[mmaimcal] Re: Polarization Teleconference

[James Lamb]

The following are my personal views of the polarization issues related to
receivers and may not necessarily
reflect the JRDG or others involved in receivers. However, I will try to
bring up some of the points at the
JRDG meeting immediately preceeding the Im/Cal meeting.

There are enormous demands on the receiver group to satisfy a wide range of
capabilities, including the
ability to make good polarization measurements. Inevitably this leads to
compromises in the design,
and the receiver group has to try to make an optimum choice based on
judgements of what is
technically feasible and on input from other groups. It has to be remembered
that we have to design
for 20 independent receiver channels (not including the WVR), even though
not all channels will be
implemented at first light (!).

In the case of polarization there have been neither clear specifications nor
detailed justifications of some
of the requests. The most complete analysis which has been issued is
Cotton's Memo 208, 1998. Although
it was very detailed and extensive, the memo did not translate into any any
firm engineering specifications
and seemed to be written more as a discussion of how one could best use an
existing instrument.
For example, it is shown that phase variations between orthogonal linear
channels will corrupt the
measurement of linear polarization and this leads to a recommendation for
circular feeds. However,
there is no resulting number for the acceptable phase stability. The
specifications on the phase for
ALMA are already quite stringent, and the interpolarization phase on any
antenna is likely to be
an order of magnitude or more better than the inter-antenna phase stability.
Is this good enough at
350 GHz, say? There are other options that may be available, such as the
proposed laser generated tone.
Calibraton of the relative linear phases on an antenna may be done by
looking at an astronomical
calibrator through a grid (which will add noise, but only during
calibration). These need to be evaluated
in detail.

In trying to decide on the best attributes for polarization measurements
there are several measures
performance including:

1.    Linearity (or circularity) of polarization channels.

2.    Orthogonality of polarization channels on each antenna

3.    Alignment of polarizations between antennas.

4.    Stability of phases and amplitudes on long timescales.

5.    Stability of phases and amplitudes on short timescales.

6.    Variation of polarization across aperture.

It is clear that perfect linear or circular polarizaton will not be
achieved. It will also not be possible to
calculate the polarization state exactly a priori so that an accurate
calibraton scheme is mandatory.
Calibration would be simplified if all antennas/receivers are identical and
stable to the level of the
final accuracy requirement. Significant differences between
antennas/receivers can in principle be
calibrated out, but this could involve significant algorithm development and
computation (but eventually
this may be needed for the most demanding observations). The temporal
stability will affect the
calibraton interval (long timescales), or measurement error (short
timescales). I have no idea if
calibraton can reduce the polarization imperfections by a factor of 2 or 100
without any clear
caibration scheme.

On the basis of this, the receiver design has not been directed at attaining
some absolute polarization
properties (although a considerable effort has been put into getting uniform
polarization over the
aperture). Instead, much of the effort has been in trying to make the
receivers as identical as possible
and achieving the highest sensitivity and stability. These goals will
benefit all the science objectives, and
the competing factors are only cost and development time. There are some
design decisions that will
have distinctly beneficial effects on polarization. For example, there are
quite stringent tolerances
on alignment of optics. Also, great efforts have been made to have large
clearances for the
beams to maximize efficiency, and this reduces the strong
polarization-dependent diffraction effects
at apertures.

Other general design goals which will be helpful are to avoid moving parts
(repeatability and
reliabilty) and to minimize the number of optical components (sensitivity,
reliaility). Furthermore,
the approach will be to do thorough testing of components and sub
assemblies, which will
avoid such problems as polarization corruption in badly electroformed feeds
or defects
in dielectric mateials.

Many issues are still unclear (at least to me). As a case in point the
beam-squint between
L and R polarizations induced by offset mirrors has been a design driver. It
always
appears to be tacitly assumed that there are two orthogonal circular
polarizations with
different pointings, but since the effect arises from polarization
variations across the antenna aperture
the two beams will have variable non-circular polarization. This effect is
much harder to measure
than the beam-squint, but may be equally important.

There is clearly an enormous amount of work to be done in understanding the
best strategy
for millimeter and submillimeter polarimetry. It is not so obvious that
there can be any
significant change in the approach to the receiver design which would have
any substantial
benefits to polarization properties. It is nevertheless important that there
be continuing technical
discussions between the Im/Cal group and the receiver and antenna groups.
Any specific
recommendations can be considered and it is important to arrive at a
procedure for testing modules
and receivers during assembly, and to include any useful calibration devices
which may be necessary.
It will require significant dedication of someone's time to be more
quantitative and definitive than
has been the case so far. It is extremely important, for example, to keep
all receivers identical
to minimize the number of 'flavors' of spares.

James Lamb
OVRO
2000-01-04


----- Original Message -----
From: "Al Wootten" <awootten at NRAO.EDU>
To: <mmaimcal at polaris.cv.nrao.edu>; <bglenden at polaris.cv.nrao.edu>;
<pnapier at polaris.cv.nrao.edu>; <jpayne at polaris.cv.nrao.edu>;
<wild at astro.rug.nl>; <lamb at ovro.ovro.caltech.edu>;
<demerson at polaris.cv.nrao.edu>; <ldaddari at polaris.cv.nrao.edu>;
<guillote at iram.fr>
Cc: <rbrown at polaris.cv.nrao.edu>; <mrafal at polaris.cv.nrao.edu>;
<kmenten at mpifr-bonn.mpg.de>; <kawabe at nro.nac.ac.jp>; <rkurz at eso.org>;
<cwhite at polaris.cv.nrao.edu>; <crutcher at astro.uiuc.edu>
Sent: Tuesday, January 02, 2001 9:39 AM
Subject: Polarization Teleconference


> ALMA/US Imaging and Calibration Group
>
> Agenda for meeting Thursday, 4 Jan 2001 at 12:30pm EST.
>
> Date: 4 Jan 2001
>
> Time: 12:30 pm EST (1730 UT, 10:30 am Socorro, 10:30 am Tucson)
>
> Phone: TBA
>
> --------
>
> Special Polarization Meeting Thursday
>
> Crutcher has proposed a special meeting on polarization on Thursday.
> Dick's discussion guide:
>
> Polarization observations have been defined by the ALMA Science Advisory
> Committee as being one of the important science drivers for the ALMA
> telescope(ref: March Report). The polarization science given the highest
> importance was the mapping of linear polarization, both over ALMA
telescope
> primary beams and over multi-pointing mosaic fields. Hence, both
> interferometer and single-dish polarization must be addressed. The March
> 2000 ASAC report proposed the requirement for 0.1% polarization mapping
> fidelity after calibration. This requirement of 0.1% polarization fidelity
> after calibration does not lead directly to specifications on antennas,
> receivers, and software, since the combination of instrumental
polarization
> induced by receivers and antennas both play a role, along with the
> procedures for calibration. In order to insure that the polarization
> requirements will be met, it is essential that coordinated planning across
> these areas take place.
>
> The present ALMA Project Book does not address how the science
polarization
> requirements will be met. Indeed, things seem to be going the way they
> usually do with radio telescope design. One discovers that to make
> polarization work well will compromise other things, so polarization is
> given low priority and polarization science is compromised. ALMA examples
> include adopting linear rather than circular polarization feeds, having
all
> feeds oriented identically rather than half rotated at 45 degrees (which
> would allow use of software to derive optimally all 4 Stokes parameters
> simultaneously), and having all feeds off-axis (which will introduce
> significant primary beam instrumental polarization). All of these design
> considerations have sound justifications in order to optimize performance
> for Stokes I observations, but at the expense of polarization
observations.
> There does not appear to be any plan for insuring that the potential of
> ALMA for polarization observations will be met.
>
> The following design features should be reviewed to see whether it might
be
> possible to give some weight to polarization in the ALMA design.
>
> 1. Having the linear polarization response of half of the antennas at 45
> degrees to the rest would provide an established method for measurement of
> all four Stokes parameters. This arrangement would require measuring 4
> correlations, which would reduce the allowable maximum bandwidth to be
> correlated if only Stokes I is desired. Is this loss of proven
polarization
> capability worth the tradeoff of higher bandwidth? If so, could it be
> recovered by having half of the receiver units rotatable through 45
> degrees?
>
> 2. Having all receivers off-axis ignores the recommendation that one (at
> least) receiver be optimized for polarization observations. Off-axis
> receivers will produce instrumental polarization that is dependant on
> position within the primary beam, making it more difficult to map
> polarization over large fields of view. Different antennas, particularly
> those of different sizes, will have different instrumental polarization
> characteristics. Is it possible to have a prime polarization receiver
> channel (345 GHz) on axis?
>
> 3. Even if the ALMA design were most carefully optimized for polarization,
> there will a variable instrumental polarization response over the primary
> beams of the antennas. To have any realistic hope of being able to do
> polarization mapping, it is essential that the polarized beam patterns be
> stable over time, temperature, antenna pointing, etc. In order to
calibrate
> polarization science maps, the polarized beam of the antennas will have to
> be measured by mapping the polarized beam pattern at the 0.1% level on the
> spatial scale of the synthesized beam. This will be a time consuming
> operation - to be practical, it will have to be done once (or
infrequently)
> and applied over extended periods of time. This will require that the beam
> pattern be stable and repeatable. Is this a formal design consideration
for
> the antennas? How will this be tested before an antenna design is
accepted?
>
> 4. How will the polarization calibrations be done? Darrel Emerson
presented
> a design for a bandpass calibration system at the Berkeley ASAC meeting
> that could also be used for polarization calibration of the receiver. This
> would consist of an amplitude and phase stable signal that would be
> broadcast into the receiver, so one could essentially continuously
> calibrate the gain and phase difference between the two nominally
> orthogonal receptors on each antenna. If both receiver gain and phase can
> be very accurately calibrated, it should be possible to derive all four
> Stokes parameters from orthogonal linear feeds with common orientation,
> with acceptable loss of fidelity in the Stokes parameter derived from $X -
> Y$. Is this a firm component of the ALMA plan?
>
> 5. Provision for testing the polarization characteristics of the
> two-antenna test array should be made. This will give the opportunity to
> test the system hardware and software for interferometer and single-dish
> polarimetry and identify problems before additional contracts are signed.
> What is the plan for a full testing of the ALMA polarization system to
> insure that the science polarization goals will be met?
>
> -------------------
>
> Richard M. Crutcher
>
> Professor & Chair, Astronomy Department
>
> Chief Application Scientist, NCSA
>
> University of Illinois
>
> 1002 W. Green Street
>
> Urbana, IL 61801
>
> Voice: 217/333-9581
>
> Fax: 217/244-7638
>
> ------
>