How CCD Cameras Work
Taking the Image
A CCD (charge-coupled device) is an electronic instrument for detecting light. In the case of an astronomical CCD camera, this light is very dim. We will see that this has certain implications for how the CCD operates.
A CCD uses a thin silicon wafer chip. The chip is divided into thousands or millions of tiny light sensitive squares (or sometimes rectangles) called photosites. Each photosite corresponds to an individual pixel in the final image and photosites are often referred to simply as pixels. For clarity in this discussion, "photosites" will refer to the CCD chip and "pixels" will refer to an image. Each photosite is surrounded by a non-conductive boundary which contains the charge that is collected during an exposure within the photosite.
Above: A CCD chip consists of an array of photosites (squares) and a serial register for reading out the image data.
So, where does this charge collected during an exposure come from? The photoelectric effect.
In 1921, Albert Einstein received the Nobel Prize in physics, not for his famous theory of relativity, but rather for explaining the photoelectric effect. It is this effect operating in a CCD chip which produces the electric charge stored in the photosites. When a photon of light strikes the surface of certain metal materials (like the silicon in a CCD chip) the energy imparted by the photon can release an electron from the metal. In a CCD, this electron is stored within the walls of a photosite. During a long exposure, photons rain down from the celestial object being imaged, are collected and focused by the telescope, and strike the CCD detector. The photosites act like wells and begin to fill up with electrons (generated by the photons impacting the chip).
Above: The photosites (squares on top) collect light and convert it to electric charge. The "wells" begin to fill with charge. Brighter areas (light grey) fill faster than dark areas (dark grey and black).
If an area of the CCD is imaging a bright object such as a star (which gives off lots of photons), the photosites in that area fill up with more electrons than those in an area imaging something dim like faint nebulosity or the black night sky. (We will see shortly that even the photosites imaging black sky will end up containing some electrons for several reasons.)
Once the exposure is finished (usually done by closing a shutter on the camera), the charge must be transferred out of the CCD and displayed on a computer monitor.
Getting the Data Out
Photosites are read out, one row at a time, into a serial register along the bottom of the CCD array. The serial register then transfers the charge from each photosite to an output node. From there the charge is sent to the camera's electronics unit where the electrons are digitized.
Above: The first row of photosites is shifted into the serial register and read out.
Above: Each row is shifted down to the next row of photosites. The bottom row is transferred to the serial register and sent to the computer.
Above: The process continues until all the photosites have been read out.
A numerical value is assigned to each photosite's charge, based on the number of electrons contained in the photosite. This value is sent to the computer and the process repeats until each photosite's electrons have been converted to a pixel value and are displayed as a raw image on the computer screen.
Above: An uncalibrated image of M51, the Whirlpool Galaxy. The specks are due to "dark current", explained below.
This image is "raw data" because it is unprocessed by any software and there are artifacts present due to the nature of the CCD. Image processing software allows the removal of most of these artifacts. Below is a processed image.
Above: A calibrated and enhanced version of the above M51 image.
Since astronomical CCDs are designed to image faint objects they must be extremely sensitive. One drawback to this sensitivity is that the photosites also pick up electrons generated by heat within the camera. To minimize this effect, astronomical CCD cameras are equipped with cooling systems. These systems are capable of lowering the camera's operating temperature 20 to 60°C (about 40 to 100°F) below ambient temperature. Even this is not enough to remove all the effects of this dark current. For this reason CCD imagers often take a dark frame. A dark frame is an image taken with the camera covered so no light gets in. This image detects only the heat-generated electronic noise. The image below shows a dark frame of the same exposure duration as the raw image above.
Image processing software can later be used to subtract the dark frame (and hence the electronic noise) from the image. Some newer CCD chips have much lower inherent noise and may not require a dark frame, but in general dark noise is still a problem with many CCD cameras.
An alternate technology to CCD is CMOS (complimentary metal-oxide-semiconductor). Both types of sensor accomplish the same fundamental task. CMOS sensors have become much more common in non-astronomical applications like DSLR cameras and smartphone cameras. They consume less power and can cost less to manufacture in large quantities.
One fundamental difference is how the charge in the photosites is converted to a digital signal to be output to a computer. In a CCD, as we saw above, each photosite is read out and sent to a single output node to be converted from analog to digital. In a CMOS sensor, each photosite has its own charge-to-voltage conversion, allowing for faster speed. This is important in video cameras as might be used for planetary imaging. Because CMOS sensors are becoming so much more common in markets like smartphone use that dwarf the astronomical camera market, they are likely the future of digital imaging in astronomy as well.