The needs of astronomers drive many of the advances in photon detectors because astronomers are looking for tiny amounts of light that have traveled great distances. Consider, for example, the infrared detectors on the James Webb Space Telescope that the National Aeronautics & Space Administration is currently developing. The telescope, which is scheduled for launch in 2013, is expected to capture a total of 4 × 10¹⁶ IR photons over five years, said James W. Beletic, director of astronomy and civil space at Teledyne Imaging Sensors, in Camarillo, Calif. Although that is a large number of photons, he noted that their cumulative energy will be only as much as that generated by a peanut M&M falling 6 inches. No wonder astronomers need such sensitive detectors.

Earthbound scientific applications could also benefit from highly sensitive optical or ion array detectors, some of which are already making their way into mainstream instruments, according to speakers at the symposium.

In an optical array detector, Beletic explained, a grid of detector elements (pixels) captures light. Energy from the captured photons shakes loose electrons trapped in crystal lattices of the pixels' light-sensitive material, thereby converting the light to electrons. The resulting signal is then amplified.

The most familiar optical array detector is the charge-coupled device (CCD), already found in many scientific instruments and in digital cameras. Individual elements in a CCD array store the electrical charge generated by incident photons. That charge is transferred through the device, read out one row at a time, and then digitized. The need for serial readout of the entire device limits the data acquisition speed.

The materials that make up the detector determine its effective wavelength regions. The most common detector material is silicon, whose optical properties make it suitable for the visible through X-ray regions of the spectrum. For other wavelengths, such as IR, designers must turn to different photoactive materials, such as HgCdTe. Varying the proportions of Hg and Cd in HgCdTe detectors tunes the effective spectral region.

CCDS HAVE excellent light-capturing abilities, but the way in which the signal is amplified and read slows them down. Manufacturers of custom detectors are now turning instead to faster, low-noise readout circuits based on complementary metal oxide semiconductors (CMOS), which are silicon-based and made in the same foundries as computer chips. The readout circuitry for CMOS detectors is built into each pixel, and these pixels, unlike those in CCDs, can be randomly addressed, enabling fast readout speeds. Low-performance CMOS detectors are already used in cell-phone cameras. Manufacturers of advanced detectors are making CMOS detectors suitable for scientific applications by improving their performance characteristics, such as noise, speed, and dynamic range.

Some companies are turning to hybrid CMOS devices, which combine a light-capture layer with CMOS signal readout and processing. The two layers are fabricated in separate wafers that are then connected to each other by indium "bumps."

The advantage of the hybrid CMOS technology relative to CCDs is that developers can use materials other than silicon for light detection, thus extending the wavelength range, and still retain the advantages of silicon for signal readout. "The hybrid, in theory, should provide the highest performance possible," said Gene Atlas, president and chief executive officer of ImagerLabs, in Monrovia, Calif.

ImagerLabs continues to work on monolithic CCD and CMOS detectors as well as on hybrids. "Right now there is no clear winner," Atlas said. "We will continue working on all three technologies until clear winners emerge."

Sarnoff Corp., in Princeton, N.J., is focusing on monolithic CMOS technology, in which light capture and signal processing occur in a single silicon wafer, said John R. Tower, technical director of the firm's imaging business unit. The main advantages of monoliths over hybrids are lower complexity and cost, Tower told C&EN. "Two pieces of silicon with interconnects are more expensive than one piece of silicon," he said.

Sarnoff is developing "backside illuminated" CMOS detectors for the scientific market. Backside illumination, which has long been available for CCDs but is only now becoming commercially available for CMOS detectors, improves device sensitivity. Sarnoff is extending backside-illuminated CMOS to new wavelength regions, including the ultraviolet and X-ray regions. They are developing UV CMOS sensors for semiconductor inspection in computer chip manufacturing and X-ray sensors for astronomy imaging.

Thermo Fisher Scientific

Take Charge In this series of charge-injection devices, each device is 30.9 mm wide by 11.4 mm tall.

Teledyne Imaging Sensors develops and produces CMOS-based sensors that use either silicon or HgCdTe layers for detection of X-ray, UV, visible, and IR wavelengths. Teledyne has developed a new type of "substrate removed" HgCdTe detector that can detect visible and IR light. The device is already being used in NASA's Moon Mineralogy Mapper, an instrument that is measuring the moon's surface with 10-nm spectral resolution from UV through near-IR.

Another type of array detector is known as a charge-injection device (CID), which is also made of silicon. The main difference between a CID and a CCD is that a CID can be read without transferring the photon-generated charge out of the pixel, making the readout faster and extending the dynamic range.

Michael J. Pilon, general manager of CIDTEC brand cameras and imagers at Thermo Fisher Scientific in Liverpool, N.Y., described a linear CID detector for handheld spark optical emission spectroscopy (OES). Traditionally, OES involves an individually positioned photomultiplier tube for each chemical element on a very large focal plane. This setup makes for an extremely sensitive technique but also for a large instrument—weighing more than half a ton, Pilon said. Linear CID, with its smaller optical system, will allow OES to fit in a much smaller handheld package.

An array detector suitable for handheld portable OES must be capable of specialized readout modes that allow time-resolved measurements, Pilon said. These specialized modes enable analysts to optimize sensitivity for each element. They also allow speciation of some elemental constituents, such as metallic aluminum and aluminum oxide in steel.

Thermo Fisher Scientific's linear array detector consists of 4,160 randomly addressable pixels arranged in two staggered rows. The detector incorporates components called row storage register capacitors, which store pixel signals as a function of time.

The spectral resolution of the handheld system will be less than that of OES systems based on photomultiplier tubes, Pilon said, but will nevertheless be "exceptional for a handheld device." In the end, the company believes the performance of the handheld instrument could be even better than that of the half-ton lab systems, Pilon said.

Researchers are also trying to improve the ion detection capabilities of array detectors. "Ion detection is an extremely important area of chemical analysis, and we're still using 1930s electron multiplier technology or 1913 faraday cup technology" to carry it out, symposium organizer M. Bonner Denton, a chemistry professor at the university of arizona, told C&EN. "We haven't seen a fundamental change in ion detector technology in all these years."

To make new ion detectors, Denton is adapting array technologies originally developed for photon detectors. His detectors include arrays of individually addressable fingerlike Faraday electrodes incorporated into a CMOS integrated circuit. These arrays are the first to have the close pixel spacing and large pixel count necessary for capturing the entire focal plane of a mass spectrometer, he claims. His newest ion detector consists of a single row of 1,696 pixels.

For ion-mobility spectroscopy, Denton is developing a single-channel detector that he said improves the technique's sensitivity several-thousand-fold over that of conventional ion-mobility instruments. When it's optimized for specific compounds, the system can detect them from more than a football field away.

Denton's next goal is to use the device to detect explosives from far distances. Even though his system is not yet optimized for explosives, he can detect TNT vapors from a distance of more than 135 feet. He has detected 100 mg of the explosive RDX from 5 feet at 25 ºC and from 70 feet when it was heated to 50 ºC.

The detector does suffer from what Denton calls the "deer-versus-hunter scenario," meaning that it needs to be downwind of an explosives source. "It requires deploying the technology such that you can force airflow from whatever you want to interrogate toward the instrument," he said. Denton suggested that such a detector could be used at vehicle checkpoints and airports and on subways, buses, and escalators.

Detector developers continue to improve the performance characteristics of scientific array detectors. The various detectors described during the symposium are not yet commercially available, but one day in the not-so-distant future they may be discussed on the expo floor rather than in the lecture hall.

Chemical & Engineering News

ISSN 0009-2347

Copyright © 2009 American Chemical Society