3.1. Low velocity forbidden line emission
From a sample of 56 Herbig Ae/Be stars we found 28 stars showing [OI] 6300 emission. This detection rate of 50% is similar to that of Böhm & Catala (1994). The equivalent widths and velocities of the [OI] 6300 lines are presented in Table 1. These 28 stars can be separated into different categories determined by the measured velocities and the profile of the line: I) High velocity ( 100 kms ) blueshifted emission, which may be accompanied by a low velocity blueshifted peak, II) low velocity (10 55 kms ) broad (FWHM 80 kms ) blueshifted emission, III) low velocity (10 kms 15 kms ) redshifted emission with narrow lines, and IV) unshifted ( 5 kms ) emission with narrow lines. Each star is identified by this type in Table 1 and an example of each type of forbidden emission line is presented in Fig. 1. A histogram showing the distribution of [OI] centroid velocities is presented in Fig. 2.
In the first category there are four stars, V645 Cyg, Z CMa, PV Cep and LkH 233, that show high velocity blueshifted emission. Centroid velocities for the HVC range from about -100 kms to as high as -450 kms in Z CMa. Velocities in the blue wing of Z CMa reach at least -700 kms . All four stars also show double-peaked or P-Cygni profiles in their H lines, an indication of a stellar wind. These four stars are also known to be associated with stellar jets and/or molecular outflows: V645 Cyg is associated with a bipolar molecular outflow (Schulz et al. 1989; Verdes-Montenegro et al. 1991) and there are at least two Herbig Haro (HH) objects in its vicinity (Goodrich 1986; Hamann & Persson 1989). Z CMa drives a stellar jet (Poetzel et al. 1989) and a molecular outflow (Evans et al. 1994). It is also believed to be a FU Orionis object with a massive and highly active circumstellar disk (Malbet et al. 1993; Barth et al. 1994) and is therefore not likely to be a true HAEBES. Rather the object's high luminosity is due to its extremely high accretion rate. PV Cep is associated with a bipolar molecular outflow (Fukui et al. 1993) and HH emission (HH 215, Neckel et al. 1987) and LkH 233 although not associated with a bipolar molecular outflow (Cantó et al. 1984; Leverault 1988) does illuminate a bipolar nebula (Staude & Elsässer 1993) and has a spectroscopically observed jet (Corcoran & Ray 1997b). All of these stars show two distinct velocity components at [OI] 6300, with the LVC showing blueshifted velocities of less than 100 kms . The LVC velocities are, however, higher on average than those of the LVCs observed in HAEBESs of category II.
In category II there are 14 stars, all showing low velocity blueshifted [OI] 6300 emission, with velocities between -55 and -10 kms . The stars in this category can be further subdivided into two groups of comparable size; those with intermediate blueshifted velocities, ranging from -55 kms - -20 kms (IIa), and those with low blueshifted velocities, ranging from -15 - -10 kms (IIb). Of the 7 stars in category IIa, 5 are known to have associated molecular outflows and/or jets: LkH 198 is associated with a highly collimated optical jet (Corcoran et al. 1995) and molecular outflow (Leverault 1988, Nakano & Tatematsu 1990). R Mon possesses a stellar jet (Mundt et al. 1987) and a molecular outflow (Cantó et al. 1981) as has MWC 1080 (Poetzel et al. 1992; Yoshida et al. 1992). V376 Cas has both HH objects in its vicinity and possibly a molecular outflow (Corcoran et al. 1995; Leinert et al. 1991; Piirola et al. 1992) while HK Ori may be associated with HH objects (Goodrich 1992). Only NX Pup and KK Oph have no observations of such outflow phenomena in the literature. In category IIb there are 7 stars, none of which is associated with a molecular outflow in the lists of Fukui et al. (1993) or a bipolar nebula in Staude & Elsässer (1993). No optical jets or HH objects are mentioned in the literature in connection with any of these 7 stars.
It seems plausible that the stars in category IIa are similar to those in category I, except for the "absence" of the HVC in the [OI] 6300 emission. This "absence" may not be real in some cases but instead may be due to the orientation of the outflow. Another possible cause is an intermittancy effect (see below). Those stars in category IIb are, for the most part, lacking any evidence for extended outflows. Of course the effect of inclination will play an important part in determining the radial velocity of any [OI] 6300 line. Some of the stars in category IIa are believed to have outflows in the plane of the sky, for example, LkH 198 and R Mon (Corcoran et al. 1995; Brugel et al. 1984). As a result the HVC and LVC emission may be unresolved. The spatial orientation of the outflow from V376 Cas is unknown, but it is plausible to assume that, like its neighbour LkH 198, its outflow axis is also in the plane of the sky, given the tendency for outflows in the same area to be aligned (Reipurth 1989).
Category III has only three stars, BD 164, MWC 297 and LkH 257, all of which showed redshifted [OI] 6300 emission at low velocities (+10 +15 kms ). None of these stars are known to be associated with jet or molecular outflow activity. Finally, category IV has seven stars all showing small shifts in [OI] 6300 velocities within the range -5 kms 5 kms . Due to the limitation of the errors in the centroid velocities it may be that categories III and IV are not distinct groups. If this is the case there may be no HAEBESs with genuinely redshifted [OI] 6300 emission in our sample.
Clearly there is not a symmetrical distribution of [OI] 6300 velocities about 0 kms in the histogram (Fig. 2). In contrast to the histogram published in Böhm & Catala (1994) the distribution indicates that most HAEBESs show low velocity blueshifted forbidden line emission.
As well as observations of the [OI] 6300 line, our spectra also include the [SII] 6716/6731 forbidden line doublet. These lines have a critical density roughly 150 times lower than the [OI] doublet (Osterbrock 1989) and can be used to probe low density regions. None of the stars observed in the sample that do not show [OI] 6300 emission display [SII] 6716/6731 emission. Of the 28 stars that do show [OI] 6300, 9 display detectable levels of [SII] 6716/6731 (see Table 2). One star, ST 202, is left out of Table 2 due to lack of information about its distance and absolute flux levels. It is interesting to note that 7 of these 9 stars with [SII] 6716/6731 emission also have definite associations with molecular outflows and/or HH jets and objects. For the other 2 stars (KK Oph and ST 202) with [SII] 6716/6731 emission, observations are lacking that would confirm or refute the hypothesis that all HAEBESs with [SII] 6716/6731 emission possess extended outflows. In any event all of the stars, except ST 202, with detectable [SII] emission either fall into our category I or category IIa.
Table 2. Mass loss rates calculated from the luminosity of the [SII] 6731 line in 8 Herbig Ae/Be stars. See Sect. 3.1 for a discussion of the calculation of the mass loss rate for the HVC and LVC. The first four stars have clearly separated HVC and LVC emission, although this is not the case for the last four stars. Here is the radial velocity of the profile centroid. Note that is larger for the [SII] line than the [OI] line for all stars except Z CMa. The values for are estimated tangential velocities. The mass loss rate are calculated for an electron temperature of 104 K and show an average increase of a factor of 2 for an electron temperature of 5 103 K. The HVC velocities marked with a are representative values assumed for the HVC. It is not clear in these cases whether we are observing a HVC, LVC or, possibly, a combination of both. The entries in the last column therefore should be seen as indicative of the mass loss range. The HVC velocity of R Mon, marked *, is taken from the tangential jet velocity of the associated jet (Mundt et al. 1987).
It is relatively straightforward to derive a value for the mass loss rate for a wind from a young stars based on the luminosity of forbidden lines. As such lines are optically thin, their luminosities are proportional to the total number of radiating atoms along the line of sight and consequently proportional to the total mass in the radiative component of the flow. The mass loss rate is found from the mass of the high velocity component, the flow speed and a scale length, l of the emission, where (see for example Hartigan et al. 1995). Using the luminosity of the [SII] 6731 line we have estimated the mass loss rates of 9 HAEBESs. This line comes from the 2 D to 4 S transition of singly ionized sulphur. Assume the total luminosity of the [SII] 6731 line is L21, then
where , are the number of S ions in the emitting region in the lower and upper states, 1 and 2, respectively; A21 is the Einstein coefficient for the 6731 transition and is 8.82 10-4 (Mendoza 1983); h is the energy of the transition and is 2.95 10-12 ergs; g1, g2 are the statistical weights of the two levels and are 10 and 4 respectively; T is the temperature at which the transition takes place, is the electron density; and , the critical density for the transition, is equal to where is the collisional de-excitation rate. The exact temperature for the transition is uncertain and can be estimated from ratios of forbidden lines with well separated energies. Unfortunately we lack observations of the appropriate lines to allow derivations of electron temperatures. The calculations of Kwan and Tademaru (1995) indicate electron temperatures in the range 3000 K to 104 K and above. For the purposes of the calculation here a representative value for the electron temperature of 104 K is used; an electron temperature of 5 103 K produces a resultant mass loss rates on average a factor of 2 higher. Table 2 presents the calculated mass loss rates.
C21, the collisional de-excitation rate is given by
where is the collision strength. The collision strength of the [SII] 6731 line is 2.79 (Mendoza 1983). The critical density is 1.47 104 cm-3 for T = 104 K, and 1.04 104 cm-3 for T = 5 103 K.
To make an estimate of the total luminosity from the [SII] 6731 line it is necessary to determine the number of S ions in the emitting region. This is done as follows: The total number of S ions at level 1 in the emitting region or aperture, , is
where (S ) is the number of S ions in the emitting region; (H) is the number of hydrogen atoms in the aperture; is the total number of atoms of all species in the aperture; and is the total mass in the aperture. For solar abundances (Allen 1973) the mean molecular weight, µ, for a neutral atomic gas is 1.24, / (H) is 1.86 , assuming all the sulphur is ionised, and (H)/ is 0.921. We assume that / (S ) 1 for T K. Combining these values gives a value of = 4.14 10-15 for in units of . Substituting this value of into the earlier equation (eqn. 2) for L21 (i.e. L6731):
for the mass within the emitting region for an electron temperature of 104 K. A similar calculation can be carried out for an electron temperature of 5 103 K.
To calculate the mass loss rate we take a length scale, , the projected aperture size on the sky, and a velocity, , the projected velocity on the sky. The mass loss rate is then . Finally then the mass loss rate as determined from the luminosity of the [SII] 6731 line is
where is the jet velocity in the plane of the sky and is the size of the aperture. For we assume an average orientation angle of the various outflows relative to the line of sight of and calculate from the measured radial velocities (Table 2), except in the case of R Mon, where independent observations suggest that the outflow is practically in the plane of the sky (Mundt et al. 1987). The tangential velocity cited by Mundt et al. (1987) for the R Mon jet (HH 39) is 300 kms while the heliocentric radial velocity they cite is -75 kms . In fact the radial velocity is actually that of the bipolar jet close to the star (4"-10", Brugel et al. 1984) and the tangential velocity that of HH 39, much further away. If we assume that the outflow undergoes no significant change in velocity between the two regions, the inferred orientation angle is . The value of for the HVC is the average slit width or 1.5". For the LVC this is clearly an overestimate of the size of the emitting region. Observations of LVC emission in TTSs by Solf & Böhm (1993) and Hirth (1994) suggest that the LVC extends a distance of 0.1 - 0.2" from the star. In our observations (Corcoran & Ray 1997b) we find that the LVC emission from the HAEBES PV Cep, for example, extends approximately 0.5" from the star. Consequently we choose an approximate value of 0.5" for here.
The approximate estimates of mass loss rates, , for the 8 stars with detected [SII] emission are listed in Table 2 (ST 202 is left out due to insufficient data about distance and absolute flux). For the 4 stars (Z CMa, PV Cep, V645 Cyg and LkH 233) with distinct HVC and LVC emission, corresponding separate values of are listed. In the case of the other 4 stars the [SII] lines cannot be separated into clear HVC and/or LVC emission. Consequently the mass loss rates in Table 2 for these stars (LkH 198, V376 Cas, R Mon and KK Oph) are calculated for two distinct cases: Pure HVC emission and pure LVC emission. Without resolving the two components no more definite values can be given.
The mass-loss rates are typically 10-100 times higher than those observed for cTTSs (Hartigan et al. 1995). We emphasize that the mass loss rates are order of magnitude estimates only, as they are strongly dependent on the temperature.
Other methods of estimating the mass loss rate exist and give varying values, examples include Hollenbach (1985) where the mass loss rate is derived from the luminosity of the [OI]63µm line assuming each atom passes through only one shock and Poetzel et al. (1992) where is derived from the visible jet parameters. See Hartigan et al. (1995) and Kwan & Tademaru (1995) for a discussion. Nisini et al. (1995) make direct estimates of the mass loss rate from the luminosities of infrared lines of neutral hydrogen. In this case there is a possibility that there is a contribution from infalling material (G. Gahm, private communication) as in the model of Sorelli et al. (1996) which would strongly affect the derived mass loss rates.
3.2. Emission line variability
The observations presented here, in conjunction with other published data (Finkenzeller 1985; Böhm & Catala 1994) provide a useful database for some comments on line variability in the [OI] 6300 line. For those stars observed by Finkenzeller (1985), comparisons of velocities and equivalent widths are possible over a time scale of roughly ten years. Of the six stars presented in Finkenzeller (1985) showing [OI] 6300 emission we have spectra of 5 stars; KK Oph, BD 4124, HD 259431, MWC 137, and HK Ori. Of the 13 stars observed by Böhm & Catala (1994), not taken from the sample of Finkenzeller (1985), we have spectra of 10 stars; BD 164, AB Aur, HD 259431, LkH 218, LkH 220, NX Pup, HD 97048 HD 200775, BD 3471 and R Mon. The variations in observed equivalent widths of all 14 stars (HD 259431 being present in all three samples) are presented in Table 3. Note that variation in the equivalent width may arise from changes in the continuum level and the variability discussed here may reflect this. However in those cases where a variation in velocity is also observed it is obvious that properties of the line emitting region are varying in addition to any contribution from variations in the stellar continuum level.
Table 3. Measurements of the equivalent width of the [OI] 6300 line in Herbig Ae/Be stars observed by Finkenzeller (1985), Böhm & Catala (1994) and ourselves. The date of observation is given in parentheses (day.month.year). Errors are 5% - 10%, depending on the individual observations. The absolute visual band variability, , where known, is taken from Finkenzeller & Mundt (1984).
Of the 14 stars, 9 show little or no variation ( Å) in the equivalent width of the [OI] 6300 line over timescales up to 10 years. The remaining five stars, LkH 218, LkH 220, BD 4124, R Mon and MWC 137, show variations ranging from a 30% reduction in ([OI]) over 10 years for BD 4124 to as much nearly a 300% increase in the line equivalent width in R Mon over six years. Given the length scales and velocities involved, it is possible to imagine a situation where the outflow characteristics vary with timescales of order 5-10 years, so perhaps the observed variations are not surprising. Note that although the sample presented here only contains one Hillenbrand Group II star (R Mon), this star shows the greatest variation in the observed equivalent width.
Of the 14 stars observed in common with Finkenzeller (1985) and Böhm & Catala (1994), published centroid velocities are available only for the data of Böhm & Catala (1994). Three stars show different centroid velocities in our observations to the observations of Böhm & Catala (1994). HD 259431 and NX Pup (CoD 3318) have changed by -20 kms and -40 kms respectively. HD 259431 shows a change from +7 kms on 22.10.1991 to -10 kms on 26.12.1991 and NX Pup shows a change from -15 kms in 30.01.1986 to -55 kms in 25.12.1991. Neither of these stars is observed to vary strongly in the equivalent width of the [OI] 6300 line. The only star showing a dramatic change in both the line profile and velocity is R Mon. From a strongly asymmetric and blueshifted line with a centroid velocity of -35 kms and a bluewing velocity of -140 kms on 30.01.1986 (Böhm & Catala, 1994) the profile had changed to a fairly symmetrical profile shifted to only -20 kms on 1.1.1991. As only R Mon shows variation both in the equivalent width and the centroid velocity of the [OI] 6300 line it is not clear whether the variability in the line width in the other stars here is due to continuum variability or an actual change in the line forming region.
3.3. Spectral type identification
Using standard star spectra taken simultaneously with the INT data from July 1993 we are able to make comparisons of the spectral types of a number of HAEBESs in our sample with those spectral types quoted in the literature (Finkenzeller & Mundt 1984; Thé et al. 1993). Of the stars examined (56), 5 stars show absorption features indicative of a distinctly different spectral type (more than 3 subclasses) to some values quoted in the literature. The spectral type of the stars was determined from the lines of HeI 4471, MgII 4481 (Lang 1991). It must be noted that a number of stars (e.g. V645 Cyg, MWC 1080) show P-Cygni profiles in many of the blue lines often used in determining a star's spectral type and more generally, levels of photospheric veiling such that we are unable to confirm the accuracy of the spectral type identifications for such stars usually quoted in the literature. The results of this investigation, and the spectral types cited in the literature are presented in Table 4.
© European Southern Observatory (ESO) 1997
Online publication: June 30, 1998