Bending the Rules

I got an amazing lesson last week on just how complex it is. First, a small digression: we have all seen the effects of refraction. That is the fancy term for the bending of light as it goes from one medium, like air, through another, like water. The "bent spoon" illusion is a classic example: put a spoon in a glass of water and look at it from the side; it appears as if the spoon is bent. Actually, it is the light from the spoon that bends as it passed from the water to the air.

On board HST is a camera called STIS, or Space Telescope Imaging Spectrograph. This is the camera I work on. STIS has several filters that help it determine just how bright its targets are, as well as what is in them (like hydrogen and carbon). These filters are plates made of glass (or other materials) and coated with a substance that only lets light of a certain color range through to the detector. We know very accurately the specific wavelengths of the light these filters let through. One of the jobs I was hired to do was to calibrate these filters; that is, to check and make sure they are letting through the right amount and color of light. Exciting, right? Well, it's very important, and we have to be very careful when we make these measurements.

A short time ago we found a small error that kept creeping up in our measurements, and we couldn't trace it. It turns out the culprit was the refraction of light through the filter. Now, we knew that light bent going through the filter, but software onboard Hubble compensates for it. Our software on the ground was also compensating for it, which means we were overdoing it. It took us quite some time to trace this little bug and fix it.

So why am I telling you about this? Because of the size of this refraction: it moved the the measurements by about 0.05 arcseconds. An arcsecond is a measure of size on the sky; the Moon is about 1800 arcseconds across (a half degree or so). This means that we had an error that was about one part in 40,000 of the size of the Moon, and this was serious enough for several people to spend many hours tracking it down! An error that small would have big repercussions down the line if we hadn't fixed it. Every scientist's measurements using that filter would have been wrong! We did fix it though, and now everything is back to normal. At least, until we find the next tiny bug that will take some sleuthing to track down!