Kuiper Belt is a region of "asteroids" beyond the orbit of Neptune - somewhat similar to the more popular asteroid belt between Mars and Juipiter. These "asteroids" called Kuiper Belt Objects (KBOs) are very, very faint because they are so far away from the sun. So it is tough to see them through a telescope, and only the biggest (and hence the brightest) ones with sizes in hundreds of kilometers have been seen so far. We want to look for the smaller objects - roughly a kilometer or so in size. Knowing how many such objects exist, and where they are located, will help us better understand the formation of our solar system.
So you take a big telescope, a sensitive camera, and go look for these objects. Right?
Kilometer sized objects dont reflect enough sunlight to be visible at such large distances. (The visual magnitudes of these asteroids are fainter than about 24, that is roughly 15,000,000 times fainter than what our eyes can see.) Hence, we have to use other methods for observing the KBOs. One such method is to look for eclipses caused by these objects.
Eclipses? We've not heard of rocks eclipsing anything...
Well, these eclipses, called "occultations" are not as well known as the solar or lunar eclipses, of course. The eclipse here consists of a KBO passing between an observer and a distant star. The shadow of the KBO falls on a small part of the earth, covering an aera equal to the cross section of the KBO. Both the KBO and the earth are moving, so the shadow passes over any observer who is looking at the star. We know the speed at which earth moves, and we know the rough speed at which objects orbiting the sun at roughly the same distance as Pluto should move. Using these, and the duration of the eclipse, we can calculate what was the size of the KBO that caused the eclipse.
But thats just one eclipse...
Yes, it is. So we have to keep looking for more. Once we see a few eclipses, we can make statistical inferences about the KBO population. So, we take a model distribution (number of KBOs of each size) of KBOs, and calculate how many eclipses we should have seen if that model was true. We can compare that with the observations and see if the model is correct.
Ok, so when will these eclipses be seen? And from where?
That is the tough part - we dont know when and where we will see the eclipses, because we are looking for yet unknown KBOs. So we just have to select a background star which may get occulted by a KBO, and keep looking. We know that most of the KBOs should be in the plane of the solar system, so we can select a star in the plane. There is no preffered place on earth where we will see the occultations from. Current knowledge about the KBOs suggests that for my setup, we will see anywhere between one occultation per night, to one occultation per month.
Has anyone seen this before? Why or why not?
People are looking for these occultations. Some people have seen a few, others have not. But there isnt enough data available to determine the number of KBOs accurately. Why is this the case? Most astronomy instruments are designed to look at fainter and fainter objects. This means that we build larger telescopes, and use the equipment to take long exposures. Its common to have exposures that are several minutes or even hours in duration. But, a KBO occultation will last just about 0.2 seconds - that is much faster than the capacity of most current devices. The solution to this is to use custom hardware. Some people are using higher speed CCD cameras, but they quickly run into problems with readout noise etc. We are building a photometer just for observing these occultations.
Photometer?
A photometer is just something that measures the amount of incoming light. I am working on a "Dual Channel Photometer" which means I will have 2 devices measuring the light coming from two sources. The hardware is a very sensitive detector called "Avalanche Photo Diode" (APD). The instrument will consist of 2 APDs pointing at 2 stars in the same field of view. They will record intensity of the star every 10 milliseconds. Thus, if we are in the path of an occultation, we will have about 20 readings of the star brightness during the occultation. Having multiple readings will help to ensure that we have seen an occultation, and will give more information about the size of the KBO and the distance to it. A KBO will occult only one star at a time, whereas something on earth (like a bat flying in the telescope dome) will eclipse everything in the field of view. Hence the instrument will observe 2 stars simultaneously - a true occultation will cause a dip in the intensity of just one star, not both. When we think of observing stars, we naturally think of taking images - but this device will not produce pictures. It will just keep telling us the intensity of the star at a speed, and with high accuracy.
Okay, so you got the observations. How do you analyze them ?
That gets a bit more technical. The intensity data given by the instrument is called a "lightcurve". The lightcurve can be compared to theoretical lightcurves, and we can infer the size of the occulting object as well as the distance to it. We have to account for a veriety of things, like diffraction effects which produce a complex lightcurve rather than a simple "bright - dim - bright" lightcurve. If you want all those details, you should take a look at the reports and presentation linked from here. Of course you can always send me an