Minutes of Polarization Teleconference Thursday, 4 January 2001

Attendees: Lamb, Mangum, Menten, Crutcher, Wootten, Guilloteau, Emerson, Baars, Radford, Butler, Napier, Brogan, Myers, Wild, D'Addario

A teleconference was arranged at the suggestion of R. Crutcher to open a line of communication between scientists and engineers within ALMA. Crutcher expressed some concerns of his to which email responses were circulated by D'Addario, Lamb and Myers prior to the teleconference. Here, I attempt to summarize the discussions.

1. One suggestion is that half of the receiver feeds be rotated by 45 degrees. This allows measurement of all four Stokes parameters. One possible drawback is that four correlations would be required all the time, reducing maximum bandwidth through the correlator. This has been implemented at Westerbork, where it appears to have worked pretty well (since 1985, feeds have been parallel, though there was a brief test of the Hamaker scheme about 5 years ago). Hamaker has discussed details of the theory of this scheme, discussed in the September ASAC report. (Johan Hamaker; see Astron. & Ap. Suppl., 143, 515, 2000)

D'Addario noted that nothing within the receiver package can be rotated in the present design. Half of a receiver complement could be installed at a 45 degree offset, but this would be a permanent arrangement. The array loses a factor of two in sensitivity.

Napier pointed out that the AT had considered this but went to a system similar to the ALMA baseline design, driven by calibration needs.

Discussion ensued on the problem of translating scientific specifications-for instance that to measure polarization to 0.1%-into engineering specifications. Lamb suggested a document which discussed this problem including effects such as receiver gain fluctuations, antenna sag and similar difficulties. D'Addario suggested that at millimeter wavelengths it would be difficult to achieve better complex gain stability than is achieved at the VLA, though calibration might be more accurate. Extended sources will still be difficult; any one of the myriad of antennas may constrain accuracy. Guilloteau pointed out that this was a computational difficulty to some extent.

Discussion turned briefly to the choice of linear over circular polarization for the receivers, and to the front end design, in which no receiver is on-axis. Guilloteau noted that the high standards for stability set for ALMA should make problems with beam squint more tractable. Lamb noted that receiver design has concentrated on trying to make the receivers as identical as possible, achieving the highest sensitivity and stability. For example, there are stringent tolerances on optics alignment. Great effort has been expended to make large clearances for the beams to maximize efficiency, reducing the strong polarization-dependent diffraction effects at apertures. D'Addario noted that refocusing optics are needed e.g. to keep the receiver horn size reasonable. Refocusing is accomplished with a pair of offset mirrors for most bands; if a lens design were adopted instead this would introduce excessive loss and loss of bandwidth at higher frequencies.

Calibration strategy dominated discussion next. The laser generated tone injected at the secondary is under development by Vaccari, Shillue and Emerson. D'Addario noted the value of a strongly phase coherent signal this provides between the two polarization channels of the same antenna; however this is accomplished at a single sweepable frequency rather than in a broadband mode. Lamb noted that the interpolarization phase on any antenna is likely to be an order of magnitude better than the inter-antenna phase stability. Calibration of relative linear phases on an antenna may be done by observing an astronomical calibrator through a grid during a calibration step. Lamb enumerated performance measures useful in trying to decide on the best attributes for polarization measurements. If antennas differ substantially(as may occur in the proposed enhanced ALMA possible with Japanese participation) the calibration task will be more difficult. Guilloteau noted that the calibration strategy is a large task, and should commence with consideration of a well-defined problem, such as linear polarization with known optics.

Myers asked when stability tests on prototype receiver optics commence. Apparently, this occurs late, after the antennas have been committed to production. It was not clear how much time is available between first tests on a prototype antenna with pre-production receivers, and release of the receivers to production. D'Addario noted that the test interferometer will certainly be operated in polarization modes, at least for the evaluation procedures. Prototype array receivers should be tested also but plans are not firm at present. Myers suggested placing some priority on tests of the primary polarization band.

D'Addario asked why the ASAC had focussed on the 350 GHz band. Crutcher noted that polarized dust and linear polarization mapping drive that choice; dust emission is stronger at higher frequencies. Guilloteau noted that perhaps 280GHz offered advantages, as one was also driven to somewhat lower frequencies by the need for substantial field of view.

What receiver optimization can be done. There are two competing designs for the 350GHz band (Band 7), involving grids or ortho mode transducers; it isn't clear which offers better performance in terms of orthogonality, stability and uniformity. Lamb noted that tolerances have been defined, but stability specifications need work. D'Addario pointed out the problems brought about by replacement of parts; it was noted that experience at BIMA is that calibration can be achieved to a few percent but must by redone whenever anything changes on an antenna. The translation from stability specification to polarization specification is very hard to define; Lamb noted that the goal now is to do the best possible job at constraining stability. This includes avoidance of moving parts to foster repeatability and reliability, minimizing optical components, to foster sensitivity and reliability, and thorough testing to avoid badly electroformed feeds or defects in dielectric materials.

There was general agreement that as ALMA progressed, there should be a person-polarization czar-assigned to keep track of the complex interplay of engineering and scientific specifications, calibration strategies and other polarization-related items. Lamb noted that he is working on a squint table, available within a few weeks. One vexing issue is the beam squint between L and R polarizations induced by offset mirrors. He noted that there is a tacit assumption that there are two orthogonal circular polarizations with different pointings, but polarization variations across the antenna aperture induce variable non-circular polarization, difficult to measure but important nonetheless. Tan noted that he would discuss polarization with Hamaker shortly.

The teleconference was thought to be useful. Continuing technical discussions between imaging and calibration, receiver and antenna groups will be scheduled.