Magnetic Fields from Molecular Lines - A First for JCMT

The line polarization arises in a similar way to that in maser sources, but can in fact occur in ANY molecular gas where there is a magnetic field. The direction of the polarization on the sky can be used to

infer the direction of magnetic fields in the clouds (see e.g. Kylafis 1983, ApJ 267, 137).

An example of the results is shown in Figure 1. The upper spectrum is of CO J=2-1 in DR21 (multiplied by a scale factor of 0.01). The two lower spectra are the 'Stokes parameters' Q and U, in polarimetry terminology, which represent orthogonal parts of the polarization vector. The direction of the polarization is 1/2 tan-1 (U/Q), which for this spectrum is approximately 125 degrees. It is clear that the polarization is detected at a good signal-to-noise. Another confirming factor is that the inferred magnetic field direction is only 10 degrees away from the net field in the area as shown by DUST polarimetry (JCMT results at 800 microns by Minchin & Murray 1994, A&A 286, 579), a difference which is within the errors.

The line polarimetry observations are quite straightforward to do. The polarimeter module, which is generally used with UKT14, is mounted onto the RxA2 frame, with a suitable waveplate for the frequency of the line. We used the 1100 micron waveplate (which has an estimated 94 % polarizing efficiency at 230 GHz), but the 800 micron waveplate

could also be used with RxB3i. The installation process is relatively straightforward. Spectra are then observed in the normal way, with an ICL routine used to move the waveplate between integrations. The data can then be analysed by subtracting pairs of spectra at different waveplate positions to find the polarized components, as in Figure 1.

Figure 1. CO J=2-1 spectra of the DR21 cloud core. (Top:) Total intensity I, multiplied by 0.01; (centre:) Stokes-Q spectrum; (bottom:) Stokes-U spectrum. The polarized components are binned to 1.6 km s-1 resolution. All intensities are in TA*, and the polarization is about 1 % (this can be seen from the U component being similar in intensity to I x 0.01).

The data in Figure 1 were obtained in about 30 minutes of integration. We are still investigating the limiting sensitivity of the system, but as an example, polarization has been detected at the 3 sigma level in a dark cloud core where the integrated TA*dv was only 50 K km s-1. Polarization levels as low as 0.4 % have been successfully measured in brighter lines. The instrumental polarization level is about 1 %, and has been measured to a 0.1 % error; it appears to be constant across the receiver passband. It is likely to be this accuracy in subtracting off instrumental effects which has made this experiment succeed, in contrast to earlier efforts (e.g. Wannier, Scoville & Barvainis 1983; ApJ 267, 126 or Lis et al. 1988; ApJ 328, 304).

This technique is potentially very useful for probing magnetic fields in interstellar clouds. In particular, where the velocity structure of the cloud is fairly well understood (a simple case such as rotation, for example), changes in the polarization across the line profile may give a three-dimensional picture of the magnetic field. This effect is clearly seen in our Galactic Centre data (paper in prep.). This information is not obtainable in any other way, and thus the line technique is both a facility unique to JCMT, and a unique probe of magnetised clouds.

Jane Greaves, Wayne Holland, Per Friberg & Bill Dent, JAC.