A&A 485, 137-152 (2008)
DOI: 10.1051/0004-6361:200809515

Molecular jets driven by high-mass protostars: a detailed study of the IRAS 20126+4104 jet

A. Caratti o Garatti¹ - D. Froebrich² - J. Eislöffel¹ - T. Giannini³ - B. Nisini³

1 - Thüringer Landessternwarte Tautenburg, Sternwarte 5, 07778 Tautenburg, Germany
2 - Centre for Astrophysics and Planetary Science, University of Kent, Canterbury, CT2 7NH, UK
3 - INAF - Osservatorio Astronomico di Roma, via Frascati 33, 00040 Monte Porzio, Italy

Received 5 February 2008 / Accepted 13 March 2008

Abstract
Context. Protostellar jets from intermediate- and high-mass protostars provide an excellent opportunity to understand the mechanisms responsible for intermediate- and high-mass star-formation. A crucial question is if they are scaled-up versions of their low-mass counterparts. Such high-mass jets are relatively rare and, usually, they are distant and highly embedded in their parental clouds. The IRAS 20126+4104 molecular jet, driven by a 10 $^4~L_{\odot}$ protostar, represents a suitable target to investigate.
Aims. We present here an extensive analysis of this protostellar jet, deriving the kinematical, dynamical, and physical conditions of the H₂ gas along the flow.
Methods. The jet was investigated by means of near-IR H₂ and [Fe II] narrow-band imaging, high-resolution spectroscopy of the 1-0 S(1) line (2.12 $\mu$m), NIR (0.9-2.5 $\mu$m) low-resolution spectroscopy, along with ISO-SWS and LWS spectra (from 2.4 to 200 $\mu$m).
Results. The flow shows a complex morphology. In addition to the large-scale jet precession presented in previous studies, we detect a small-scale wiggling close to the source, which may indicate the presence of a multiple system. The peak radial velocities of the H₂ knots range from -42 to -14 km s^-1 in the blue lobe, and from -8 to 47 km s^-1 in the red lobe. The low-resolution spectra are rich in H₂ emission, and relatively faint [Fe II] (NIR), [O I] and [C II] (FIR) emission is observed in the region close to the source. A warm H₂ gas component has an average excitation temperature that ranges between 2000 K and 2500 K. Additionally, the ISO-SWS spectrum reveals a cold component (520 K) that strongly contributes to the radiative cooling of the flow and plays a major role in the dynamics of the flow. The estimated $L_{\rm H_2}$ of the jet is $8.2 \pm 0.7~L_{\odot}$, suggesting that IRAS 20126+4104 has a significantly increased accretion rate compared to low-mass YSOs. This is also supported by the derived mass flux rate from the H₂ lines ( $\dot{M}_{\rm out}$(H $_2)\sim 7.5\times 10^{-4}~M_{\odot}$ yr^-1). The comparison between the H₂ and the outflow parameters strongly indicates that the jet is driving the outflow, at least partially. As already found for low-mass protostellar jets, the measured H₂ outflow luminosity is tightly related to the source bolometric luminosity.
Conclusions. As for a few other intermediate- and high-mass protostellar jets in the literature, we conclude that IRAS 20126+4104 jet is a scaled-up version of low-mass protostellar counterparts.

Key words: stars: pre-main-sequence - infrared: ISM - ISM: jets and outflows - ISM: kinematics and dynamics - individual objects: IRAS 20126+4104

  
1 Introduction

Protostellar jets and outflows are a ubiquitous phenomenon among young stellar objects (YSOs) of different masses and luminosities (see e.g. Shepherd 2003). They are usually explained as a consequence of accretion from a disc around the protostar (see e.g. Pudritz & Norman 1986; Camenzind 1990). This is particularly true for low - and, partially, for intermediate - and high-mass YSOs up to $L_{\rm bol}\sim 10^4$ $L_{\odot}$ (or spectral type B0), where collimated outflows, often driven by protostellar jets, have been observed (Shepherd 2003). In contrast, no highly collimated outflow or circumstellar disc has been observed in high-mass protostars exceeding 10 $^5~L_{\odot}$ (or O-type stars, the spectral type and $L_{\rm bol}$ of the M17 disc silhouette are not clear yet) (see e.g. Arce et al. 2007; Zinnecker & Yorke 2007), and the formation mechanism of these latter objects is still being debated. The observations are, however, strongly limited by the large distance of the massive star-forming regions, the considerable extinction, and the short lifetime of massive YSOs. In addition, these objects are often grouped in small clusters, which confuses the morphology of massive star-forming regions even more. Therefore, optical and IR studies of intermediate- high-mass protostellar jets are very rare, and only a few examples are present in the literature: e.g. IRAS 18162-2048 ( $L_{\rm bol}\sim 2\times10^4~L_{\odot}$, Martí et al. 1995), IRAS 20126+4104 ( $L_{\rm bol}\sim 10^4~L_{\odot}$, Ayala et al. 1998), IRAS 16547-4247 ( $6~\times~10^4~L_{\odot}$, Brooks et al. 2003), IRAS 18151-1208 ( $2\times 10 ^4~L_{\odot}$, Davis et al. 2004), and IRAS 11101-5829 ( $10^4~L_{\odot}$, Gredel 2006), the M17 disc silhouette ( $M_{\rm YSO}\sim 15~M_{\odot}$, Nürnberger et al. 2007). Expanding the number of observations of intermediate- and high-mass jets and comparing their general properties with those of low-mass protostellar jets is therefore important for understanding whether differences exist, or if high-mass protostellar jets are just scaled up versions of their low-mass counterparts. We therefore investigated the kinematical and physical properties of the IRAS 20126+4104 jet by means of NIR narrow-band imaging, high-resolution and low-resolution IR spectroscopy. We then compared our findings with those of other high- and low-mass protostellar jets available in the literature.

IRAS 20126+4104, at a distance 1.7 kpc, is a very well studied high-mass YSO ( $M\sim 7~M_{\odot}$, Cesaroni et al. 1997, 2005) in a very early stage of evolution. It is accreting mass at a very high rate ( $\dot{M}_{\rm acc} \sim 2\times 10 ^{-3}~M_{\odot}$ yr^-1, Cesaroni et al. 2005) and it gives birth to a large poorly-collimated CO outflow. It harbours the first H₂ jet detected from a high mass YSO (Ayala et al. 1998), which had previously been seen in SiO emission close to the source (Cesaroni et al. 1999). The H₂ jet extends for about 1 pc. Its ``S'' shape morphology suggests that it is precessing with a period of $\sim$60 000 yr and with a wide precession angle of about 37 $\hbox{$^\circ$ }$ (Shepherd et al. 2000). As a consequence, the inclination of the flow with respect to the plane of the sky changes strongly from $\sim$9 $\hbox{$^\circ$ }$, close to the source, up to $\sim$45 $\hbox{$^\circ$ }$ in the outer part of the flow.

The structure of this paper is as follows. In Sect. 2 our observations are presented. Section 3 reports an overview of our results, including the morphology of the H₂ jet, its kinematics, the physical parameters of the gas, its energy, and mass flux rate. In Sect. 4 we briefly consider the cause of the newly detected small-scale precession mode. Then, we discuss our H₂ data in relation to the CO outflow literature data. Finally, we compare the properties of the IRAS 20126+4104 jet with other high- and low-mass protostellar jets.

  
2 Observations and data reduction

Our data were collected at the UK Infrared Telescope (UKIRT), and at the 3.5 m Italian Telescopio Nazionale Galileo (TNG). More data were retrieved from the Subaru and ISO archives. The relevant information on the observational settings is summarised in Table 1.

	Figure 1: Close up of the TNG H2 mosaic, showing the entire flow with superimposed positions of the slits and the ISO-SWS FoV. The positions of the H2 knots and the IRAS source ( $\alpha =20^{\rm h}14^{\rm m}26.03^{\rm s}$ and $\delta =41\degr13\hbox{$^\prime$ }32\hbox{$.\!\!^{\prime\prime}$ }$58 [J2000], Hofner et al. 2007) are also indicated.
Open with DEXTER

Table 2: Observed lines in the IRAS 20126+4104 outflow: knots A, B, C, X, jet-1, jet-4.

In Table 4 we report the results of the high-resolution H₂ spectroscopy. For each knot in each slit we indicate the H₂ radial velocities ( $v_{\rm rad}$), corrected for the cloud speed with respect to the LSR ( $v_{\rm LSR}=-3.5$ km s^-1, Cesaroni et al. 1997). Two or more velocity components are given where detected. The first reported value is the peak value. The last column gives the full width at zero intensity (FWZI) of the line profile, measured where the flux reaches a 2$\sigma$ background noise level (see e.g. Davis et al. 2001).

Position-velocity (P-V) diagrams are presented in Fig. 6 (CGS4 Slits 1 and 2) and in Fig. 7 (CGS4 Slits 3 and 4). In Fig. 8 the emission line profiles are shown. It is worth noting that our spectra often cannot resolve the single structures inside the wiggling jet, due to the limited spatial resolution. As a result, radial velocities are often an average over more than one substructure visible in the Subaru image.

Almost all the knots also have more than one velocity component in their spectral line profile, depending on the position of the slit with respect to the knot. However, such components are never completely resolved (with the exception of knot 2 along the ``jet'', see Fig. 8), but they instead show up like ``bumps'' along the smooth profile. When single-peaked profiles are observed, the lines are never symmetric, but evidence of extended line-wing emission (opposite to the blue- or redshifted peak) is always detected. This may be indicative of a bow shock morphology of the knots, producing both the wings and the different velocity components (see e.g. Davis et al. 2001; Schultz et al. 2005), as also indicated by the sub-millimetric observations of Su et al. (2007). Remarkably, knot B has four velocity components (see Table 4 and Fig. 8). The fourth component located at $\sim$64 km s^-1, and visible along the profile as a ``bump'' on the redshifted wing (see also knot X profile) cannot be explained in a bow shock context. This component could originate from a different flow, but we have no further evidence in support of this hypothesis.

The flux peaks of the knots range from -42 to -14 km s^-1 in the blue lobe, and from -8 to 47 km s^-1 in the red lobe (see also Table 4). Noticeably, knots X and A in the red lobe (Fig. 6, central panel) have a slightly negative peak velocity. This can be understood by considering that close to the source the axis of the flow has an inclination of $\sim$9 $\hbox{$^\circ$ }$ with respect to the plane of the sky, and the aperture angle of the precessing jet is $\sim$37 $\hbox{$^\circ$ }$. Consequently, even if the knots are located in the ``red'' lobe, they could have a negative $v_{\rm rad}$ (see also Su et al. 2007).

In both lobes, the absolute peak radial velocities of knots close to the source (A...C) are lower (0-30 km s^-1) than those located at greater distances (knot D and knots 1...4, 40-50 km s^-1). On the other hand, the FWZI of the line profiles decreases with distance (see Table 4 and Fig. 8). The line profiles of the knots close to the source are broader on average (110-180 km s^-1) than those far from the source (70-90 km s^-1). This could confirm that the inclination of the flow axis (with respect to the plane of the sky) is different in the two regions; i.e., it changes from $\sim$9 $\hbox{$^\circ$ }$, close to the source, to $\sim$45 $\hbox{$^\circ$ }$ further out, at knots 1...4 and knot D. On a smaller scale, radial velocities appear to oscillate. This is very visible in P-V diagrams of Slit 3, and marginally, for the redshifted knots in Slit 1.

Assuming two different values for the inclination, the spatial velocity of the knots ranges between 50 and 80 km s^-1.

	Figure 4: The 0.9-2.5 $\mu$m low-resolution spectrum of knot C in the IRAS 20126+4104 jet. An asterisk near the line identification marks the detections between 2 and 3 sigma. Telluric lines are indicated by the symbol ``$\oplus$''.
Open with DEXTER

3.4 Physical parameters of the gas

The wealth of molecular hydrogen lines detected in both MIR and NIR spectra allows us to perform a detailed study of the H₂ excitation in a high-mass jet, also taking the pure rotational lines into account, for the first time.

As a result, employing all the available H₂ ratios in our ro-vibrational diagrams, we have derived column densities, extinction, and temperature of the gas. Combining these parameters with the 2.12 $\mu$m flux obtained from the narrow-band imaging, we determined an accurate measurement of the H₂ luminosity ( $L_{\rm H_2}$) for each knot and for the entire flow (see Caratti o Garatti et al. 2006 for a detailed description of this procedure).

  
3.4.1 Ro-vibrational diagrams from NIR lines

As a first step in our analysis, we only employed the NIR lines. Line pairs originating from the same energy level should lie in the same position of the ro-vibrational diagram. By varying the extinction value ($A_{\rm v}$) and increasing the goodness of the fit (maximising the correlation coefficient), extinction and temperature can be evaluated simultaneously (see e.g. Giannini et al. 2004; Davis et al. 2004; Caratti o Garatti et al. 2006). To reduce the uncertainties, only the transitions with S/N >3 and not affected by blending with other lines were used. The result of this analysis is shown in Table 5, where the average excitation temperature, the A_v, and the column densities of the warm gas component are listed for each knot (Cols. 2-4). The ro-vibrational diagrams are shown in Fig. 9.

To compute the extinction in each knot, we selected all the available pairs of lines. From the spectrum of knot C 41 lines were used, giving 29 different ratios (i.e. the ratios are among the following groups of transitions: 1-0S(i),1-0Q(i+2); 2-0S(i),2-0Q(i+2),2-0O(i+4),2-1S(i); 3-1S(i),3-1Q(i+2),3-1O(i+4); with i=0, 1, 2...). The best fit was obtained excluding the 1-0S(i)/1-0Q(i+2) pairs, which deviated more than 3$\sigma$ from the average value. The final value is $A_{\rm v}=7.6\pm0.2$, and it has a very small error. For the other knots, the $A_{\rm v}$ ranges from 6 to 10 mag, with errors between 1 and 3 mag (see Table 5, Col. 3). For these knots, a few pairs of lines were available making the errors larger. With the exception of knot C, all the lines in our ro-vibrational diagrams are well-fitted by a single straight line (see Fig. 9), indicating a uniform temperature of the gas. The observed temperatures range from 2000 to 2500 K.

Table 3: Observed lines in the IRAS 20126+4104 outflow: ISO lines.

	Figure 5: ISO-SWS and LWS detected lines.
Open with DEXTER

On the other hand, the different H₂ lines detected in knot C cannot be fitted by a single line. The ro-vibrational diagram exhibits a typical curvature indicating the presence of a stratification in the gas temperature. A more elaborate model of a mixture of gas at two different temperatures can describe these data (see e.g. Giannini et al. 2002; Caratti o Garatti et al. 2006; Gredel 2007). The population densities of the lines coming from levels with an excitation energy up to $\sim$12 000 K are consistent with an excitation temperture of 2050 K. These lines indeed trace the warm component of the gas. Conversely, lines with a higher excitation energy (>12 000 K) are thermalised at a higher temperture of $\sim$5200 K. These lines trace the hot component of the gas. A single fit through all the lines only gives a measure of the ``averaged'' temperature ($\sim$3300 K) (see Fig. 9, top left panel). In our model the hot component is a fraction of the gas of about 8%.

Table 4: H₂ radial velocities of individual knots in the IRAS 20126+4104 flow from CGS4 observations.

Another important parameter that derives from ro-vibrational diagrams is the column density of the gas. The column densities of the warm H₂ component are reported in Table 5. They are very similar to those measured in other high-mass protostellar jets (10¹⁸-10¹⁹ cm^-2) (see e.g. Davis et al. 2004; Gredel 2006), and they are 1-2 orders of magnitude higher than the values observed in low-mass jets.

Figure 6: P-V diagram for slits 1 and 2. The left panel shows the H2 (TNG) image with the slits superimposed, and the position of the knots. Y-offsets are displayed from the position of the source. The contour levels are 3, 30, 50, 100, 130, 210, 235, 300, 420$\times$ the standard deviation to the mean background ( $\sigma \sim 2\times 10^{-17}$ erg s-1 cm-2 arcsec-2). The central and right panels display the uncalibrated H2 (2.12 $\mu$m) spectra from CGS4 (slits 1 and 2, respectively). The contour levels are 5, 20, 60, 125, 180, 300, 450$\times$ the standard deviation to the mean background.

Open with DEXTER

Finally, from the analysis of Fig. 9, it is worth noting that the H₂ gas along the flow is fully thermalised. That we detect no lines with $v \ge 6$ indicates that fluorescence mechanisms do not play an important role in the excitation (see e.g. Black & van Dishoeck 1987). Excitation by non-thermal processes like fluorescence in the presence of a UV field would cause the vibrational temperature to be different from the rotational temperatures, and thus lines belonging to different vibrational series should not be located on the same line (see e.g. Hora & Latter 1994). As a consequence, we should observe strong deviations from the smooth (or linear) Boltzmann distribution of our ro-vibrational diagrams (see also Gredel 2007). Apparently, this contrasts with results of Sect. 3.2.2, where the presence of a PDR was inferred. Such a PDR could produce enough far-UV radiation to excite the higher vibrational levels of the H₂ and induce a fluorescent emission (see e.g. Burton 1992), which we do not detect. To evaluate a possible contribution of the PDR to the observed lines, we used the ``PDR toolbox''. The observed [C II] 158 $\mu$m diffuse line intensity (considering an LWS beam-width of 80 $\hbox{$^{\prime\prime}$ }$) is $\sim$ $2.5 \times 10^{-4}$ erg s^-1 cm^-2 sr^-1. Assuming that all the emission arises from the PDR and that the density is 10⁵-10⁶ cm^-2 (Cesaroni et al. 1999), we obtain a relatively faint FUV flux of G₀ between 10 and 100 (where G₀ is measured in units of $1.6 \times 10^{-3}$ erg s^-1 cm^-2). Under these circumstances the PDR surface temperture does not exceed 100 K (Kaufman et al. 1999) and the contribution to the overall H₂ emission is negligible. For example, the diffuse line intensity of the 0-0 S(1) is between 10^-6 and 10^-5 erg s^-1 cm^-2 sr^-1, while we obtain $\sim$ $6\times 10 ^{-4}$ erg s^-1 cm^-2 sr^-1 from our observations (dividing by the ISO-SWS FoV). In the NIR the emission from the shock (at 2.12 $\mu$m) is four orders of magnitude greater. These results clearly indicate that the detected H₂ arises mostly from shocks.

Figure 7: Left P-V diagram for slit 3. On the left panel the H2 (TNG) image indicates the position of the slit and knots. Y offsets are displayed from the position of the star in the centre of the image. The contour levels are 3, 6, 12, 20, 30, $50 \times \sigma$. On the right, the uncalibrated H2 (2.12 $\mu$m) spectrum from CGS4 is reported. The contour levels are 3, 5, 7, 10, $13 \times \sigma$. Right P-V diagram for slit 4. Y offsets are displayed from the position of knot D. The image contour levels are 3, 6, 30, 50, 100, 130, 210, 235, 300, $420\times \sigma$. The levels of the spectra are 3, 5, 7, 10, $13 \times \sigma$.

Open with DEXTER

3.4.2 The cold H₂ component from MIR lines

As a second step in our analysis, we also used the ISO-SWS lines to study the cold component of the gas, which is usually traced by lines with excitation energy lower than 5000 K, namely the 0-0 S lines (with $J_{\rm fin}\le 5$) between 6 and 28 $\mu$m in the ISO-SWS spectrum.

Since we do not know the spatial extent of the emitting region, we matched the column densities of the 1-0 Q lines present in both NICS and SWS spectra, to intercalibrate the data in the ro-vibrational diagram. Indeed the measured ISO-SWS fluxes of the 1-0 Q lines are very close to those measured in knot C. Therefore we have assumed that its column density is representative of the warm H₂ component detected with ISO, as well. The result is shown in Fig. 10. The SWS lines coming from the v=0 and v=1 levels are overplotted on the ro-vibrational diagram of knot C (shown in Fig. 9, top left panel). Although there is some scatter among the v=1 data points (two 1-0 Q and 1-0 O lines have slightly higher and lower column densities, respectively), the overall fit is quite satisfactory. In particular, the 0-0 S lines reveal the presence of a cold gas component at about 520 K, which has a higher column density than the other components ( $N_{\rm H_2}\sim 9.7 \times 10 ^{21}$ cm^-2, see also Fig. 10).

To reproduce the observed ro-vibrational diagram, we calculated a theoretical H₂ spectrum for a mixture of three H₂ layers in LTE condition at a temperature of 520, 2050, and 5200 K. The adopted LTE code (see also Caratti o Garatti et al. 2006) computes the line intensities involving levels with $0 \le v \le 14$ and $0 \le J \le 29$ ( $E_{v,J}\le 50$ 000 K). Ro-vibrational energies were taken from Dabrowsy (1984) and the Einstein coefficients from Wolniewicz et al. (1998). We assumed an ortho/para ratio equal to three (as also evinced from our ro-vibrational diagrams). The resulting model is plotted in Fig. 10, and properly fits our observations. In this case the warm and hot components are only a small fraction of the total H₂ gas, i.e. less than 1%.

3.4.3 H₂ luminosity

In the last step of our analysis, we inferred $L_{\rm H_2}$. Once the physical parameters ($A_{\rm v}$, T) for each knot were derived, we dereddened the 2.12 $\mu$m flux obtained from the imaging (see Table 5, Col. 5), adopting the Rieke & Lebofsky (1985) reddening law, and computed the line ratios with the other H₂ lines by applying our radiative LTE code at $T=T_{\rm avg}$. From these ratios, the absolute intensities of individual lines were computed and the H₂ luminosity of each knot was derived (see Table 5, Col. 6). The sum of these values gives a total $L_{\rm H_2}$ of the flow of $4.6\pm0.3~L_{\odot}$. An average temperature reproduces the contribution of both the warm and hot gas components to the radiated energy in H₂ well (see Caratti o Garatti et al. 2006). We are not taking the luminosity of the ``cold'' component into account, which, however, here can be considerable due to the extremely high column density of the gas. In order to evaluate this, we assumed that the measured SWS flux from the 0-0 S lines is the total flux emitted by the shocked cold H<sub>2</sub> gas of the flow and that the contribution of non-thermal emission is negligible, as seen in Sect. 3.4.1. We then computed the luminosity of an LTE gas at T=520 K, using the 0-0 S(1) observed dereddened intensity to derive the absolute intensities of the remaining lines. The luminosity of the ``cold'' gas is $3.6 \pm 0.6~L_{\odot}$. Added to the previous estimate, this gives a total $L_{\rm H_2}$ for the entire flow of $8.2 \pm 0.7~L_{\odot}$.

  
3.5 Mass flux, mass, and energy of the H₂ jet

The physical and kinematical parameters inferred in the previous sections allow us to evaluate important kinematical and dynamical properties of the flow. Therefore, we estimated the mass flux rate of the flow ( $\dot{M}_{\rm out}$) from the H₂ in order to compare it with the mass outflow rate and the mass accretion rate of the protostar previously obtained from CO observations and models (see e.g. Shepherd et al. 2000; Cesaroni et al. 2005; Lebrón et al. 2006). The value of $\dot{M}_{\rm out}$(H₂) can be written as $\dot{M}_{\rm k} = 2\; \mu m_{\rm H} N_{\rm H_2} A v_{\rm t} / l_{\rm t}$, where $\mu$ is the average atomic weight, $m_{\rm H}$ the proton mass, $N_{\rm H_2}$ the H₂ column density, A the area of the H₂ knot, $v_{\rm t}$ the tangential velocity, and $l_{\rm t}$ the projected length of the knot (see e.g. Nisini et al. 2005; Podio et al. 2006; Antoniucci et al. 2007).

	Figure 8: Line profiles of the knots in the 1-0 S(1) line of H2 at 2.12 $\mu$m.
Open with DEXTER

	Figure 9: Rotational diagram of knots in the IRAS 20126+4104 jet obtained from the low-resolution NIR spectroscopy. Different symbols indicate lines coming from different vibrational levels, as coded in the upper right corner of the boxes. The inferred temperature and extinction are indicated in the upper right corner of the box.
Open with DEXTER

Using the 2.12 $\mu$m column density of Table 5, the radial velocities of Table 4, and assuming an average inclination for all the knots of 20 $\hbox{$^\circ$ }$ with respect to the plane of the sky, we estimate mass fluxes between 10^-6 and 10 $^{-8}~M_{\odot}$ yr^-1 (see Table 5, Col. 7). In fact, these values represent lower limits. From the ro-vibrational analysis, the cold gas column density is two orders of magnitude higher than the warm component (e.g. 0-0 S lines have a column density $N_{\rm H_2}\sim 9.7 \times 10 ^{21}$ cm^-2). Using this value, and assuming the same velocity and extension of the emitting area of the 2.12 $\mu$m line, we obtain $\dot{M}_{\rm out}$(H₂)  $\sim 7.5 \times 10 ^{-4}~M_{\odot}$ yr^-1.

It is worth noting, however, that this value should only be considered as an estimate of the mass flux, for two reasons. Firstly, we have set $l_{\rm t}$ and $v_{\rm t}$ as equal for both H₂ components. Probably, such an estimate is a lower limit, since $l_{\rm t}$ of the cold gas can be higher than the $l_{\rm t}$ measured at 2.12 $\mu$m. More important, the adopted mass flux formula depends on the assumption that $l_{\rm t}/v_{\rm t}$ is the cooling time ($t_{\rm c}$) of the shock. In this case, the kinematically determined $t_{\rm c}$ of knot C is $1.9 \times 10^{10}$ s. Also other methods have been used in the literature to derive $t_{\rm c}$. Davis et al. (2000) adopt the approximation $t_{\rm c} \sim 3 \times 10^{8} n_{6}^{-1}T_{3}^{-2.3}$ (in seconds), where n₆ is the density in units 10⁶ cm^-3 and T₃ the temperature in units of 1000 K (Smith & Brand 1990). In this way, adopting $n_{\rm H_2} =10^5$ cm^-3 (Cesaroni et al. 1999) and two different temperatures, for knot C, we would obtain $t_{\rm c}$(warm) = $6.1\times10^{8}$ s (with T=2000 K), and $t_{\rm c}$ (cold) = $1.5 \times 10 ^{10}$ s (with T=500 K). The $\dot{M}_{\rm out}$ of the cold component would then be remarkably similar to the previous estimate. Moreover, it is very close to the mass outflow rates derived by Shepherd et al. (2000) and Lebrón et al. (2006) from the CO mm analysis ( $8.1 \times 10 ^{-4}~M_{\odot}$ yr^-1, and $3.4 \times 10 ^{-3}~M_{\odot}$ yr^-1, respectively).

In addition, we can derive a another estimate of the mass flux from the luminosity of the [O I] line (at 63 $\mu$m) using the approximate formula $\dot{M}_{\rm out}\sim 10^{-4}\times$(L(63 $\mu$m)/ $L_{\odot})~M_{\odot}$ yr^-1 (see e.g. Hollenbach & McKee 1989; Liseau et al. 1997; Cabrit 2002). As a result, we get $\dot{M}_{\rm out}$(OI) =  $2\times10^{-4}~M_{\odot}$ yr^-1. Such an estimate is an upper limit, however, since part or all of the [O I] emission could originate from the PDR.

To derive the mass of each knot, we assumed $M_{\rm k} = 2 \mu m_{\rm H} m_{\rm H} N_{\rm H_2} A$. For knot D, where $N_{\rm H_2}$ was not measured, we assumed an average value of $4.4 \times 10^{18}$ cm^-2 from Table 4. The total mass ( $\Sigma_{k}M_{k}$) of the warm gas is $\sim$ $10^{-3}~M_{\odot}$. Assuming an average value for the inclination of the flow of 20 $\hbox{$^\circ$ }$, the dynamical timescale of the flow ( $\tau_{\rm d}$) is $\sim$ $1.3 \times 10 ^{4}$ yr, the total momentum $P(\Sigma_{k}M_{k}~v_{k}$) is 0.06 $M_{\odot}$ km s^-1, the kinetic energy E_k( $\frac{1}{2}\Sigma_{k}M_{k}~v_{k}^2$) is $5 \times 10 ^{43}$ erg, and the momentum flux $\dot{P}$ (P/$\tau_d$) is $5 \times 10 ^{-6}~M_{\odot}$ yr^-1 km s^-1. We note that these values must be considered as lower limits, since the mass of observable shocked warm H₂ is only a small fraction of the total shocked molecular hydrogen mass, as seen previously. If we compute these quantities by taking the column density of the cold component of the gas into account, we obtain the following crude estimates: $M \sim 0.6~M_{\odot}$ km s^-1, $P\sim 50~M_{\odot}$ km s^-1, $E_{\rm k} \sim 4 \times 10 ^{46}$ erg, $\dot{P}\sim 4 \times 10^{-3}~M_{\odot}$ yr^-1 km s^-1.

  
4 Discussion

  
4.1 A precessing jet model

As mentioned in Sect. 3.1, the inner region of the H₂ jet clearly shows a wiggling morphology, possibly indicating a small-scale precession, different from the long-period precession derived from the positions of the outer knots (Shepherd et al. 2000; Cesaroni et al. 2005).

Figure 10: Rotational diagram of knot C of the IRAS 20126+4104 jet obtained from the low-resolution spectroscopy. Different symbols indicate lines coming from different vibrational levels, as coded in the upper right corner of the box (v=0 and v=1 ISO lines are indicated by open and filled pentagons, respectively). The dashed line represents the theoretical population distribution for a combination of thermalised gas at three different temperatures (520, 2050, and 5200 K, see text). The derived temperature from the different vibrational levels is indicated in the upper right corner of the box. The extinction is reported as well.

Open with DEXTER

To investigate this new morphological feature, we used a simple jet model to fit the observed position of the inner knots (see e.g. Eislöffel et al. 1996; Texeira et al. 2007):

$$\left(\begin{array}{cc} \alpha \\ \delta \end{array} \right... ...l \sin(2 \pi l / \lambda + \chi_0 ) \\ l \end{array} \right)$$ (1)

where $\alpha_0$, $\delta_0$ are the coordinates of the source, $\psi$ the rotation angle, $\phi$ the precession amplitude (in radiants), l the distance from the source position, $\lambda$ the precession length scale, $\chi_0$ the initial phase at the source. The resulting model, shown in Fig. 11, was been obtained for values of $\psi= -60\hbox{$.\!\!^\circ$ }9$, $\phi=0.133$ ($\sim$7 $.\!\!^\circ$6), $\lambda\sim$10 $.\!\!^{\prime\prime}$9, and $\chi_0=81\hbox{$.\!\!^\circ$ }6$. The inclination of the flow with respect to the plane of the sky was not considered in the fit, since its value (close to the source) is relatively small ($\sim$9 $\hbox{$^\circ$ }$). For a constant jet velocity of 80 km s^-1, at a distance (D) of 1.7 kpc, we obtained a precession period of $\sim$1100 yr, which appears short in comparison with the value of $\sim$64 000 yr derived in Shepherd et al. (2000) and Cesaroni et al. (2005). We then tried to reproduce the complex motion of the flow, adding a second precession mode to our model, that should take the longer period traced by the outer knots into account. We were, however, not able to fairly match the pattern of the entire jet.

Finally, we can also give a very rough estimate of the density for the cold component, as well. If we assumed that the size for knot C along the line of sight were similar to the observed length ($\sim$10¹⁷ cm), we would obtain $n_{\rm H_2}\sim$10⁵ cm^-3, as in Cesaroni et al. (1999).

This could indicate that the origin of the two observed `precessions' is different or that the precession period changed over time. The long-period precession was interpreted by Shepherd et al. (2000) and Cesaroni et al. (2005) as caused by the interaction between the disc of IRAS 20126+4104 and a stellar companion of a few solar masses. However, it is unlikely that the small-period precession we observe is caused by the orbital motion of the jet source around its companion. Since the orbital radius is given by the following equation (see Anglada et al. 2007) $r=\frac{\lambda {\rm tan}(\phi) D}{2\pi}$, the resulting value is $r \sim 400$ AU. The radius of IRAS 20126+4104 disc was estimated around 800 AU (Cesaroni et al. 1997), meaning that the companion would intersect the disc. Alternatively, the precession could have changed with time. The ejection of a third companion in a hierarchical triple system, for example, might have led to the formation of a tighter binary system and to shortening of the orbital period. However, this hypothesis does not explain the different position angles on the sky measured in the two precession modes (-60 $.\!\!^\circ$9 and -37 $\hbox{$^\circ$ }$, respectively).

It seems more likely, then, that the complex axis-wandering is induced by tidal interactions of multiple stellar companions. Indeed IRAS 20126+4104 is not a single compact source, but is composed of a small cluster of YSOs (De Buizer 2007); therefore, the dynamics of such a system can be extremely complicated. It is indeed beyond of the scope of this paper to obtain a rigorous modelling for the dynamics of such a system.

Table 5: Physical parameters of the H₂ knots in IRAS 20126+4104 derived from the low-resolution spectroscopy and imaging.

	Figure 11: Precession model plotted over the H2-Br$\gamma$ image from Subaru. The model is for a precession angle of 7 $.\!\!^\circ$6 and a precession scale of 10 $.\!\!^{\prime\prime}$9. A cross indicates the position of the source.
Open with DEXTER

  
4.2 H₂ jet vs. CO outflow

Observations of protostellar jets in low-mass YSOs support the idea that their molecular outflows are driven and accelerated by the highly-collimated bipolar jets. Outflows from intermediate- and high-mass sources, on the other hand, are often poorly collimated, and only in a few cases is there clear evidence of a driving jet. This has often led to the suggestion that these outflows could be driven by a different mechanism, such as, a wide-angle radial wind (see e.g. Arce et al. 2007).

Apparently, IRAS 20126+4104 represents a particular example of poorly collimated outflow powered by a protostellar jet. The poor collimation of this outflow has been explained by the severe precession of the jet. It is thus interesting to investigate whether the properties of the IRAS 20126+4104 H₂ jet meet the physical parameters of the CO outflow, i.e. if the jet is powerful enough to accelerate the surrounding medium and drive the outflow. This is indeed a fundamental question, because the YSO accretion rate is often estimated from the mass flux rate of the outflow. When comparing CO and H₂ parameters, however, we have to keep in mind that the CO lines only give a time-integrated response over the lifetime of the jet, whereas the H₂ emission provides an ``instantaneous'' measure of these quantities.

Large-scale outflow properties from CO mm observations were extensively studied by several authors (see e.g. Shepherd et al. 2000; Lebrón et al. 2006). Since their computations are based on different assumptions, however, the results of Lebrón et al. (2006) are at least one order of magnitude larger than those of Shepherd et al. (2000).

Table 6: Comparison between H₂ jet and CO outflow physical properties of the IRAS 20126+4104 flow.

In Table 6 we compare our H₂ flow parameters (obtained including the cold component) with the parameters of the CO outflow from Shepherd et al. (2000) and Lebrón et al. (2006). From this table, we can infer that the jet is driving the outflow, at least partially. The radial velocities of H₂ are consistent with the CO velocities, including the high-velocity component up to the 60 km s^-1 detected by Lebrón et al. (2006). On the other hand, the total mass of the H₂ bullets is just a fraction of the total CO mass. As a consequence, the momentum (P) of the jet is at least one order of magnitude less than that of the outflow. However, our estimates of the mass flux, kinetic energy, and momentum flux are, indeed, almost coincident with Shepherd et al. (2000), supporting the hypothesis that the H₂ jet is fully driving the flow. In Contrast, the values from Lebrón et al. (2006) are from one to two orders of magnitude more than ours (see Table 6). This would imply that the H₂ jet is only partially driving the outflow and that there is an undetected component along the flow, likely a neutral jet, that is supplying further momentum flux to the outflow. However, it is worth noting that the mass flux obtained by Lebrón et al. (2006) is slightly higher than the mass accretion rate of IRAS 20126+4104 inferred by Cesaroni et al. (2005) ( $2\times10^{-3}~M_{\odot}$ yr^-1), and this is quite unlikely. This discrepancy could be then explained by the presence of (undetected) multiple jets (winds) driven by other YSOs near the source that would substantially contribute to power the CO outflow.

  
4.3 Shock conditions along the jet

As the previous results point out, the cold H₂ component of the jet plays a major role in the kinematics and dynamics of this outflow. It is therefore desirable to discuss the origin of this component in more detail.

A prevalence of C-type shocks along the flow could explain our findings. Indeed C-type shocks produce a high column density in the 0-0 lines (see e.g. Flower et al. 2003; McCoey et al. 2004), which are very sensitive to these shocks. The column densities in our 0-0 lines could not be explained by a J-type shock (see e.g. Le Bourlot et al. 2002). Except for the [Fe II] and the [O I] emission close to the source, the emission observed along the flow only comes from the H₂. Only the inner region near the source position exhibits hints of ionic emission. There is no evidence of the ionised jet detected in the radio by Hofner et al. (2007), probably confined to a region very close to the source and highly extincted. On the other hand, the complete absence of any ionic emission in the other knots of the flow should not be attributable to a high extinction value, since the measured $A_{\rm v}$ values along the jet are relatively low. Such a behaviour has already been observed in several low-mass protostellar jets. They usually show only H₂ emission, or, at most, a hint of ionic emission close to the exciting source (see e.g. Giannini et al. 2004; Caratti o Garatti et al. 2006). It is also clearly observed in the high-mass IRAS 18151-1208 jet spectra (Davis et al. 2004). They are interpreted as C-type shocks or as J-shocks with magnetic precursors, which are J-type shock waves evolving into ``continuous'' C-type shock waves (see e.g. Giannini et al. 2004). Our kinematical observations also seem to indicate the presence of C-type shocks. The H<sub>2</sub> velocities in C-type shocks can exceed 50-60 km s<sup>-1</sup> up to 80 km s<sup>-1</sup> (see e.g. Le Bourlot et al. 2002). Accordingly, the FWZI of H<sub>2</sub> bow shocks can be up to twice this value, depending on the inclination of the jet axis with respect to the plane of the sky (see e.g. Davis et al. 2004), i.e. the broadest lines are expected for a bow shock located in the plane of the sky. Actually, if we consider the different inclinations for the inner and the outer knots, our observations seem to indicate a high shock velocity in both regions, whereas the higher value of the FWZI is observed in the inner knots (closer to the sky plane), and a narrower shape in the outer knots. From these values we can roughly estimate a shock velocity between 40 and 80 km s<sup>-1</sup>, where the fastest speed would be attained in the inner region. Such high shock velocities (50-80 km s<sup>-1</sup>) imply, however, low pre-shock densities of 10<sup>3</sup>-10<sup>4</sup> cm<sup>-3</sup> (Le Bourlot et al. 2002) in C shocks, assuming a magnetic field B($\mu$G) =  b[n<sub>H</sub>(cm<sup>-3</sup>)]<sup>-0.5</sup>, with b=1. Higher densities lead to a lower maximum shock speed ( $v_{\rm diss}$), which is the maximum shock velocity that can be attained prior to the collisional dissociation of H<sub>2</sub>. In contrast, both CO and H<sub>2</sub> observations seem to indicate a higher pre-shock density (10<sup>5</sup>-10<sup>6</sup> cm<sup>-3</sup>) in the medium (see e.g. Cesaroni et al. 1999). In this case, $v_{\rm diss}$ would decrease to 30-50 km s<sup>-1</sup> (Le Bourlot et al. 2002). Thus a ``standard'' C-type model cannot entirely explain our findings close to the source. A larger transversal magnetic field (b >1) or a different shock model (such as a J-shock with a magnetic precursor) is probably needed. This last hypothesis would also explain the presence of the ionic emission in the knots close to the source.

  
4.4 L_H2, accretion and ejection rates

Several authors have argued that a tight correlation between the evolutionary properties of YSOs and their outflows exists. This is particularly true for those objects forming by disc accretion, since accretion and ejection should be regulated by the same mechanism. Recently, it has been shown for low-mass YSOs that the total H₂ luminosity of the jets is proportional to their accretion rates (see e.g., Smith 2002; Froebrich et al. 2003). In particular, an empirical relationship between $L_{\rm H_2}$ and $L_{\rm bol}$ ( $L_{\rm H_2} \propto~L_{\rm bol}^{0.58}$, Caratti o Garatti et al. 2006) was derived for a large sample of low-mass protostellar jets. This correlation holds for very young YSOs (Class 0 and some Class I), where the bolometric luminosity is mostly coincident with the accretion luminosity of the object and implies that $\dot{M}_{\rm acc}$ increases with the luminosity (i.e. mass) of the protostar. In Fig. 12 we report the results obtained from the sample of Caratti o Garatti et al. (2006). The diagram compares the measured outflow H₂luminosity versus the bolometric source luminosity, both on a logarithmic scale. We have included the new data from IRAS 20126+4104 (positioned in the upper right corner). The best fit previously obtained fits our object, perfectly, suggesting that the relationship also applies to more massive jets in their earliest stage of formation. According to the dynamical timescales of the outflow and the jet (a few times 10⁴ years), IRAS 20126+4104 has not yet reached the main sequence (MS) and has not yet developed any hypercompact HII (HCHII) region, which may affect the collimation of the jet/outflow system (Beuther & Shepherd 2005). Most importantly, the bolometric luminosity of the source mainly comes from accretion (see Cesaroni et al. 1999, 2005). In these works, the authors obtain for this source an accretion luminosity of $1.2\times 10 ^{4}~L_{\odot}$ for a mass accretion rate of $\sim$1- $2\times10^{-3}~M_{\odot}$ yr^-1. Moreover, that we obtain an ejection rate of 30-40% for the accretion rate from the H₂ strengthens the reliability of our findings, since for high-mass sources we would expect a ratio of the mass ejection/accretion rate higher than the value ($\sim$0.1) usually derived for the low-mass objects (see e.g. Cabrit 2007).

Figure 12: $L_{\rm H_2}$ vs. $L_{\rm bol}$ including the IRAS 20126+4104 jet (also considering the luminosity of the cold component, the new datapoint is located in the upper-right corner of the diagram). Values for IRAS 18151-1208 and IRAS 11101-5829 jets have been included as lower limits, as well (see discussion in the text). A dashed line indicates the previous fit from Caratti o Garatti et al. (2006).

Open with DEXTER

Finally, we compared our results with the $L_{\rm H_2}$ of two high-mass jets previously investigated by means of NIR spectroscopy, i.e. IRAS 18151-1208 ( $L_{\rm H_2}=0.7~L_{\odot}$, Davis et al. 2004) and IRAS 11101-5829 ( $L_{\rm H_2}\ge2~L_{\odot}$, Gredel 2006). These sources have almost the same $L_{\rm bol}$ as IRAS 20126+4104. The $L_{\rm H_2}$ estimates are quite close to the H₂ luminosity of IRAS 20126+4104 obtained from the warm H₂ component (i.e. from our NIR analysis, $L_{\rm H_2}=4.6 \pm 0.3~L_{\odot}$). For comparison, the values of the IRAS 18151-1208 and IRAS 11101-5829 jets were also included as lower limits in Fig. 12. It is also worth noting that the slightly lower value found by Davis et al. (2004) could be caused by the high extinction values (10-30 mag) observed towards this flow. Indeed an MIR investigation could reveal whether the cold H₂ component plays a major role in the cooling of those jets, as well.

  
4.5 Comparing high- and low-mass jets

The three high-mass protostellar jets spectroscopically investigated up to date share similar characteristics. Like the low-mass jets, they are collimated and powerful enough to drive their outflows. The jet is formed close to the source, as we observe e.g. in IRAS 20126+4104 knot X (located $\sim$1700 AU from the source) or in IRAS 11101-5829 knot HH136 J2 (located $\sim$3000 AU from the source), indicating that the collimation of the jet occurs close to the source. They show molecular and ionic shocked emission along the flow, and no evidence of fluorescent excitation. The observed jet velocities are similar to those of the CO outflow. Where the extinction is low, some jets (e.g. IRAS 11101-5829, IRAS 18162-2048) also show HH objects. Some kinematical and dynamical quantities, such as mass flux, momentum flux, luminosity, and kinetic energy, are greater, however, than in low-mass jets, because the powering YSO is more massive. Moreover, a precession-wiggling morphology is observed in most of the massive collimated H₂ jets, often with large precessing angles (up to $\sim$40 $\hbox{$^\circ$ }$ in IRAS 20126+4104) (e.g. IRAS 16547-4247, Brooks et al. 2003; IRAS 18151-1208, Davis et al. 2004; IRAS 11101-5829, Gredel 2006; M17 disc silhouette, Nürnberger et al. 2007; IRAS 07427-2400, IRAS 20293+3952, and IRAS 23033+5951, Nanda Kumar et al. 2002). Precession is also observed in low-mass jets, but it is neither as frequent nor as pronounced (usually the precession angles are less than 10 $\hbox{$^\circ$ }$). All this indicates that the dynamical interactions among massive stars are stronger and more frequent than in low-mass star-forming regions. Such complex dynamics could also explain the confused H₂ morphology of some jets and the lack of collimation in some massive outflows. Finally, the lack of jets and the poor outflow collimation observed in several massive young sources fits the evolutionary outflow scenario proposed by Beuther & Shepherd (2005) well. In this context, those sources that have jets detected toward them are very young (well before the MS turn-on), while those without detectable jets near the protostar have ultracompact HII (UCHII) regions. If the disappearance of a collimated jet in early B protostars stems from the presence of enhanced ionising radiation from an accreting early main sequence star, then all early B stars may be formed via accretion. In this sense, they are scaled-up versions of low-mass protostars at early phases (such as for IRAS 20126+4104, IRAS 18151-1208, IRAS 11101-5829, etc.). As YSOs evolve, developing UCHII regions that destroy the disc, the jets finally disappear, and the outflows look more like poorly collimated wind-blow bubbles (see also Arce et al. 2007).

In conclusion, the three high-mass protostellar jets spectroscopically studied appear to be scaled-up versions of the low-mass ones. Furthermore, a morphological analysis of the few intermediate- high-mass H₂ jets known up to now partially supports an evolutionary jet/outflow scenario. A larger sample of intermediate-/high-mass jets is, however, needed, and it would be premature to jump to the conclusion that the disc accretion-ejection paradigm can be extended to the intermediate- and high-mass protostars.

  
5 Conclusions

The IRAS 20126+4104 H₂ jet has been extensively investigated through near-IR H₂ and [Fe II] narrow-band imaging, H₂ high-resolution spectroscopy, along with low-resolution spectroscopy (0.9-200 $\mu$m) throughout the infrared wavelength range. The kinematical, dynamical, and physical conditions of the H₂ gas along the flow were probed. The main results of this work are the following:

-

A high angular resolution H₂ continuum-subtracted image from Subaru reveals a small-scale precession of the jet close to the source, with an angle of ($\sim$7 $.\!\!^\circ$6) and a period of $\sim$1100 yr. This is about a factor 50 shorter than the precession period deduced from large-scale H₂ images.

-

H₂ and [Fe II] narrow-band images show the appearance of a new knot, labelled X, roughly 1 $\hbox{$^{\prime\prime}$ }$ from the source position. No further ionic emission is detected along the flow in the narrow-band imaging, indicating that the jet is mainly molecular.

-

Low-resolution spectra are rich in H₂ emission, and no ionic emission is detected along the flow, with the exception of faint emission of [Fe II] close to the source position (in knots X and B). Faint [O I] and [C II] emissions are observed in the ISO-LWS spectrum. They could arise from an embedded PDR region around the source.

-

The peak radial velocities of the knots range from -42 to -14 km s^-1 in the blue lobe and from -8 to 47 km s^-1 in the red lobe. Their line profiles, very broad and often with two or three velocity components, seem to indicate a bow shock structure. In both lobes, the absolute peak radial velocities of knots close to the source (A...C) are slower (0-30 km s^-1) than those located at greater distances (i.e. knot D, and knots 1 to 4, 40-50 km s^-1). This possibly confirms a change in the flow inclination angle (with respect to the sky) from $\sim$9 $\hbox{$^\circ$ }$ (close to the source), to $\sim$45 $\hbox{$^\circ$ }$ (in the outer knots). Assuming these inclination values, the spatial velocity of the knots is between 50 and 80 km s^-1.

-

The ro-vibrational diagrams indicate H₂ excitation temperatures between 2000 and 2500 K. Stratification of temperature is detected only in knot C, which can be modelled combining a warm ( $T_{\rm ex}=2050$ K) and a hot ($\sim$5200 K) H₂ component. Additionally, the ISO-SWS spectrum reveals the presence of a cold component (520 K) with a high column density.

-

Furthermore, our analysis seems to indicate that the H₂ is mostly excited in C-type shocks, and no evidence of fluorescent excitation has been observed.

-

The estimated $L_{\rm H_2}$ is $8.1 \pm0.7~L_{\odot}$, where the cold component contributes about 50% to the whole radiative cooling. The high H₂ luminosity suggests that IRAS 20126+4104 has a significantly increased accretion rate compared to the low-mass YSOs. This is also supported by the measured mass flux rates from H₂ lines ( $\dot{M}_{\rm out}$(H $_2)\sim 7.5\times 10^{-4}~M_{\odot}$ yr^-1) matching the previous CO estimates. Our analysis also indicates that the cold H₂ component plays a major role in the kinematics and dynamics of this flow.

-

Comparing the H₂ and outflow parameters strongly indicates that the jet is driving the outflow, at least partially.

-

By comparing the measured luminosity of the H₂ jet with the source bolometric luminosity (assumed representative of the accretion luminosity), we show that IRAS 20126+4104 fits the correlation well between these two quantities already found for low-mass protostellar jets (Caratti o Garatti et al. 2006).

-

Considering our results and the literature data of a few intermediate- and high-mass protostellar jets, we conclude that these few jets appear to be scaled-up versions of their low-mass protostellar counterparts.

Acknowledgements

We would like to thank the referee, Debra Shepherd, for her helpful suggestions, which improved the manuscript.

The present work was supported in part by the European Community's Marie Curie Actions-Human Resource and Mobility within the JETSET (Jet Simulations, Experiments and Theory) network under contract MRTN-CT-2004 005592. Our observations have been funded by the Optical Infrared Coordination network (OPTICON), a major international collaboration supported by the Research Infrastructures Programme of the European Commission's Sixth Framework Programme.

The version of the ISO data presented in this paper corresponds to the Highly Processed Data Product (HPDP) set called 35500738 by W. Frieswijk et al. and HPDP set called 04300333 by Lloyd, Lerate, & Grundy, available for public use in the ISO Data Archive.

The United Kingdom Infrared Telescope is operated by the Joint Astronomy Centre on behalf of the Science and Technology Facilities Council of the United Kingdom.

Based in part on observations made with the Italian Telescopio Nazionale Galileo (TNG) operated on the island of La Palma by the Fundación Galileo Galilei of the INAF (Istituto Nazionale di Astrofisica) at the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofisica de Canarias.

Based in part on data collected at Subaru Telescope and obtained from the SMOKA, which is operated by the Astronomy Data Center, National Astronomical Observatory of Japan.

This research has also made use of NASA's Astrophysics Data System Bibliographic Services and the SIMBAD database, operated at CDS, Strasbourg, France, and the 2MASS data, obtained as part of the Two Micron All Sky Survey, a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation.

References

• Anglada, G., López, R., Estalella, R., et al. 2007, AJ, 133, 2799 [NASA ADS] [CrossRef] (In the text)
• Antoniucci, S., Nisini, B., Giannini, T., & Lorenzetti, D. 2008, A&A, 479, 503 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Arce, H. G., Shepherd, D., Gueth, F., et al. 2007, in Protostars and Planets V, ed. B. Reipurth, D. Jewitt, & K. Keil, 245 (In the text)
• Ayala, S., Curiel, S., Raga, A. C., et al. 1998, A&A, 332, 1055 [NASA ADS] (In the text)
• Baba, H., Yasuda, N., Ichikawa, S., et al. 2002, ADASS XI, ed. D. A. Bohlender, D. Durand, & T. H. Handley, ASP Conf. Ser., 281, 298
• Baffa, C., Comoretto, G., Gennari, S., et al. 2001, A&A, 378, 722 [NASA ADS] [CrossRef] [EDP Sciences]
• Beuther, H., & Shepherd, D. 2005, in Cores to Clusters: Star Formation with Next Generation Telescopes, ed. M. S. N. Kumar, M. Tafalla, & P. Caselli, 105 (In the text)
• Black, J. H., & van Dishoeck, E. 1987, ApJ, 322, 412 [NASA ADS] [CrossRef] (In the text)
• Bragg, S. L., Smith, W. H., & Brault, J. W. 1982, ApJ, 263, 999 [NASA ADS] [CrossRef]
• Brooks, K. J., Garay, G., Mardones, D., & Bronfman, L. 2003, ApJ, 594, L131 [NASA ADS] [CrossRef] (In the text)
• Burton, M. 1992, AuJPh 45, 463 [NASA ADS] (In the text)
• Cabrit, S. 2002, in Star Formation and the Physics of Young Stars, ed. J. Bouvier, & J.-P. Zahn, EAS Publ. Ser., 3, 147 (In the text)
• Cabrit, S. 2007, in Jets from Young Stars, Lect. Notes Phys. (Berlin, Heidelberg: Springer-Verlag), 723 (In the text)
• Camenzind, M. 1990, RvMA 3, 234 [NASA ADS] (In the text)
• Caratti o Garatti, A., Giannini, T., Nisini, B., & Lorenzetti, D. 2006, A&A, 449, 1077 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Cesaroni, R., Felli, M., Testi, L., et al. 1997, A&A, 325, 725 [NASA ADS] (In the text)
• Cesaroni, R., Felli, M., Jenness, T., et al. 1999, A&A, 345, 949 [NASA ADS] (In the text)
• Cesaroni, R., Neri, R., Olmi, L., et al. 2005, A&A, 434, 1039 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Clegg, P. E., Ade, P. A. R., Armand, C., et al. 1996, A&A, 315, L38 [NASA ADS]
• Davis, C. J., Berndsen, A., Smith, M. D., et al. 2000, MNRAS, 314, 241 [NASA ADS] [CrossRef] (In the text)
• Davis, C. J., Hodapp, K. W., & Desroches, L. 2001, A&A, 377, 285 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Davis, C. J., Varricatt, W. P., Todd, S. P., & Ramasy Howat, S. K. 2004, A&A, 425, 981 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Dabroski, I. 1984, Can. J. Phys., 62, 1634 [NASA ADS] (In the text)
• de Graauw, T., Haser L. N., Beintema, D. A., et al. 1996, A&A, 315, L49 [NASA ADS]
• De Buizer, J. M. 2007, ApJ, 654, L147 [NASA ADS] [CrossRef] (In the text)
• Eislöffel, J., Smith, M. D., Davis, C. J., & Ray, T. P. 1996, AJ, 112, 2086 [NASA ADS] [CrossRef] (In the text)
• Flower, D. R., Le Bourlot, J., Pineau des Forêts, G., & Cabrit, S. 2003, MNRAS, 341, 70 [NASA ADS] [CrossRef] (In the text)
• Froebrich, D., Smith, M. D., & Eislöffel, J. 2002, A&A, 385, 239 [NASA ADS] [CrossRef] [EDP Sciences]
• Froebrich, D., Smith, M. D., Hodapp, K. W., & Eislöffel, J. 2003, MNRAS, 346, 163 [NASA ADS] [CrossRef] (In the text)
• Giannini, T., Nisini, B., Caratti o Garatti, A., & Lorenzetti, D. 2002, ApJ, 570, L33 [NASA ADS] [CrossRef] (In the text)
• Giannini, T., McCoey, C., Caratti o Garatti, A., et al. 2004, A&A, 419, 999 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Giannini, T., McCoey, C., Nisini, B., et al. 2006, 459, 821
• Gredel, R. 2006, A&A, 457, 157 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Gredel, R. 2007, A&A, 474, 941 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Hofner, P., Cesaroni, R., Olmi, L., et al. 2007, A&A, 465, 197 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Hollenbach, D., & McKee, C. F. 1989, ApJ, 342, 306 [NASA ADS] [CrossRef] (In the text)
• Hora, J. L., & Latter, W. B. 1994, ApJ, 437, 281 [NASA ADS] [CrossRef] (In the text)
• Hunt, L. K., Mannucci, F., Testi, L., et al. 1998, AJ, 115, 2594 [NASA ADS] [CrossRef]
• Kaufman, M. J., Wolfire, M. G., Hollenbach, D., & Luhman, M. L. 1999, ApJ, 527, 795 [NASA ADS] [CrossRef] (In the text)
• Le Bourlot, J., Pineau des Forêts, G., Flower, D. R., & Cabrit, S. 2002, MNRAS, 332, 985 [NASA ADS] [CrossRef] (In the text)
• Lebrón, M., Beuther, H., Schilke, P., & Stanke, Th. 2006, A&A, 448, 1037 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Liseau, R., Giannini, T., Nisini, B., et al. 1997, in Herbig-Haro Flows and the Birth of Low Mass Stars, ed. B. Reipurth, & C. Bertout, 111 (In the text)
• Martí, J., Rodríguez, L. F., & Reipurth, B. 1995, ApJ, 449, 184 [NASA ADS] [CrossRef] (In the text)
• McCoey, C., Giannini, T., Flower, D. R., & Caratti o Garatti, A. 2004, MNRAS, 353, 813 [NASA ADS] [CrossRef] (In the text)
• Mountain, C. M., Robertson, D., Lee, T. J., & Wade, R. 1990, in Proc. SPIE, 1235, 25
• Murakawa, K., Suto, H., Tamura, M., et al. 2003, in Proc. SPIE, 4841, 881
• Kumar, M. S. Nanda, Bachiller, R., & Davis, C. J. 2002, ApJ, 576, 313 [NASA ADS] [CrossRef] (In the text)
• Nisini, B., Caratti o Garatti, A., Giannini, T., & Lorenzetti, D. 2002, A&A, 393, 1035 [NASA ADS] [CrossRef] [EDP Sciences]
• Nisini, B., Bacciotti, F., Giannini, T., et al. 2005, A&A, 441, 159 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Nürnberger, D. E. A., Chini, R., Eisenhauer, F., et al. 2007, A&A, 465, 931 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Podio, L., Bacciotti, F., Nisini, B., et al. 2006, A&A, 456, 189 [NASA ADS] [CrossRef] [EDP Sciences] (In the text)
• Pudritz, R. E., & Norman, C. A. 1986, AJ, 301, 571 [NASA ADS] [CrossRef] (In the text)
• Ramsay Howat, S. K., Todd, S., Leggett, S., et al. 2004, in Proc. SPIE, 5492, 1160
• Rieke, G. H., & Lebofsky, M. J. 1985, ApJ, 288, 618 [NASA ADS] [CrossRef] (In the text)
• Rousselot, P., Lidman, C., Cuby, J-G., Moreels, G., & Monnet, G. 2000, A&A, 354, 1134 [NASA ADS]
• Schultz, A. S. B., Burton, M. G., & Brand, P. W. J. L. 2005, MNRAS, 358, 1195 [NASA ADS] [CrossRef] (In the text)
• Shepherd, D. S. 2003, in Massive Star Birth: A Crossroads of Astrophysics, ed. R. Cesaroni, M. Felli, E. Churchwell, & M. Walmsley (Cambridge: Cambridge University Press), Proc. IAU Symp., 227, 237 (In the text)
• Shepherd, D. S., Yu, K. C., Bally, J., & Testi, L. 2000, ApJ, 535, 883 [CrossRef] (In the text)
• Smith, M. D. 2002, in The Origins of Stars and PLanets; the VLT View, ESO Astrophysics Symposia, ed. J. Alves, & M. McCaughrean (In the text)
• Smith, M. D., & Brand, P. W. J. L. 1990, MNRAS, 314, 108 [NASA ADS] (In the text)
• Sridharan, T. K., Williams, S. J., & Fuller, G. A. 2005, ApJ, 631, L73 [NASA ADS] [CrossRef]
• Su, Y.-N., Liu, S.-Y., Chen, H.-R., et al. 2007, ApJ, 671, 571 [NASA ADS] [CrossRef] (In the text)
• Teixeira, P. S., McCoey, C., Fich, M., & Lada, C. J. 2007, MNRAS, in press (In the text)
• Wolniewicz, L., Simbiotin, J., & Dalgarno, A. 1998, ApJS, 115, 293 [NASA ADS] [CrossRef] (In the text)
• Zinnecker, H., & Yorke, H. W. 2007, ARA&A, 45, 481 [NASA ADS] [CrossRef] (In the text)