It’s been a busy week and there’s a lot to write about, but this time I’ll focus on all the calibrations that have been taking place. Our camera is sensitive to both the temperature of the microwave sky and its polarization. (Polarization is sort of the orientation of light - it’s actually defined as the direction of its electric field). Since polarization is orientation-dependent we need to know to high precision the orientation of each pixel in the camera. If we get this “detector polarization angle” wrong then we’ll get the wrong answer for the polarization we measure in the CMB - that’s bad. And since we’re trying to measure very tiny polarization signals getting the wrong answer, even at a very tiny level, could trick us into thinking we see something in the sky that isn’t there. Really bad.
To measure the polarization angle of our detectors we made a “PolCal source.” It’s really just a hot light source behind a grid made of many parallel wires. The grid defines the polarization of the light coming from the source. At first the light has no net polarization: there’s about as much light oriented up-down as there is left-right. The grid acts to block light oriented in a certain way. Light parallel to the wires in the grid can excite electrons to oscillate in the same direction as the light (up and down the wire) and therefore that light gets absorbed. Electrons aren’t free to move between wires, so light perpendicular to the wires passes through as it can’t excite the electrons. The result is light with a well known polarization (perpendicular to the wires in the grid), which is exactly what we need to test our detectors.
The PolCal source is 3 km away from the telescope. A hot source with a wire grid is behind a giant reflector that bounces power up to the sky.
The station (center) and the telescope (left of center) from the PolCal source.
Our detectors are only sensitive to one light direction - essentially either up-down or left-right. If the light going through the wire grid is lined up in the same direction as a detector then the detector will pick up the light. If the light is perpendicular to the direction of the detector then the detector can’t see the light. You can imagine, then, that if we rotate the wire grid in front of our source we’ll change the intensity of the light the detectors will pick up - maximum intensity when the polarization of the light is lined up with the detector and minimum intensity when the light is 90 degrees out of phase with the detector. By measuring the strength of light the detectors see as a function of the PolCal source’s wire grid angle we can then fit a simple model to obtain the detectors’ actual polarization angles. Pretty nifty.
Here’s a plot of the same type of measurement Jay and I took in the lab before heading to the Pole. As the angle of our wire grid changes, the strength of the light seen by our detectors is modulated in a roughly sinusoidal pattern. The red and blue lines are for two different detectors in the same pixel. By design the angles should be 90 degrees apart and that’s just about what they are in this example. That’s great, but not every detector plays as nicely as this one, which is why we have to measure all the angles individually.
PolCal measurements we took in the lab. The strength of the light absorbed depends on the angle of the wire grid in the PolCal source.
There’s been a PolCal team down at Pole for the past three weeks or so making these measurements for as many of the ~1500 detectors in our camera as they can get. It’s painstaking and difficult work but they did a great job! It’s a lot harder to do this down at the Pole than it is in our lab in Boulder. For one, the PolCal source is 3 km away from the telescope and barely off the ground. That means the detectors are looking through a lot of atmosphere, which dumps a lot of power on them - almost so much that they completely lose sensitivity. That makes it difficult to understand the signals we’re getting from each detector. Another complication is pointing - the source is really tiny (so that the power of the light coming from it doesn’t totally blow our detectors out of the water). A tiny source means you have to be pointing at that source very accurately. Move off the source just a little bit and the signal strength drops really fast, which can look like a rotation in polarization angle, as in the plot above. REALLY confusing. And guess what makes the pointing difficult? The Sun (among other things). The Sun will heat one side of the telescope more than the other, which bends and flexes the whole structure changing where the telescope is actually pointing compared to where we WANT to point it. The PolCal team had to come up with ways around all of these problems and more, but they did and we have excellent data! Awesome job Ryan, Amy, JT, Tijmen, and Nicholas!
The PolCal team adjusting the source for more observations.
A map of the PolCal source as seen by our detectors. The light blue strip on the bottom left is the road out to the source. In this map dark blue is hotter and red is cooler. Low on the horizon you're looking through a lot of atmosphere, so it's warmer. The giant reflector in front of the hot source reflects cooler atmosphere from above down into the optical path of the detectors, which is why the reflector is red.
Before the end of PolCal observations I got to go out to the source. The telescope itself is a kilometer from the station, and the source is a further 3 km. The station and telescope are little specs on the horizon that far away - makes you really glad for all the cold weather gear as well as the snowmobile we use to get out there because it’s an hour walk one way. Yuck. But the really great thing about being that far away from station is that there is absolutely nothing around you. No buildings, no marker flags. Just a big, white, empty horizon. Breathtaking.
An endless blank horizon behind the PolCal source.