AO Polarimetry with IRCAL

Note: IRCAL Polarimetry mode is still undergoing development! This mode is not officially supported and you should almost certainly talk to me if you're contemplating using it. The following notes are in no way comprehensive.

Observing Procedure Summary

(This is just the overview; read on below for the full gory details.)

Getting Started

Waveplate Setup

First you need to mount the waveplate on IRCAL. When not in use, it's kept in the IRCAL/GEMINI lab room at Lick, on the shelf in a cardboard box labelled "IRCAL Waveplate". It mounts onto the front of IRCAL using four screws, which should already be in the screw holes. Just unscrew them, mount the waveplate, and screw back on. This can be done by one person but it's easier with two pairs of hands and somebody holding a flashlight. The two control cables plug in to labelled sockets on the AO bench to the left of IRCAL; both mechanisms are currently at their zero positions. Ellie knows how to do all this stuff if you need a hand.

Right now we're just using the waveplate rotation motor; there's no need to even cable up the waveplate insert motor as we tend to just leave it inserted all the time on the nights we're doing polarimetry. In any case, when taking observations you will need the waveplate rotation motor powered on, and the waveplate insert motor powered off (because it creates lots of electronic interference noise if left on...)

Warning: For some unexplained reason, installing the waveplate on IRCAL sometimes causes the IRCAL read noise level to increase dramatically, even when all motors are turned off. To fix this, you have to power cycle IRCAL. Thus after installing the waveplate, check the noise levels by subtracting two darks, and if necessary, power cycle the camera.


Don't forget about focus! The waveplate adds about 5000 counts of focus to IRCAL. You should either just leave the waveplate in place all night even for non-polarimetric obs (this is easiest but costs a few percent of the light at K, only 1% at J or H) or else you have to refocus IRCAL every time you move the WP in or out. This is a pain and is easy to get wrong at 4 am so be warned (A lot of the data from May was out of focus for this reason...). The best course by far is just to leave the waveplate in place all night long.

Flats and Darks

You need flats and darks of course. Darks are taken in the regular way at the end of the night. I have been taking two different types of flats for polarimetry observations. Make sure to tell the telescope operator ahead of time that you want the dome open early for twilight flats. First, take regular IRCAL sky flats, most easily done with the "getflats" script. You can actually take flats starting before the sun goes down if you use a short integration like 1s. These flats should be taken with the waveplate in place but not the half field mask or wollaston. Then insert the Wollaston and half field mask, and jog the half field mask to be positioned nicely (not falling off the edge of the detector). Once you have positioned the half field mask, do not under any circumstances move it again! The positioning is highly unrepeatable and a real pain. If somehow it does get moved, carefully try to get it back to the same position as closely as you possibly can. The polarized flat fielding depends a lot on your getting this right....

Now run the "polarim" script at Ks, H, and J to obtain polarized flat fields. Do Ks first, then H, then J because the sky gets dark much faster at K. When running polarim scripts on the sky, set NEXPOSE=1.. This is important so that the corresponding pairs of images with the waveplate rotated are taken as close in time as possible. This is important since the overall polarization level of the sky changes rapidly with time as the twilight sky brightness changes. Of course, get as much flat field data as you can before the sky gets too dark.

In addition to being useful for flattening the data, this second set of flats will serve as a position angle calibration, when compared with the known polarization of the sky relative to the sun.

Observing Techniques

Observing is easy. Waveplate in, Wollaston and ap mask in, fiddle with ap jog if necessary (Woll generally does not need to be jogged I don't think).

The "polarim script"

The "polarim" script takes the 8 polarimetric exposures. Set up the exp time and coadds and position the object appropriately, then just run polarim and sit back while it does all the work. You should dither around a few field positions and execute "polarim" at each. Ircaldisplay's subtract feature is very useful for looking at the data and seeing if there's anything there. just subtract two sequential polarized ims and you get a low-S/N Stokes image.

WARNING: Currently the waveplate rotator mechanism is misbehaving. Maybe a third or a half of the time, it fails to return to 0 after the script completes, because the motor has stalled around 1500-2000 counts (15-20 degrees). To fix this, on the AO control computer in the appropriate motor control panel, turn the power to this motor off then on again, then send it to zero.

Taking Skies

Make sure to take some sky frames! As polarized emission is often quite extended, using the median of the science frames as a "sky" does not always work well. You should take a polarim sequence or two out on clear sky 20-30 arcsec away. You don't necessarily need to repeat this for each and every object during the night. Under stable conditions, skies obtained near one object usually work fairly well for reducing other objects as well. A note of caution: Taking highly overexposed images of a bright star (such as the 30-60 s exposures which are often necessary to bring out faint circumstellar scattering) will temporarily burn afterimages of the star onto the detector due to charge trapping in pixel walls. This means if any skies taken right after a series of 30 s exposures are likely to have ghost afterimages of the star in them, faintly. The only way around this is to just wait a bit for the excess charge to dissipate. On the other hand, this isn't that big a deal in the end, since the afterimages are in the same location as saturated data which is already lost to you...

Photometric Standards

Observe these like normal - just make sure to do so with the waveplate and Wollaston in place! You can search for standards using
this web page at Gemini

NOTE: For Ks band, be sure to dither the standard around on the FOV so that one can use a median sky frame to remove the background! This step is much more important at K than it is for J and H.

Unpolarized Standards

For calibration you should also observe an unpolarized standard star. I've been using the list on this UKIRT web page but any nearby star should be unpolarized (i.e. no interstellar polarization). The UKIRT list is in 1950 coords and they're all high proper motion stars, so look up the new coords from SIMBAD or you can waste ten minutes trying to find a star in completely the wrong spot. The observation of the unpolarized standard star should be unsaturated - and can actually be done open loop since we're just doing aperture photometry on it.

There are also Serkowski polarized standards available on the same web page. I haven't found these to be all that useful for our purposes since most were measured with very large apertures (20" or larger) and it's often hard to compare the data to our spatially resolved polarimetry.

Data Reduction

...needs to be written up, but hasn't been yet.
This page last updated by Marshall on 2003-11-15