Electronic Detectors

Since their first astronomical use in 1976, electronic sensors have steadily gained ground as the detectors of choice in astronomy. In amateur astronomy, chargecoupled devices (CCDs) have been the dominant type. CCDs can detect light over a broad range of wavelengths, and they offer both high quantum efficiency and low noise. Furthermore, dark current and nonuniformities in sensitivity can be subtracted or divided out, thereby minimizing these shortcomings.

Challenging the CCD is another class of electronic sensor, the CMOS device. CMOS stands for complementary metal-oxide semiconductor, referring to the method of making them. While CMOS offers lower manufacturing costs, the resulting sensors are less sensitive and noisier than CCDs—a situation that will almost certainly change as the technology improves.

Today, electronic sensors incorporating CCDs and CMOS devices are used widely in amateur astronomy. Electronic cameras designed for astronomical work almost exclusively use CCDs, as do high-end digital cameras; but consumer-grade digital cameras and webcams increasingly rely on CMOS devices.

How CCDs Work. CCDs consist of an array of identical metal-oxide semiconductor capacitors formed on a silicon substrate. Each element in the array is called a photo-detector junction or photosite. The charge in each photosite is isolated from the others by a voltage applied through conductive channels on the surface of the silicon. At the beginning of an exposure, the capacitors are charged positively and then disconnected. As photons enter the silicon crystal lattice and are absorbed, they raise electrons from a low-energy valence-band state to a high-

Figure 1.8 Electronic detectors have revolutionized astronomy. They are compact, highly sensitive to light, and—unlike a photographic emulsion—can be used over and over again. Shown here is a Texas Instruments TC237, a 640x480-pixel CCD in a package just half-an-inch long.

energy conduction-band state, partially discharging the capacitors. The degree of discharge of each capacitor is proportional to the number of photons that strike it during the exposure. At the end of the exposure, the electrons remaining in the photosites are sequentially shifted (or "clocked") to a charge-sensing node, and amplified; then the signals are passed to external circuitry to be digitized and stored.

The detector characteristics of CCDs derive directly from their construction. In silicon, the energy gap between the valence band and the conduction band is 1.1 electron volts, so that only photons with energies higher than that can boost an electron into the conduction band and be detected. This energy limit corresponds to a wavelength of 1100 nanometers, in the near infrared part of the spectrum. At shorter wavelengths, however, silicon becomes progressively more reflective so that the photons never enter the crystalline lattice, and hence cannot be absorbed. Silicon CCDs, therefore, reach peak quantum efficiencies of 40% to 90% between 500 and 950 nanometers wavelength.

In CCDs, the number of electrons boosted into the conduction band is directly proportional to the incident flux of photons. CCDs are highly linear as long as the total charge that collects in a photosite is too small to leak past the charge barriers that separate each photosite from its neighbors. In practice, most CCDs are

Figure 1.9 CCDs consist of adjacent charge wells, each collecting photo-electrons during integration. To shift the electrons down to a charge detection circuit on the CCD chip, the charge on the electrodes is cycled so that accumulated photo-electrons move from one charge well to the next.

linear as long as the photosites retain at least half their original charge.

Photosites on a CCD are arranged in columns and lines. The size of a photosite ranges from 3 to 25 microns depending on the design of the detector. The lower limit is set by manufacturing difficulties, and the practical problem is that small photosites have small collection areas and therefore intercept few photons, and they can hold only a limited amount of charge. At the upper end, there is no need for photosites that are too large to capture all of the information present in images formed by camera lenses and telescopes.

CCD imagers range in size from 1.3 mm to approximately 70 mm across the diagonal, containing arrays of 102 x 102 to 8192 x 8192 photosites—and even bigger CCDs are being designed all the time. Unfortunately for amateur astronomers, the price of CCDs rises exponentially with the physical size of the chip. Typical CCDs for amateur use measure 4 to 24 mm across the diagonal, contain between 32,000 and 12,580,000 photosites, and cost between $60 and $10,000+. Although many CCDs can be read out 60 times per second, astronomical CCD cameras are seldom read faster than 50,000 photosites per second to allow precise measurement and digitization of the signal. Readout times vary from a fraction of a second to about 60 seconds to read a complete image.

The electronics industry makes CCDs in a variety of configurations, from robust camcorder CCDs designed for readout at 60 fields per second, to digital camera CCDs designed to produce sharp megapixel images, to science-grade sensors optimized for high quantum efficiency and low noise. The configurations encoun-

Progressive-Scan CCD M Masked Photosite

Figure 1.10 CCDs used in astronomy are almost always one of three types: interline-transfer, frame-transfer, and progressive-scan, shown schematically above. In all three types, charge is shifted line-by-line from the array to a serial register, and then shifted to a charge-sensitive amplifier.

Progressive-Scan CCD M Masked Photosite

Figure 1.10 CCDs used in astronomy are almost always one of three types: interline-transfer, frame-transfer, and progressive-scan, shown schematically above. In all three types, charge is shifted line-by-line from the array to a serial register, and then shifted to a charge-sensitive amplifier.

tered most often in amateur astronomy are the progressive-scan, frame-transfer, and interline-transfer. In addition, CCDs can be either front-illuminated (as most are) or back illuminated (as are high-performance scientific CCDs).

Progressive-scan CCDs contain a rectangular array of photosites with a special row of high-capacity photosites called the serial register at the bottom line in the array. After exposure, the electrons in all of the columns in the array are "clocked" down one line, and the bottom line enters the serial register. The serial register is then clocked one element at a time into the charge detection node. When the serial register has been emptied, all columns are clocked again to refill the serial register. The sequence is repeated until all charges on the array have been read and digitized. Progressive-scan CCD cameras must have a shutter to prevent light from generating new photo-electrons during the readout phase.

Frame-transfer CCDs have the same basic architecture, except that the lower half of the array is covered by an opaque mask. After the exposure, all of the columns are clocked rapidly to move the accumulated charge in the top half of the array to its bottom half. The bottom of the array is then clocked and read out just like a progressive-scan CCD. Frame-transfer CCDs do away with the need for a mechanical shutter, since charge in the upper half of the array can be transferred to the lower half in about 1 millisecond.

Interline-transfer CCDs are designed for even faster shuttering than frametransfer ones. In this configuration, alternate columns on the CCD are masked, and the electrodes overlying the CCD are arranged so that the charge in the uncovered

Figure 1.11 In this highly enlarged view of one corner of a CMOS sensor, you can see the regular array of individual pixels. In a CCD, accumulated charge from the whole chip is transferred to a single amplifier; in a CMOS device, each pixel has its own individually addressable amplifier.

columns can be clocked into the covered columns in a microsecond. The covered columns are then read out slowly. The primary difficulty with interline transfer is that half the detection area of the CCD is covered with masked columns, but a new generation of interline CCDs is being made with tiny integral lenses that redirect light from the masked columns to the active sensing columns.

The vast bulk of CCDs are front-illuminated models, meaning that the photosites are formed on a "thick" silicon wafer, and the electrically conductive gates necessary for charging and clocking charge are laid on top. This construction means that to reach the light-sensitive bulk silicon, photons must penetrate the gate structures, which are sometimes nearly opaque to short-wavelength (blue) light. Because of this, front-illuminated CCDs seldom exceed a peak quantum efficiency of 60% at the long-wavelength (red) end of the spectrum.

Back-illuminated CCDs are made the same way as front-illuminated ones, but the silicon wafer is etched to a thickness of around 10 microns, and the silicon mounted so that photons enter the array from the back side. Thinned back-illuminated CCDs are expensive, but their quantum efficiencies are high across the spectrum from 500 to 950 nm, with peak values approaching 90%.

One of the greatest advantages that CCDs enjoy is that the same detector is used again and again, and the output is highly repeatable. This means that the odd-

Figure 1.12 For color images, electronic sensors can be made with a checkerboard array of red, green, and blue filters (a Bayer Array). Each filter covers just one photosite, so the resulting raw image has a checkerboard of pixels made with different filters. The final color image must be reconstructed in software.

Figure 1.12 For color images, electronic sensors can be made with a checkerboard array of red, green, and blue filters (a Bayer Array). Each filter covers just one photosite, so the resulting raw image has a checkerboard of pixels made with different filters. The final color image must be reconstructed in software.

ball characteristics of a particular CCD can be calibrated out. An image from a CCD contains the following components:

• a bias voltage that is constant,

• systematic variations in the bias voltage,

• random variations in the bias voltage,

• a temperature-dependent dark current,

• systematic variation in the dark current,

• random variations in the dark current,

• a random variation (readout noise) in the output amplifier, and

• a signal generated by photons falling on the CCD.

In addition, photosites vary from one to the next in quantum efficiency, so the sensitivity of the array is not constant. However, because you can use the same CCD over and over, all of the constant and systematic effects can be removed. After calibration, a CCD image faithfully reproduces the amount of light that fell on each photosite in the array.

CMOS Devices. Like the CCD, the CMOS device consists of a large array of photosites on a silicon substrate. Unlike the CCD, however, the photosites in a CMOS device are individually addressable; that is, by activating a grid of conductors, any photosite can be read at any time. Instead of reading out the entire image, a small group of photosites can be read; or the CMOS sensor can be read out one line at a time while the remaining lines continue to accumulate signal.

Offsetting these advantages, however, each photosite must have an amplifier to sense charge and transistor switches to return the output signal. The auxiliary

Figure 1.13 In CCD cameras designed for astronomical imaging, the CCD is isolated in a sealed chamber behind a clear glass window. This allows the detector to be cooled, thereby reducing its dark current, while avoiding problems from water condensing and freezing on the detector.

electronics take space that on a CCD would be collecting light, and hard-to-control variations in amplifier characteristics make CMOS devices less uniform than CCDs. Nevertheless, CMOS devices have proven their worth in webcams and a growing number of digital cameras.

From a practical point of view, once an electronic sensor has passed an image to the computer, the image data from CCDs and CMOS devices look very nearly the same. To get the best possible results from either type of detector, it is necessary to make calibration frames and apply them to raw images.

Color Imaging with Electronic Sensors. Although they are intrinsically sensitive to light over a wide range of wavelengths, CCDs and CMOS devices are monochrome sensors; that is, they record total incident flux of photons with no color information. To obtain color, observers must use one of three methods:

1. Make separate exposures through color filters—usually red, green, and blue. Each of the filtered images records the photon flux in one band of wavelengths, or one color channel. Observers often back up a tricolor set of images with an unfiltered luminance image that records all three color channels. To construct a color image, the three separated color channels and luminance must be merged into a single image.

2. Make a single exposure using a CCD or CMOS device with an integral color filter matrix, usually called a Bayer array. The filter array is a tiny checkerboard of red, green, and blue filters, each large enough to cover just one photosite. Thus, one exposure records information for all three color channels, at the expense of reduced image sharpness. To construct a color image, the image data must be resampled to provide every pixel with all three color channels.

3. Make a single exposure using a special CMOS device that has three sensing layers. The top layer responds to blue, the middle layer to green, and the bottom layer to red. At the time of this writing, multi-layer sensors are a new technology and not yet available for use in astronomy.

Making separate filtered exposures is by far the most flexible technique because it allows the observer to select a set of filters suited to the imaging task at hand. A sensor with a Bayer array is, however, the easiest way to make color images because "generic" red, green, and blue filters are built into the detector.

Was this article helpful?

0 0
Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook

Post a comment