Charge coupled devices, or CCDs, are sensitive detectors of photons that can be used in telescopes instead of film or photographic plates to produce images. CCDs were invented in the late 1960s and are now used in digital cameras, photocopiers and many other devices. Its inventors, Willard Boyle and George E. Smith received the Nobel Prize in physics in 2009 for their work.
A CCD is a tiny microchip onto which the light that the telescope collects is focused. The microchip consists of a large grid of individual light sensing elements called pixels. There are 2048 pixels along each side of the chip in the Merope Camera in Faulkes Telescope North. Each pixel is a 13.5 micrometers(µm) square printed on a cracker sized piece of silicon 50µm thick. Tissue paper is about the same thickness. The images below are of astronomical CCDs from one of LCOGT's telescopes and shows the front and back of a CCD.
When light falls onto one of the pixels, electrons are released from atoms in the pixel. To measure the amount of light that fell onto each pixel, the number of electrons that was released has to be counted. This is done by measuring the charge on the pixel at the end of the last row in the grid. Then that charge is discarded and all the other charges in the row are made to move along to that one corner pixel. The next charge in line is then measured, and so on – until all the charges in that row have been dealt with. Then all the charges in all the remaining rows are made to move over one row, and the whole process is repeated. Amazingly, the entire chip can be "read" in less than 10 seconds. It is this method of read out that distinguishes CCDs from other devices (such as photodiodes and CMOS devices) that convert photons to electrons.
CCDs are increadibly powerful tools for astronomers because when a telescope's motion is synchronized with the Earth's rotation, the camera can “stare” at one spot in space for hours at a time. The longer the CCD is exposed to the sky, the more photons will land on it, and fainter, more distant objects can be imaged than are otherwise visible. CCD exposures are so long in astronomy (seconds, minutes or even longer) compared to digital cameras (normally a fraction of a second), that CCDs in telescopes are usually kept very cold (−50° to -100°C). Keeping the CCD at a very low temperature minimizes the effects of thermal noise. At any given temperature, a certain fraction of the electrons in the atoms of the CCD itself will will have enough thermal energy to liberate themselves. They are then indistinguishable from electrons liberated by the interaction of the CCD with incoming photons from the telescope, so they get counted as if they were light from a star.
Astronomical CCDs are similar to the sensors in your digital camera in that they both use the same underlying physics to detect light, Einstein's photoelectric effect, but that's where the similarities end. The CMOS device that's likely in your digital camera is really an array of millions of independent light sensitive photodiodes called pixels bounded by structure etched into the silicon itself. Each pixel is in turn connected to 3-7 transistors which together make up an electronic structure called a source follower, buffer, or simply an amplifier. In an astronomical CCD the boundaries of individual pixels are, in a sense, defined electronically, so that the charge created by the photoelectric effect and stored in each pixel can be moved about the sensor to a single larger and much more precise source follower. The end result is that the astronomical CCDs are more flexible and have less noise than your digital camera. The downside is that CMOS devices in digital cameras are much faster than CCDs. The other important difference is that astronomical CCDs are typically used upside down. All of the electronic structure etched on to a CCD or CMOS device is on a single surface of the silicon wafer. If you think of this structure as wires, and bridges, and walls, its easy to see that the more of it you have, the less room there is to collect incoming light. Astronomical CCDs are illuminated from the back where there is no structure, and through a process called depletion, electric fields are created within the silicon which rapidly move the charge created by the photoelectric effect on the back into the "buckets" on the front. To do this efficiently, the silicon needs to be very thin, anywhere from the thickness of an index card to the thinness of the thinnest tissue paper, so most astronomical CCDs are termed "back thinned." The end result is that astronomical CCDs can often detect 90-95% of the incoming light compared to typcially 40-60% for CMOS sensors in digital cameras.