Constraints on Magnetized Outflow Models of Protostars

Magnetic fields are believed to play an important role in the collimation and driving of bipolar outflows from protostars. Polarimetric observations of star formation regions can now be used to trace the magnetic field directions, and thus test these models. Some ideal sources to observe are the youngest low-mass protostars (Class 0 sources, defined by André, Ward-Thompson & Barsony 1993), which have highly collimated and energetic flows. However, the cool protostars are only detected at far-infrared and submillimetre wavelengths, and their low flux densities present a considerable technical challenge for polarimetry. Only two Class 0 sources, VLA1623 and NGC1333-IRAS4A, have previously been detected in polarization (Holland et al. 1996; Minchin, Sandell & Murray 1995; Tamura, Hough & Hayashi 1995), both at the JCMT. In this article, we present our recent JCMT observations of four further Class 0 protostars, which together with the two earlier detections, represent almost all of the reasonably bright known sources (flux densities of a few Jy at 800 microns).

Our observations of IRAS16293-2422, L1448-IRS3, NGC1333-IRAS2 and HH24MMS were made in February 1995, using the Aberdeen/QMW polarimeter with UKT14. Integration times of about 1 to 2.5 hours per source were needed to detect polarization at the 1-4 % level. The directions of the magnetic fields could then be constrained to within typically +/- 10 degrees.

Figure 1: Correlation of , the difference in angle between the magnetic field and outflow directions, with , the angle between the bipolar outflow and the line-of-sight to the observer (only the most collimated outflow system is included for IRAS16293 and NGC1333-IRAS2, both of which have secondary outflows also).

Field orientations for the protostars

Models of magnetic fields around protostars generally show a net field that is symmetric, and aligned either with the outflow, or in the perpendicular direction, along the presumed disk plane. If the outflows in the six sources share a single magnetic collimation mechanism, we might expect a single preferred orientation, such as the field always lying along the flow axis, or always along the disk plane. Surprisingly, we found instead two examples of the former, three of the latter, and one intermediate case (which was HH24MMS, a rather anomalous source with an IR jet but no well defined molecular flow).

We considered the possibility that the observed field direction might depend on the orientation of the outflow - i.e. the angle of inclination to the observer's line of sight. So we estimated the angle between the bipolar outflow and the line-of-sight, , for each source, using the method of Cabrit & Bertout (1986), which involved modelling the spatial appearance of the blue and red lobes. Figure 1 shows the results of this analysis, by plotting against , where is the angle between the observed field and the outflow direction (HH24MMS is omitted as the method of finding the inclination cannot be applied to the IR jet). We found that, for outflows close to the plane of the sky, the magnetic field tends to lie perpendicular to the outflow direction. In contrast, for outflows directed closer to the line-of-sight to the observer, the magnetic field tends to lie parallel to the outflow direction.

This result can in fact be explained by a viewing angle effect, as illustrated in Figure 2. The field configuration shown represents a generic class of models which typically have a circular or spiral field in or near the disk plane, and either helical or linear field lines along the outflow axis. In the first sketch, the highly ordered disk field, seen edge-on, is dominant, so and are both ~90 degrees. In the second sketch, the disk field is seen almost face-on and, as it is circularly symmetric, produces negligible polarization summed over the beam. Then only the field component along the flow is important, and and are both ~0 degrees. Thus this model reproduces the results seen in Figure 1.

Figure 2: Sketch of magnetic field configuration which can reproduce net polarization either along the outflow axis or in the disk plane. The first sketch shows the system seen almost edge-on, and the second sketch shows the system seen almost face-on. The thin lines represent the magnetic field, and the thick lines the outflow boundary. The net observed field direction is shown below each sketch.

Evolution of the magnetic field.

The results above show for the first time that one class of magnetic models can explain the observed field orientations towards protostars. Another interesting new result is that the field evolves with protostellar age, even though all the sources are estimated to be only a few 10(4) years old. Figure 3 plots percentage polarization, p, for the Class 0 sources, versus the ratio L/L(1.3mm), which is a measure of source evolutionary stage (André, Ward-Thompson & Barsony 1993). Figure 3 shows a tendency for p to decrease in the more evolved protostars (higher L/L(1.3mm)).

Figure 3: Correlation between p and L/L(1.3mm), a measure of source evolutionary stage. Older protostars are observed to have a lower net polarization. The correlation is significant at the 95% confidence level.

In general, p is not related to field strength (Hildebrand 1988), and varies with field structure and grain properties in ways that are not well understood (Goodman 1996). Here, a simple interpretation of the magnitude of the polarization is offered: the field structure around the protostars may initially be very ordered, but then the field becomes progressively more disrupted as the outflow sweeps up ambient material. In a more disordered field, polarization components in different directions will tend to self-cancel, producing a lower polarization at later times, as observed. Recent observations (Bontemps et al. 1996) have shown that bipolar outflows evolve with time, becoming less energetic and less well collimated as they evolve. Our results show that the outflows also interact with the magnetic field, causing it to become less well ordered as the source evolves.

Conclusions

These new results imply that future models for bipolar outflow collimation will have to consider the greater tangling of the fields with time. As a follow-up, we hope to study the environments of a larger number of fainter protostars, using the next generation polarimeter in conjunction with SCUBA. We will then be able to map the magnetic field structures, and gain a much better understanding of how energetic outflows are collimated, and how they evolve and affect the magnetic field structure within their parent clouds.

J. S. Greaves(1), W. S. Holland(1) and D. Ward-Thompson(2)

(1) Joint Astronomy Centre, Hawaii

(2) Royal Observatory, Edinburgh

References

André, P., Ward-Thompson, D., Barsony, M., 1993, ApJ, 406, 122

Bontemps, S., Andre P., Tereby, S., Cabrit, S., 1996, A&A, in press

Cabrit, S., Bertout, C., 1986, ApJ, 307, 313

Goodman, A. A., 1996, in Polarimetry of the Interstellar Medium, ASP Conference Series Vol. 97, p.325

Hildebrand, R. H., 1988, QJRAS, 29, 327

Holland, W.S., Greaves, J.S., Ward-Thompson, D., André, P., 1996, A&A, 309, 267

Minchin, N. R., Sandell, G., Murray, A. G., 1995, A&A, 293, L61

Tamura, M., Hough, J. H., Hayashi, S. S., 1995, ApJ, 448, 346