From the 3.6 cm continuum VLA maps of the Serpens triple source, we have derived the parameters for an ejection velocity variability and a precession of the outflow axis that would be necessary for explaining the observed proper motions, knot spacings, and curved structure of the NW lobe of this object. With these parameters, we then compute a numerical simulation of a jet, from which we obtain predicted 3.6 cm free-free continuum maps which can be directly compared with the VLA observations.
The predicted maps show a morphology of emitting knots aligned along a curved trajectory, which agrees qualitatively well with the observations of the Serpens jet. This is not surprising, as the ejection velocity variability and the precession of the model jet were specifically chosen so as to obtain as good an agreement as possible.
However, we should point out that the 3.6 cm fluxes predicted from our model jet are considerably fainter than the ones obsered for the Serpens jet. For example, the brighter knots in the model jet would have fluxes (at a distance of 300 pc) of mJy/beam (if the knots are unresolved). This is approximately a factor of lower than the fluxes of the brighter knots in the NW lobe of the Serpens jet (see Fig. 1).
This discrepancy between the model and the observations can be removed by increasing the jet density, and/or by increasing the mean velocity and amplitude of the time-dependent ejection velocity variability. For example, an increase of an order of magnitude in the jet density, and an amplitude for the ejection velocity variability about 50% larger than the one used in the numerical simulation, would bring the predicted fluxes up to the values observed in the Serpens jet. This results from the fact that the radio continuum scales approximately like the H intensity, which is approximately proportional to the preshock density times the shock velocity to the power 3.8 (see Raga & Kofman 1992). Such changes in the jet parameters would be reasonable, since for our numerical simulation we have chosen the lower boundary of the density range determined by Curiel et al. (1993), and for the velocity we have chosen the proper motions determined for the Serpens jet (the real velocities of course being higher as a result of projection effects).
Interestingly, the knots in the model jet show strong variabilities over periods of -10 years (see Fig. 3-Fig. 5). Due to the complex shock structures that result from the ejection variability, a given knot can have a series of brightening and fading away episodes over a time span of a few decades.
These complex time histories for the different knots result in qualitatively different morphologies for the jet. For example, at some times the jet has a bright, leading condensation, trailed by a chain of fainter knots closer to the source. Quite rapidly, this configuration can evolve into a structure in which one of the intermediate knots is completely dominant, or into a structure with several knots of comparable intensities (see Fig. 3-Fig. 5).
The VLA maps of the Serpens triple source appear to show dramatic variabilities, which qualitatively resemble the ones of the model jet. For example, in 1990 the leading condensation of the NW lobe was much brighter than the rest of the knots, and by 1998 one of the intermediate knots had become dominant (see Fig. 1). The timescale of this variability is in good agreement with the model predictions.
This result is most interesting, because one of the important features of models of jets from variable sources is the resulting variability of the predicted jet structure. Even though detailed models for at least one HH jet have been computed (HH 34, Raga & Noriega-Crespo 1998), the comparison with the observations has not included an analysis of the observed time-evolution. Actually, very few observations of the time-variability of the optical emission of HH jets have been made, since the relevant timescales are of a few decades (Herbig 1969).
The situation is different for the very compact radio continuum jet in Serpens, for which the variability timescales are of only a few years. These timescales are similar to the ones found in some of the "microjets" from T Tauri stars (Burrows et al. 1996; Lavalley-Fouquet et al. 2000). The shorter timescales relevant for these objects do allow observational studies of the evolution of the intensities of the knots along the jets, and direct comparisons with predictions from time-dependent jet models. As far as we are aware, the work described above is the first effort at carrying out such comparisons between the predicted and observed time history of a jet from a young star.
© European Southern Observatory (ESO) 2000
Online publication: January 29, 2001