|
|||
| Joint Astronomy Centre |
|
|
|
| Show document only |
The UK/Japan SCUBA Polarimeter is an add-on instrument used to detect linear polarization, typically of dust or synchrotron emission. It consists of a rotating quartz half-waveplate and a fixed analyser. The waveplate is housed in a module with a motor and an optosnsor, which control the waveplate movement. In operation, SCUBA sees only one plane of rotation (because the other is absorbed by the analyser) and the source rotation plane is rotated with respect to the acepted plane as the waveplate spins. The resulting modulated signal is used to deduce the source polarization charcteristics - the percentage and direction of linear polarization. The current observing modes are photometry (850 and 450 microns) and jiggle-mapping (typically just 850 microns, although 450 microns is also possible in good weather). A general explanation of these observing modes can be found here. Photometric polarimetry is the most suitable mode for faint point-like sources. There is now sufficient test data that instrumental effects can be subtracted down to levels of less than 0.1% at 850 microns. For jiggle-map polarimetry we have been able to compile a large database to assist in removing instrumental polarization on a range of scales, to accuracies of about 0.25% at 850 micron, and up to 1% at 450 micron due to lack of instrumental data in good weather. Generally we know less about the quality of observations at 450 microns than at 850 microns, and this should be borne in mind when considering proposing for time. Scan-map polarimetry is not a supported observing mode at present. The sensitivity of SCUBA with the polarimeter fitted is of course worse than without the polarimeter. Generally, sources fainter than 0.1 Jy (or Jy/beam) at 850 microns are unlikely to be feasible, unless very polarized. The SCUBA Integration Time Calculator estimates integration times for polarimetry in photometry mode, and you should enter the rms polarized-flux (e.g. for 5-sigma on a 2% polarized / 1 Jy source, the rms is 1000 x 0.02 /5 = 4 mJy). The observational overhead included in the calculator result is probably optimistic - add an observational overhead of ~ 150% to the integration time. For jiggle-map polarimetry, as with ordinary jiggle-maps, multiply the photometry integration times by 16 (for half-beam resolution, as comes from a standard jiggle-map) or 4 (if smoothing data to beam-width spacing) and then add an observational overhead of ~ 50-100%. For both modes, the likely calibration overhead depends on whether you are interested in the absolute flux as well - in which case you should add about 50% of your total time. If you are only interested in the relative polarization of your source rather than its flux rightness, you only need add about 1 hour for every anticipated shift, which will account for the time required to fix the polarimeter in place as well as pointing, focusing observations etc.
During an engineering period in July 2003, while raising the membrane in order to gain access to the secondary mirror, the membrane split at the seams, requiring its replacement with a spare. The instrumental polarization (IP) resulting from the new membrane was found to be substantially (~40%) higher than with the old one. Tests conducted at the same time (Aug/Sep 2003) found that by mounting a "depolarizer" (essentially a piece of membrane mounted in the optical path with fibers oriented perpendicular to the membrane over the observatory opening), the IP was found to decrease at 850 µm by ~1/3, and at 450 µm by a factor of 2-3. However, later data suggest that the depolarizer does not have a zero IP itself, and most importantly, that it is non-uniform in strength and direction across the array. We are currently doing tests to evaluate the effect of the depolariser. Hence our current recommendation is not to use the depolariser when observing. We are also still in the process of measuring the IP due to the new membrane across the array at 850 and 450 µm. (Note that this means that 450 µm polarimetry will be taken simultaneous with 850 µm, although the 450 µm will not be fully sampled.) Any data taken before the IP is completely characterized will need to be re-reduced, but all PIs affected will be alerted when this is completed. To reduce polarimetry data, follow one of the following links:
|
