4.1. Shock structure
4.1.1. HH 29
In HH 29a, the steepest intensity gradient (indicating the location of the shock front or the apex of the bow shock) is found on the side towards or closer to the originating source, IRS5. The same is found upon close inspection of features HH 29b & (very clearly) HH 29f and - to some extent - for feature HH 29c. This is what would be expected from a 'shocked cloudlet' (Schwartz, 1978) and is indicative of a thinner, higher velocity, wind acting on denser, lower velocity material. Comparing Figs. 1, 2 and 3 we see that the opposite is true for feature HH 29d, where the steepest gradient is found on the side of the object furthest away from IRS5. This is more what would would be expected from an 'interstellar bullet' (Norman & Silk, 1979), or from a denser wind interacting with a thinner, slower moving medium (Hartigan, 1989). Interpreting our result qualitatively in terms of these models, we can then hypothesize that HH 29 is best explained as an aggregate of density enhancements in the wind traveling outwards into the ambient medium. It interacts with the (thinner) medium that it is plowing into, and this causes the shock feature HH 29d on the side of HH 29 away from IRS5 (the originating source). At the same time HH 29 is overtaken by a thin, faster wind component (possibly identical to the atomic component permeating the outflow lobe cavity - see Giovanardi et al. 1992), which causes the shocks at the position of features HH 29a, HH 29f and HH 29b to have the appearance of 'shocked cloudlets', i.e. with the shocks steepest gradient on the side of the feature towards IRS5. Inspection of the H line profiles displayed in Fig. 3 at the positions of HH 29a and HH 29d show that they are entirely consistent with this scenario.
The shock in HH 29a has a velocity of 90 km s-1 as determined by flux ratios (FLP utilizing the models of HRH). This is significantly slower than the velocity of HH 29 found from proper motions, 150 km s-1 (Cudworth & Herbig, 1979), the 190 km s-1 derived from the line profile of H by Stocke et al. (1988), or the FWZI (240 km s-1) of H , at the position of HH 29a, given in this paper. The latter two results are thus indicative of shock velocities more than a factor 2 higher than what is derived from the flux ratios (using the same models of HRH where it is inherently assumed that width of the full line profile is equivalent to the shock velocity at the apex of the shock). The discrepancy between the VS determined from line flux ratios and that determined from spectral profiles is a common phenomenon when interpreting data of this nature in the context of bow shock models, and has hitherto not received an adequate explanation (e.g. Böhm, 1995). Comparing the data in Tables 4 and 6, we see that although the integral flux ratio H / [SII] for HH 29a is 2.5, when we look at the same ratio as a function of velocity it varies between 1 and 6, and this suggests one possible explanation.
4.1.2. Knot HH 29f and the two narrow components
Following the arguments in Sect. 3.1, we assume that the low velocity component (-15 km s-1) is (mostly) intrinsic to the object. It is found at the rest velocity of the molecular cloud L1551, and this suggests that it is due to recombination in a precursor to the bright shock in HH 29a. Since this component is quite faint in our data the level of excitation and electron density can only be estimated from an individual inspection of the [SII] and H spectra. We find values of ne ranging from 10 cm-3 to 300 cm-3. The level of excitation in the -15 km s-1 component is as high as 0.3, as defined by the ratio between [SII] to H .
As what concerns the high velocity component, we find from both the morphological appearance (compare Fig. 1 and Fig. 2) and from its velocity structure that it is most likely due to emission from the 'bridge' between HH 29a and the separate knot we have designated HH 29f. As mentioned above, inspection of the H spectrum obtained through slit 8 show the steepest gradient to appear also in this case on the side facing towards IRS5. HH 29f thus appears as a case of a 'shocked cloudlet'. The FWZI of the H profile at knot f is 150 km s-1. In the region between HH 29a and HH 29f, we find ne to be a few hundreds cm-3.
4.2. Excitation and ionization
For a first determination of the excitation conditions, we can use the results of FLP. Table 2 shows the values of the [OIII]/H and the [SII]6717Å+6731Å/H ratios. Comparing these values with the compilation of Raga et al. (1996) of 45 HH condensations, we find that HH 29 is one of the highest excitation Herbig Haro objects. This is also true individually for the a - d knots. HH 29b & c have unusual high [SII]6717Å+6731Å/H ratio of 1.3 & 0.9 respectively given the [OIII]/H value of 0.6 & 0.7 where the former show more of low excitation character. Strong [SII] relative to H is usually taken as an indication of low excitation shocks (see the Introduction), but is rarely found in combination with high values of [OIII]. We take this to indicate a combination of excitation conditions along the line of sight to the HH 29b & HH 29c features. The relatively strong [CaII] emission from the region of features b & c is consistent with this. Displaying the [SII]6717Å+6731Å/H ratio as a function of velocity (see below) confirms this hypothesis and show varying excitation conditions over velocity dispersions of 30 km s-1 - 50 km s-1.
The present results alone do not allow a determination of , but we can try to estimate it from the ratio of the [OII] 3726Å+3729Å flux (taken from FLP) to the 7318Å+7319Å+7330Å+7331Å flux. The blend between the red [OII] lines and [CaII] at 7323.88Å can be 'untangled' because of our high spectral resolution. A check of this procedure is allowed by a comparison between the theoretical and the measured value of the [CaII]7323.88Å to [CaII]7291Å ratio which should be equal to 0.67 (Brugel et al. , 1981). This is indeed the case (within our errors). Now, taking the UV-lines of [OII] fluxes from FLP, we find results (depending on the reddening applied to the UV-lines) of 5000K 7500K (Aller, 1984). In the paper of LHFC, we have shown HH 29 to be variable on time scales of less than 6 months in the satellite UV, however, and the range of derived here can only be considered meaningful as an indication.
4.3. H flux, the level of excitation and the electron density, ne, in a three dimensional representation
Based on the hypothesis that the emission is optically thin, the velocity profile of the lines in question contain information about the third dimension, i.e. along the line of sight. In this paper, we have determined the flux of H and the [SII]6717Å,6731Å lines in 20 km s-1 1 2 (the slit width) bins. Data for one of the slits (No. 1) is presented in table form in Tables 5, 6 & 7. F(H ) is a representation of the emission measure, albeit with the proviso that we can have departures from a pure case B recombination spectrum in regions of weak shock conditions. As a first approximation, we can assume that case B is valid where we either have (relatively) strong [OIII]5007 emission (from FLP) or high excitation/ionization conditions as determined by the H / [SII] ratio. This ratio is the second parameter that we have determined from our data. Finally, the [SII]6717 /[SII]6731 ratio is given as a measure of ne.
Table 5. Observed line fluxes in the H line along slit 1 at 1 intervals. The flux is integrated in 1 2 bins. Units of 10-15 erg cm-2 s-1
Table 6. Observed level of excitation along slit 1 at 1 intervals. The flux was integrated in 1 2 bins. The level of excitation is defined as the ratio of H to [SII]
Table 7. Observed ratio of the [SII] lines along slit 1 at 1 intervals. The flux was integrated in 1 2 bins.
We again see a structure where the physical conditions (as defined by these three quantities) change very rapidly. The highest value of ne (as defined by low values of the [SII]6717 /[SII]6731 ratio) is found at -50 to -90 km s-1 for knots a and d. The highest level of excitation (as defined by the H / [SII] ratio) for knot a & b is also found within this velocity band while lower levels are found both at more negative and more positive velocities. Knot d, however, have its highest level of excitation at velocities between +10 km s-1 and -50 km s-1. For knot b, the highest values of ne is found between -10 km s-1 and -50 km s-1. This all very strongly indicates that the 'clumps' are small structures.
The electron densities are derived by using a model for the 6717Å/[SII]6731Å ratio that incorporates 5 levels, Einstein A coefficients from Keenan et al. (1993) and new collisional coefficients calculated by Ramsbottom et al. (1996). This model was calculated for 3000K K and for 1 cm-3 ne cm-3.
Maps of the density as a function of velocity were created by integrating the line intensities (in arbitrary units) over velocity intervals of 10 km s-1. Then the ratio of the 6717Å to the 6731Å was taken and the above mentioned density model for a fixed temperature of 104 K was applied. A 2-dimensional cubic spline was then applied to each map which decreased the spatial resolution by a factor of 3. The maximum density at the position of HH 29a is 104 cm-3. Since this is the maximum density measurable with this method, this does not exclude higher values at this position. The value of ne drops rapidly to lower values behind HH 29a, so that 3 further 'downstream' from the intensity maximum we find densities of 500 cm-3. The general structure of HH 29 appears very 'clumpy', with smoothed size scales of typicaly 2 - 4 (corresponding to 300 - 600 AU). The average density is also 300 cm-3, with the density enhancements going well above several 103 cm-3 for positions near features b,c,d and e. At these positions themselves, as defined by FLP, we find lower densities, however. The appearence of these 'clumps' is morphologically very similar to the map of FLP, where density enhancements of several magnitudes where found with the same size scales. A comparison between Fig. 3 of FLP and our data shows the enhancements to be well offset (i.e. more than 3 - 4) from the peak ne 'clumps'.
The values of ne are also velocity dependent, (as can be clearly seen from e.g. Table 7 so that the most extreme positive and particularly negative radial velocities ( -100 km s-1), are represented by low values of ne.
A clumpy structure with ne varying between 300 cm-3 and 104 cm-3 in the peak of the densest clump is thus indicated. Structures found in HST imaging of the L1551 IRS5 jet (HH154) (Fridlund et al. , 1997a & 1997b) are an order of magnitude smaller than what is resolved in the ground based observations, and our calculation above can not exclude structural variations on such a small scale also being present in HH 29. On the contrary, the time series UV data of LHFC strongly suggests even smaller spatial scale structures being present. Nevertheless, our observations indicate a large filling factor whether originating in (relatively) large structures or in agglomerates of smaller structures. When we analyze the [SII]6717/6731 ratios as a function of velocity, we see large variations on scales as small as a few 10 km s-1. We take also this to be a suggestion of small (spatial) scale structural variations along the line of sight.
© European Southern Observatory (ESO) 1998
Online publication: January 8, 1998