Many high-performance scientific-grade CCD cameras are now designed to meet the needs of specific applications.
Gary McAnally
In applications ranging from biological imaging to Raman spectroscopy, the scientific-grade, high-performance charge-coupled-device (CCD) camera offers advantages for low-light detection. In response to market needs, manufacturers are developing an ever-widening range of CCD sensors and integrated cameras, each designed to provide the most economical solution for a specific application (see Fig. 1). Three application-specific cameras discussed below illustrate this trend.
Microarray scanning
Microarray scanners are making possible new applications in biology and biomedicine, from genetic research to the development of new pharmaceuticals. One such instrument is produced by Applied Precision (Issaquah, WA). This device provides an instant global picture of how a cell is functioning and how it is affected by an external agent, such as a new drug candidate, according to Carl Brown, a senior scientist at Applied Precision.
FIGURE 1. Charge-coupled-device sensors must often be customized for optimum use in demanding applications. Such sensors are now available in a wide array of formats, pixel sizes, dynamic ranges, and readout rates.
The most common type of microarray is a glass microscope slide covered with thousands of closely spaced spots of single-stranded DNA. Each spot contains many copies of a given DNA segment; most commonly each spot is loaded with the DNA from a different single gene.
A typical experiment involves two types of cellscontrol cells and cells that have been exposed to some agent or chemical. The cells are killed and ruptured to release their contents. The extent to which each gene in the cell is being expressed at that time can be estimated by assaying the amount of free messenger-RNA (m-RNA) associated with that gene. This m-RNA is converted into fluorescently labeled DNA using a combination of fluorescently labeled bases and an enzyme called reverse transcriptase. The control and experiment are labeled with different fluorescent markers, and the resultant fluorescent mixtures are each applied to the surface of a microarray. The ratio of fluorescence intensity at the two characteristic wavelengths quantifies the extent to which the gene associated with a given spot has been turned on or off.
The microarray scanner is a fully automated instrument designed to read out these arrays. According to Brown, early array readers used a focused laser beam to sequentially excite fluorescence from each spot, with the light being detected by a photomultiplier tube (PMT). This serial processing was time-consuming and could yield artifacts due to fluctuations in laser amplitude from spot to spot. Wide-field microscope optics and a CCD allow the user to sample many spots simultaneously. Furthermore, the quantum efficiency (QE) and noise floor of the CCD are both much better than for a PMT detector. The broad spectral response of the CCD and the use of metal-halide white-light illumination with a filter wheel allow sophisticated experiments involving a variety of different fluorophores.
The scanner's CCD camera is designed specifically for the instrument. A large-area detector is needed to maximize the number of spots imaged in each acquisition; another requirement was a 14-bit dynamic range. The mechanics of the thermoelectrically (TE) cooled camera are specific to the application, with cooling to -15°C required to increase dynamic range. Because the light is gated elsewhere in the system, no camera shutter is needed. More than one-quarter of the entire array can be imaged in a single data acquisition (see Fig. 2).
Amateur astronomy
FIGURE 2. Using a CCD-based array reader, biologists and geneticists can simultaneously survey the activity of thousands of individual genes. In this false-color image, the red-green color ratio indicates which genes have been turned on or off relative to a yellow normal control.
One of the first applications of CCDs was wide-field imaging with large telescopes. For many years, however, their high cost precluded CCDs from being used by amateur astronomers around the world.
Affordable CCD cameras have been available for amateur astronomy for about 10 years. But until very recently, these were low-performance cameras with limited dynamic range and low-grade sensors, according to amateur astronomer Tim Puckett (Atlanta, GA). "The inherently low QE of these front-illuminated chips was a significant limitation, extending the image-acquisition time necessary for capturing faint objects," he says. Minimizing data-acquisition time enables a single telescope to image more sky in a given night and avoids blur due to errors in tracking. Amateur astronomers want high QE but in a low-cost format; most amateur astronomers cannot pay $30,000 for a high-end scientific camera, explains Puckett.
FIGURE 3. Deep-ultraviolet imaging of the Sun's chromosphere reveals it to be highly structured and dynamic, containing knots and loop-like structures that change with time.
The highest QE is obtained with thinned, back-illuminated CCDs. With such devices, the light directly enters through the rear of the silicon substrate rather than through the gate electrodes that cover the front side, raising QE to more than 80% for 550-800-nm wavelengths. However, manufacture of thinned CCDs is much more costly than their front-illuminated counterparts.
To address the needs of the mid-range astronomy market, a CCD camera made by Apogee Instruments has been optimized for low-cost manufacturing. Extensive use of surface-mount electronics and custom gate arrays minimizes the number and the complexity of PC boards needed in the host computer. As a result, the entire system consists of just a TE-cooled camera head and an interface card. The efficiency of the electronics means that even with the shutter operating and the TE cooler running at maximum throughput, the interface card draws 25 W or less from the backplane of the host PC.
Puckett operates a 24-in. Ritchey-Chretien reflector with such a camera, as well as several other smaller telescopes; with the large telescope, he has imaged objects as faint as magnitude 20.5 in a 60-s exposure. In a typical night, Puckett uses his Ritchey-Chretien to capture images of hundreds of galaxies, then reimages these same galaxies one week later. Careful comparison of the images reveals the presence of any new supernovasof which Puckett has discovered three.
Deep-ultraviolet solar imaging
Some applications require a high degree of customization with less regard to cost. An example is a recent space-borne experimentthe Very Advanced Ultraviolet Telescope (VAULT)designed to capture deep-ultraviolet (UV) high-resolution images of the Sun at 122 nm. Its camera must withstand space flight as well as provide response well outside the usual silicon detector spectral range of 400-1100 nm.
The VAULT mission flew in May 1999, directed by a group of Naval Research Laboratory scientists, including Clarence Korendyke, Angelos Vourlidas, and Norman Moulton. "We wanted to obtain high-resolution images of the chromosphere, which is a very sensitive and complex layer of the solar atmosphere," explains Korendyke. "It is the zone through which heat, momentum, and mass are transported from the photosphere to the corona. Since the chromosphere is as hot as 20,000 K, one way to directly image this is to detect Lyman alpha hydrogen emission at 121.6 nm. But because of absorption by the Earth's atmosphere, it is impossible to record useful images at this wavelength using an earth-bound telescope." For this reason, VAULT was flown on a suborbital flight of a NASA sounding rocket that provided five minutes of viewing.
The weaker deep-UV emission is separated from the much stronger UV and visible emission using a pair of diffraction gratings and an interference filter. The principal requirements for the CCD camera were for a large-area chip to capture the widest possible field of view, small pixels for maximum resolution, low power consumption, ruggedization, and high sensitivity at 122 nm. The target dynamic range was 12 bits. With an intended exposure time of 2 s/frame and a satellite-to-earth download rate of 15 s/frame, a high readout rate was not required.
The chosen camera was based on a 2046 x 3096-pixel chip coated with lumogen, a phosphor that efficiently absorbs deep-UV light and re-emits in the green. Conversion efficiencies for this phosphor can be as high as 35%. Other modifications included customized packaging designed to endure the rigors of space flight and a modified board designed for ultraefficient cooling.
"This mission was highly successful, providing the first high-resolution deep-UV images of the chromosphere." Says Korendyke, "The camera recorded a succession of images of the same 396 x 256-arcsec field of view with a resolution better than 0.5 arcsec. At this high resolution, the filamentary nature of the chromosphere is quite clear, with both bright knots and extended loop-like structures (see Fig. 3). Our task now is to interpret this data in relation to observations of the corona and photosphere from other experiments."
GARY MCANALLY is a vice president at Apogee Instruments, 1350 N. Kolb Rd., Tucson, AZ 85715; e-mail: [email protected].