanced density. To check this possibility we have compared our HCO+ observations with those of another high-density tracer, CS (see the CS and HCO+ spectra within 27" - the beam FWHM - around the nebula center in Sánchez Contreras et al. 1997). The remarkable differences between the spectra of both species (note the maximum CS intensity around the systemic velocity) lead us to rule out a density effect as responsible for the strong HCO+ emission enhancement in the lobes of OH 231.8+4.2. On the other hand, we have also studied the excitation conditions in OH 231.8+4.2 from the HCO+ (J = 1-0)/(J = 3-2) line intensity ratio assuming that both transitions are optically thin. The brightness ratio of these lines is found to vary between 2 (for velocities between -10:+50 km s-1) and 0.8 (for velocities between 50-90 km s-1), corresponding to (representative) excitation temperatures () of 8 K and 12 K, respectively. Note that the strong dependence of the previous ratio on (e.g. for = 6 K the (1-0)/(3-2) ratio is 5.7, ten times higher than for = 15 K) limits the possible values of to a relatively small range. Excitation variations within this range cannot explain the strong intensity differences observed between the central, slow clumps and the outer, faster ones.
On the other hand, we note that the values of obtained for the HCO+ lines are very similar to the kinetic temperature in the nebula estimated from the CO data (Sanchez Contreras et al., 1997). This could suggest that HCO+, like CO, is thermalized and, therefore, that the gas density is, at least, 106 cm-3 (the HCO+ critical density).
In view of the previous discussion, we conclude that the observed intensity variations of HCO+ are dominated by the real abundance variations of this molecule along the nebula: for some reason, this molecule is efficiently formed in the outer, accelerated clumps of OH 231.8+4.2. (For an estimate of the HCO+ abundance in different clumps see Sánchez Contreras et al. 1997; note that the there assumed, 10 K, is very similar to the values here obtained.)
In our opinion, the efficient formation of HCO+ in the lobes of OH 231.8+4.2 is very probably induced by shocks. The passage of shock fronts through the nebula (of which observational evidence exists, see Sect. 1) can modify the abundances dissociating stable molecules, followed by recombination favoring the formation of certain species like HCO+, that is very rare in equilibrium chemistry (e.g. Neufeld and Dalgarno, 1989; see also discussion by Morris et al., 1987). A shock-induced chemistry would then explain the increase of the HCO+ abundance in the post-shocked gas and, consequently, the particularly intense emission from this molecule in OH 231.8+4.2. We note that most of AGB envelopes (which usually do not show shock signs) show very weak or null HCO+ emission (Cox et al., 1992, and references therein). This fact suggests that the high abundance of HCO+ in OH 231.8+4.2 cannot be attributed to ionization by galactic UV photons since this source is not immersed in a particularly intense UV radiation field.
Ionization by stellar UV photons should also be ruled out as responsible for the HCO+ abundance enhancement in the outer, fast flowing regions in OH 231.8+4.2. In fact, in the case of stellar photoionization the most intense emission should be observed in the inner regions of the molecular shell (see the case of CRL 618, Cox et al., 1992). On the other hand, the UV radiation by the central star of OH 231.8+4.2 is known to be very weak (Cohen, 1981; Cohen et al., 1985; Taylor & Morris, 1993). Reactions involving high-energy particles have also been considered, since they are an efficient mechanism for producing HCO+ in dense molecular clouds (Glassgold, 1996). Nevertheless, these processes, due to the large penetrating power of cosmic rays, would lead to an HCO+ distribution similar to that of the other molecules, i.e. with the most intense emission arising from the most dense regions (the central clump in this case). Moreover, high energy particles, like galactic UV photons, should act in a similar way in AGB envelopes, that, as we have mentioned, show very weak or null HCO+ emission.
The distribution of the SO emission is found to be similar to that of CO (Sánchez Contreras et al., 1997; Paper III), although significantly more compact (see Fig. 5). The SO emission occupies a narrow region ( 1017 cm broad) that extends about 3.5 1017 along the symmetry axis of the nebula. The total line width is about 100 km s-1, indicating that SO is present in the accelerated lobes of OH 231.8+4.2. The peak of the SO emission (in contrast to HCO+ and similarly to CO) arises from the central, slowly expanding component. From the SO p-v diagram along the nebula axis (Fig. 6) we can see that the SO emission follows the general velocity gradient along the symmetry axis of the nebula (the emission from the north/south lobe being blue/red-shifted). In addition, the SO emission reveals the presence of an expanding disk (or ring) that surrounds the central star. The presence of this inner, expanding disk is indicated by the inversion of the slope of the velocity gradient at the nebula center. In fact, in the PA 21o p-v diagram, we can see that in the densest, central 2" of the nebula, the velocity gradient has opposite sign (the north/south emission is red/blue-shifted) to that observed in the outer (weaker) regions of the nebula. Equatorial, expanding disks have been also detected in the inner molecular envelopes of other PPNe, leading to similar spectral features in p-v diagrams (see e.g. the case of M 1-92, Bujarrabal et al., 1998). The characteristic radius and the expansion velocity of the equatorial disk (or ring) of OH 231.8+4.2 deduced from our maps are 2 1016 cm and 6-7 km s-1, respectively. These values yield a kinematical age for the disk of about 1000 yr, very similar to that found for the bipolar molecular outflow (Sanchez Contreras et al., 1997). Nevertheless, the low velocity of the disk seems to indicate that this structure is a remnant of the old AGB envelope that has not been affected by the two-wind interaction that presumably accelerated the gas in the lobes of OH 231.8+4.2. No sign of rotation has been found in the disk (the rotational velocity should be less than 2 km s-1; see the p-v diagram in the direction perpendicular to the nebula axis, Fig. 6), but the spatial resolution of these observations does not allow us to rule out some rotation in the innermost regions of OH 231.8+4.2 (see below). We have estimated the SO abundance (relative to H2) in the equatorial disk and in the outflow (procedure and assumptions described in Sánchez Contreras et al. 1997) and have not found significant differences between the two components, for which we obtain a value of 10-6.
3.3. SiO maser
In Fig. 7 we show the SiO (v = 1, J = 2-1) maser spectra obtained in six different epochs. Three main spectral components can be distinguished at (LSR) velocities 26, 33, and 40 km s-1. Since SiO masers usually arise from regions very close to the star, the centroid of the line is considered to be a reliable indicator of the stellar (or systemic) velocity (), see Jewell et al. (1991). In the case of OH 231.8+4.2 we get 33 km s-1, that approximately coincides with the peak of the CO emission (Sanchez Contreras et al., 1997) and the centroid of the SO line emission. Each of the three SiO features is formed of several subcomponents, indicating the complex and, probably, clumpy distribution of the gas in the vicinity of the central star. We are not able to spatially separate the regions from which the three (blue, systemic and red) components arise, and only an upper limit for the size of the emitting region of about () 2 1015 cm can be given. This value has been estimated from the direct analysis of the visibility phases for each spectral feature.
The blue and red spectral components of the SiO maser line roughly coincide with the two intensity peaks of the H2O maser at 22 GHz (Bowers & Morris, 1984). The spectral distribution of the OH maser (1667 MHz) is relatively flat in the LSR velocity range [-20:+80] km s-1 and possesses a narrow spike at +19 km s-1 (Bowers & Morris, 1984). The difference between the OH and SiO maser profiles is not surprising, since in OH 231.8+4.2 the OH maser emission occupies an extended region of 10" (Morris et al., 1982).
We would like to note that the blue and red spectral components of the maser roughly coincide with the blue and red intensity peaks of the SO disk emission (see Fig. 6). This fact, could suggest that the maser emission originates in the innermost regions of the expanding disk seen in the SO maps. The relatively stable SiO line structure (note that the same three spectral features appear in all the epochs in spite of the relative intensity variations between them) is in agreement with that idea. In this scenario, the systemic spectral component could arise from tangential maser amplification, and the blue and red ones, from the front and back regions of the disk, respectively. Our data would, indeed, indicate a small (or null) velocity gradient between the inner and outer parts of the disk traced by the maser and the SO thermal emission respectively. If it is confirmed that the SiO masers are distributed in an expanding disk close to the star, then the models requiring inner accretion disks or orbiting structures in the stellar vicinity (due to some kind of companion) to explain the onset of bipolarity in PNe (Soker & Livio, 1994; Jura et al., 1995; Mastrodemos & Morris, 1998, and references therein) should be substantially revised. However, the similar velocities of the blue and red intensity peaks of the maser and SO lines could be just an unlucky coincidence. In fact, for OH 231.8+4.2, rotation or infalling velocities are expected to be of the same order of magnitude ( 5-8 km s-1) at the distances where the masers originate (a few stellar radii). VLBI observations with about resolution are still needed to unveil the structure and kinematics of the inner regions of OH 231.8+4.2.
The relative intensity between the different spectral components and subcomponents as well as the total flux of the SiO line are found to strongly vary with time. In Fig. 8 we have plotted the maser intensity integrated over the total width of the line and over each of the three main (blue, systemic, and red) spectral features. In this figure, the epochs at which the maximum and minimum intensity of the near-infrared (n-IR) should have to occur (based on the n-IR light curve by Kastner et al., 1992) have been indicated by dotted-lines. We have found a relative minimum of the SiO flux 60 days before the n-IR minimum. Assuming that maser pumping is radiative, this phase lag would correspond to a distance of 1017 cm between the maser (very close to the star) and the nebular dust reflecting the n-IR starlight. This value is in agreement with the measured distance from the center to the region from which the maximum n-IR emission arises (Kastner et al., 1992; Paper III).
3.4. H13CN and NS
We have found the H13CN (J = 1-0) emission to be distributed along the molecular outflow, sharing the general velocity gradient (see Fig. 9). The spectral and spatial distribution of this molecule is similar to that of SO, with the most intense emission arising from the slow, central component. Note that in this case we are not able to distinguish the inner expanding disk or ring seen from the SO emission. The H13CN intensity peak does not lie at the systemic velocity (33 km s-1) but at 28 km s-1, that approximately corresponds to the blue peak of the SO disk emission.
From the H12CN/H13CN (J = 1-0) intensity ratio (H12CN low-resolution maps were obtained with the 30 m IRAM radiotelescope, Sánchez Contreras et al. 1997) and assuming that both lines are optically thin, we deduce an isotopic 12C/13C ratio of 5-10. This value is in agreement with that found from the 12CO/13CO (J = 1-0) intensity ratio of the line wings, which are optically thin (Sanchez Contreras et al., 1997). Such a value is lower than that predicted by standard evolutionary models ( 20), and is often observed in evolved stars (Charbonnel, 1995; Palla et al., 1998). The 13C abundance enhancement has been attributed to a nonstandard mixing mechanism that operates during the red giant phase.
We also report the first detection of nitrogen sulfide (transition , J = 5/2-3/2, parity-e) in circumstellar envelopes (Fig. 10). The NS emission is distributed in a compact central region and in the outflow (note the relatively large width of the line). The total flux of the line (integrated over all the hyperfine components) is relatively high, 15 Jy km s-1. The intensity ratio between the 3 highest and the 2 weakest hyperfine components is in agreement with the theoretical value (considering the different strength of each hyperfine line component, McGonagle et al. 1994) suggesting that the line is optically thin. This fact allows us to estimate the NS abundance (relative to H2) in the central clumps of OH 231.8+4.2. Assuming = 10 K and the mass of the central clumps (-30:+80 km s-1) to be 0.5 (see Sánchez Contreras et al. 1997) an abundance of 2 10-8 is found.
The large variety and abundance of nitrogen- and sulfur-bearing molecules in OH 231.8+4.2 is a clear sign of an active chemistry that is probably induced by shocks. Shocks would initiate (endothermic) reactions that trigger the N and S chemistry (e.g. N+ + H2 ..., and S+ + H2 ...; Lada et al. 1978) and could also extract additional S from the surface of dust grains (Jackson & Nguyen-Q-Rieu, 1988).
3.5. Continuum at = 3 mm
The 3 mm ( 87.8 GHz) continuum map of OH 231.8+4.2 is shown in Fig. 11. The continuum is found to slightly extend along the nebula axis. From our mapping we obtain that the 3 mm continuum can be separated in an extended ( 18 mJy) and a point-like ( 7 mJy) component. The continuum total flux obtained from these observations, is in agreement with that obtained from single-dish millimeter observations (see below), showing that flux losses by the interferometer are negligible. The extended component is roughly elongated along the symmetry axis of the nebula, with a total length of 8". The previous separation in components has been established by fitting different structure models directly to the visibilities. The best fit was obtained when assuming a point+`elliptic gaussian' source. The structure and the fluxes obtained for the two components of the continuum at 3 mm are in agreement with the general properties of the continuum emission at millimeter wavelengths in this source (Sanchez Contreras et al., 1998). These authors attribute the extended mm-continuum emission to cold dust ( 20 K) distributed along the lobes of the nebula. The point-like emission would be due to a compact region of warmer dust ( 55 K) around the central star.
© European Southern Observatory (ESO) 2000
Online publication: June 5, 2000