It was discovered that silicone release electrons when exposed to light. With a few pieces of electronics attached to a silicone chips, a CCD is made. As an aside, the CCD was invented by Bell Laboratories in 1970 - back when corporations believed in pure research.

A CCD, shown here on the right (image from SBIG), shows a 1024x1024 CCD. The 1024 is the numbers of pixels per row and per column. For a total pixel pount, use the formula for the area of a rectangle - or row x height = 1048576 pixels. This qualifies as a "mega-pixel" CCD chip.

A CCD is composed of smaller parts called pixels - or picture elements. The resolution of a CCD chip is the total number of pixels. They are organized in rows and columns. When the CCD receives light, an electric charge is collected by each pixel and stores them until the image is downloaded.

As seen in the image above, the charge of the pixels is read a row at a times. The image below shows an analogy of how the data from a CCD is read.

There are many definitions used to define the capabilities of a CCD. The most important of these is called Quantum Efficiency - or the chips ability to see light. One of the factors that make CCD's superior to film is its linear capability - that is twice the exposure means twice the light, unlike the reciprocity failure of film. Quantum efficiency is defined by a percent out of 100 and can be produced in graphical form:

Image Credit

This graph also demonstrates back-side illumination to front-side illumination. These are terms that define the orientation of the CCD chip. A CCD is coated with silicone and is usually thicker in the front than the underside (back side). This serves two purposes: protection and stability. It was discovered that flipping the chip over allows for more direct light on the pixel elements, and shaving additional silicone from the back allows for for light to reach the pixels, but this process is exact and results in many failed chips - hence the cost being in some cases twice that of the usual chip. But the graph says it all, they are more sensitive. Another type of CCD is called "anti-blooming." If a pixel stores too many electrons, they can spill over resulting in a grossly over-exposed image.

The above example shows what blooming does. An "anti-bloom" CCD contains electronics that limit this behavior, but limits the light grasp of the chip.

Other terms describing a CCD:

• ABG - Anti-Blooming Gate - reduces the blooming effects

• ADC - Analogue to Digital Converter - converts the electronic signal to binary information

• ADU - Analogue to Digital Unit - a single count from a CCD

• Bias - the background level of the CCD

• Bias Frame - a single image used to counter the Bias from the image

• CTE - Charge Transfer Efficiency - the electronics ability to transfer the electronic signal

• Dark Current - thermal noise due to heat build up on the chip

• Dark Frame - an image used to counter the effects of the Dark Current

• Flat Field - an image used to define the flaws of the CCD

• Full-Frame Device - the entire area of the CCD is exposed to light

• Frame Transfer Device - only half of the CCD is exposed to light

• Gain - number of electrons per ADU

• Lumigen - fluorescent coating used to improve UV response

• Overscan - read out more pixels than images, used for calibration

• QE - Quantum Efficiency

• Saturation - another name for blooming

• Shift Register - mechanism of moving a charge around the CCD

• Readout Noise - accuracy of the pixel can measure the charge

• Readout Register - location where the charge is measured

• Trapping Site - defect on the CCD preventing electron flow

• Well Depth - how many electrons a pixel can hold prior to saturation, or blooming

A couple of terms probably caught your attention: Dark Frame, Flat Field, and Bias Frame. These are images taken in addition to the normal image. The purpose of these is for additional calibration and reduction. Since the image from a CCD is digital - hence binary - any defect in the image can be cancelled out by using reduction techniques. By subtracting the Dark, Bias, and Flat frames from an image, a better image (and more accurate image) results.

For more information of image calibration, see my Advanced Topics.

Back to Top