2. Molecular Line Observations
The line observations were made in January 1996 using the 10.4 m telescope of the Caltech Submillimeter Observatory (CSO) on Mauna Kea, Hawaii. The telescope is equipped with SIS junction receivers cooled to liquid helium temperature. The observations were made of the CO(2-1) and CO(3-2) lines at 230 and 345 GHz and of CS and lines at nearby frequencies. The atmospheric conditions were variable during the observing run, and the higher frequency observations were made during the better weather. The zenith atmospheric opacity at 225 GHz, measured by the NRAO tipping radiometer at the site, was between 0.1 and 0.15 for the 345 GHz observations and between 0.2 and 0.25 for the 230 GHz observations.
The frequencies of the observed spectral lines are given in Table 2. Three lines, CO(2-1), CS(5-4), and (25-24) were observed in the 230 GHz band. The telescope half-power beamwidth was , the main-beam efficiency was 76% and the single-sideband receiver temperature was 100 K. Two lines were observed in the 345 GHz band, CO(3-2) and CS(7-6). The telescope half-power beamwidth was , the main beam efficiency 65%, and the receiver temperature 130 K. The effective single sideband system temperatures, including the effects of atmospheric emission and absorption, were 500 K at 230 GHz and 800 K at 345 GHz. The receivers for both bands are double-sideband, and spectral lines in the image sideband, 3 GHz below the observed frequency, are also detected.
Table 2. Molecular line observations of V Hya
The spectral lines were observed by three acousto-optic spectrographs (AOS) of which two at a time could be used. The first AOS has 2048 channels over a total bandwidth of 1.5 GHz, of which only the inner 1 GHz (the receiver bandwidth) produces usable data. The channel - to - channel spacing is and the resolution is at 345 GHz. The second has 1024 channels over a bandwidth of 500 MHz, giving a resolution of at 345 GHz and at 230 GHz. The third has 1024 channels over 50 MHz, giving a resolution of at 345 GHz and at 230 GHz. The spectrometer frequency was calibrated using an internally-generated frequency comb, and the velocity scale is corrected to the Local Standard of Rest (LSR).
The observations were made by chopping between the star position and an adjacent sky position with the secondary mirror. The chop throw was in azimuth at a rate of 1 Hz. Pairs of chopped observations were made with the source placed alternately in each beam. The spectral baselines resulting from this procedure are linear to within the r.m.s. noise. The CO emission from V Hya is strong enough that it can be used to measure the telescope pointing offsets, and the pointing was corrected every hour or so by mapping the integrated CO line flux from the star.
The temperature scale and the atmospheric opacity were measured by comparison with a hot (room temperature) load. The line temperature was corrected for the main beam efficiency, and the resulting scale is the Rayleigh-Jeans equivalent main beam brightness temperature , i.e. that measured by a perfect 10.4 m antenna above the atmosphere.
The remarkable new feature in V Hya revealed by these observations is a very fast molecular wind, seen as weak emission extending to about with respect to the velocity of the brighter CO emission. This weak emission is also seen in the CO(3-2) observations of V Hya. Such weak wings can be artificially produced in observations with acousto-optic spectrometers when they are slightly out of focus. We examined the shapes of the frequency calibration spikes; these show no detectable broad wings. Quantitatively, = 0.02 for both the CO(2-1) and CO(3-2) line observations, where is the central velocity of the CO profile where the emission is brightest. The same quantity measured in the frequency comb spectra is 0.001. Further, no "wings" are seen on observations of the much stronger CO(3-2) and CO(2-1) emission from IRC+10216, which was also observed during this observing run. We are thus convinced of the reality of the fast wind from V Hya.
This fast wind has not been seen in the many observations of V Hya made previously, and its detection now can be attributed to the great improvements in sensitivity and baseline stability made possible by the new receivers and the chopping secondary mirror at the CSO. It can be identified with the high negative velocity CO 4.6 line absorption observed by Sahai & Wannier (1988) and Sugerman et al. (1997), the high velocity KI line absorption and fluorescent emission found by Plez & Lambert (1994), and the high negative velocity shocked-gas emission lines observed by Lloyd-Evans (1991). The CO(2-1) and, especially, the CO(3-2) observations (see below) show that this wind is symmetric in velocity and extends to higher velocities than observed previously. V Hya is thus one of the small group of evolved stars with molecular wind speeds in excess of 100 .
The observational results are presented in Table 2 and Figs. 1-7. Table 2 lists the rest frequencies of the observed lines, the channel-to-channel r.m.s. noise (as measured by the 500 MHz AOS, except for the CO(3-2) and CS(7-6) lines, where the results from the 1.5 GHz AOS are used), the integrated line flux in units of , the peak line temperature, the central velocity and the terminal wind outflow speed , which is measured from the half width of the line at zero intensity. The parameters for the CS and lines were found by fitting parabolic profiles to the data (cf. Knapp & Morris 1985). The CO line shapes however are complex. Several components are identified in the profiles and are discussed below; the quantities listed in the last three columns of Table 2 are eye estimates.
The CO(2-1) line observed with the 500 MHz AOS is shown in Fig. 1. The feature at " " is the (25-24) line at 227.4 GHz, detected in the image sideband. This line seems to have roughly the parabolic shape expected for spherical uniform outflow, and the parameters in Table 2 were found by fitting a parabola to the line profile after recalculating the velocity scale and removing the broad CO emission as a "baseline".
The CO line profile does not have the steep sided, roughly parabolic shape expected for uniform outflow; rather, its shape is roughly gaussian, and the line also shows asymmetric, double-peaked structure (Zuckerman & Dyck 1986; Kahane et al. 1988; Olofsson et al. 1988; Jura et al. 1988; Tsuji et al. 1988; Zuckerman & Dyck 1989; Nyman et al. 1992; Olofsson et al. 1993; Bieging & Latter 1994; Bujarrabal et al. 1994b,c; Stanek et al. 1995; KABM). The CO(2-1) line in Fig. 1 is considerably broader than the line and has a full width at zero power, after removal of the broad CO emission, of , suggesting a wind outflow speed of (cf. KABM), much larger than the speed of given by the line. The profile shape in Fig. 1 suggests that the 165 and 45 components are kinematically distinct.
Fig. 2 shows the CS(5-4) line profile compared with the CO(2-1) profile. The CS line is narrower than the CO line and is double peaked; there is some suggestion that the velocities of the peaks relative to the star are smaller than are those of the CO peaks. There is no indication of the fast () wind in the CS profile, even when it is smoothed to a fairly coarse resolution, 10 .
Fig. 3 shows the CO(3-2) line observed with the 1.5 GHz AOS. The excess noise beyond is due to the 1 GHz bandwidth. The feature at "-180 " is the CS(7-6) line in the image sideband. That at "+320 " is probably the (4-3) line. The line shapes of the features at "-240 " and "-280 " suggest that these lines are in the image sideband; they show slight asymmetry which is the mirror image of the asymmetry seen in the CO line. We could find no plausible identification for them; they appear also in the spectrum of the bright carbon star IRC+10216. The CO(3-2) line has a very similar shape to the CO(2-1) line. The fast wind is even more pronounced at this frequency, and suggests an outflow speed of at least 200 . Again, it seems to be kinematically distinct from the 45 component. Fig. 4 shows the CS(7-6) line compared with the CO(3-2) line. The lines have similar shapes but the CS line is somewhat narrower and has steeper sides.
Observations of the CO(3-2) line were made at offsets of (the full beamwidth) in right ascension and declination to investigate whether the envelope is significantly spatially extended. We did not attempt a fully sampled map; the IRAM telescope, with its beam at 230 GHz, has made the best single-dish map currently available (KABM). Fig. 5a shows the CO(3-2) line profiles observed at in declination, and Fig. 5b the profiles at in right ascension.
The profiles observed north and south of the stellar position are roughly parabolic and centered on the stellar velocity. The difference in intensities may be due to spatial structure, but we cannot rule out the effect of small pointing errors. The expected peak line temperature from the center of V Hya is 0.2 K if the molecular line emission is a point source on the scale of the CO(3-2) beam, compared with the observed values of 0.36 K at south and 0.77 K at north, showing that the envelope is slightly elongated in the north-south direction, in agreement with the results of KABM. The central velocity of the emission at both the north and south positions () is roughly the same as that observed at the stellar position, and the line width is similar to that measured for the CS and lines (Table 2).
By contrast, the line profiles observed at the right ascension offsets of have quite different shapes. They show broad emission with a roughly parabolic shape at a peak strength T = 0.27 K () and T = 0.14 K (), close to the values of 0.2 K expected if the CO emission in the envelope is a point source. In addition, these profiles have two sharp spikes, at velocities of -8.4 at and -24.8 at . The mean velocity of these spikes is -16.6 , very close to that measured from the CS and lines, and their velocity difference is 16.4 , giving an outflow speed of 8.2 . Such line shapes cannot be produced by a spherical envelope expanding at constant velocity (cf. Fig. 5a); the spikes in Fig. 5b are almost certainly emitted by the same gas which produces the double-peaked shape of the CO(2-1) line (Fig. 1). The velocities of the peaks in the CO(2-1) line profile are -24 and -10 , centered at -17 and suggesting an outflow speed of .
Fig. 6 shows the CO(3-2) line profiles measured at the offset positions smoothed to a resolution of 10 . The only profile showing believable high velocity emission is at (), which may have weak emission to a velocity of about +130 with respect to the star. There is a suggestion in the profiles at () of fast bipolar emission, both for the 45 wind and the 200 wind, and that the bipolarity of the fast winds are in the opposite sense to that of the 8 feature shown in Fig. 5b; i.e. the 8 feature is at negative velocity at () while the faster winds are at positive velocity, and vice versa.
As these data show, V Hya's circumstellar molecular envelope does
not have the simple geometry which seems to be the case (though less
so as observations get better) for most envelopes, which are modeled
reasonably well by a spherically symmetric wind expanding at constant
velocity and mass loss rate. Rather, it appears as though several
kinematic components, all centered at the stellar velocity, can be
identified in the V Hya wind:
Fig. 7 shows the ratio of the CO(3-2) and CO(2-1) intensities at the stellar position as a function of velocity. The data were averaged in bins of . The velocity ranges of the four wind components described above are shown. Fig. 7 shows that within the noise the line ratio at all velocities is close to the value of 2.25 expected if the CO(2-1) and CO(3-2) line emitting regions have the same brightness and same size and are smaller than the CO(3-2) beam. The ratio is slightly smaller close to the line center, suggesting that the envelope is slightly resolved at , consistent with the discussion above.
The ratios of the total intensity and peak brightness temperature of the CS(7-6) and CS(5-4) lines, respectively 2.39 and 2.22, are also consistent with a molecular envelope with the bulk of its emission within the telescope beam. Because the emission from circumstellar molecular envelopes is centrally concentrated, the derivation of the angular size is not straightforward, but the present observations and those of KABM suggest that the envelope is only slightly extended and probably has an outer radius of or less. The IRAS observations of the circumstellar shell also find the envelope to be unresolved (Young et al. 1993a, b). At a distance of 380 pc, the linear radius of the envelope is only about cm and the envelope age, at an outflow speed of 15 , is about 1000 years.
What is the underlying geometry? The mapping observations described by KABM and those in the present paper show that the CO emission near the systemic velocity (which we find from the measurements in Table 2 to be -16.3 ) is slightly elongated in the north-south direction. The outflow speed from the envelope, as given by the CS and observations at the central position, and from the CO(3-2) lines observed north and south of the star, is about 15 . The 4.6 absorption line observations by Sugerman et al. (1997) show a strong feature at -30 with respect to the LSR which also provides evidence for an envelope outflow speed of 15 . However, the CO(3-2) lines observed east and west of the center, and the CO(2-1) line observed at the central position, show strong narrow spikes at , which are spatially displaced from each other. Together, these observations suggest that the molecular outflow from V Hya has a velocity of 15 but that it is in a flattened structure (sketched in Fig. 8) inclined to the line of sight, with the projected major axis lying approximately north-south. The spikes then arise from the gas along the minor axis, moving at 15 and inclined at about to the line of sight. The CS and lines arise from denser gas in the inner regions of the expanding disk, and so are not spatially resolved. This interpretation differs from that of KABM, who describe a model in which the basic envelope outflow is spherical with a velocity of 7.5 , and with large cones gouged out by the fast winds.
We suggest that the most likely configuration for the high - velocity gas, both the strong 45 wind and the weaker 200 wind, is flow from the poles of the disk (Fig. 8). The evidence for this is admittedly sketchy, consisting of the mapping observations in Fig. 6 (where fast gas at positive velocities is tentatively seen in the profile at ) and the optical emission line maps presented by Plez & Lambert (1994). If this model is correct, the fast gas is likely to have a space velocity of at least 230 , and this star joins CRL 618 (Cernicharo et al. 1989; Gammie et al. 1989) and OH231.8+4.2 (Alcolea et alet al. 1996) in having a molecular outflow with velocity in excess of 200 . The mapping observations in Fig. 6, and the line shapes, suggest that the outflow velocity increases with distance from the star. Perhaps the 200 gas is moving freely, while the 45 component is slow circumstellar gas swept up by the fast bipolar outflow. The fact that blue-shifted CO absorption is seen to about -130 (Sugerman et al. 1997) suggests that the fast flow has a fairly large opening angle, so that much of the total velocity range is seen in absorption.
In summary, the observations described in this section suggest that V Hya is surrounded by a flattened molecular shell of radius cm expanding at . The star is also ejecting a fast molecular wind with at least two components, the first moving at about 50 and the second at . Our observations provide very tentative evidence that the fast winds are bipolar and are flowing from the poles of the circumstellar shell, with speed increasing with distance from the star.
© European Southern Observatory (ESO) 1997
Online publication: April 20, 1998