Multicolor focal plane arrays provide critical image information.
Richard Gaughan, Contributing Editor
A missile launches from a remote silo, streaks upward through the clouds, then shoots above the atmosphere. There is no chance to stop the re-entry vehicle from hitting its target unless it can first be identified — picked out from the natural Earth background and from the clutter created by decoys it releases. This technically challenging task demands a technologically advanced solution: multicolor focal plane arrays (FPAs).
All physical objects emit infrared radiation. Because the intensity of the emitted IR depends only on properties of the object itself, rather than on ambient illumination, IR detection is a key component of many surveillance systems. Detection in a single band can alert a system to the presence of a warm object, but it won’t necessarily provide enough detail to identify the object because the radiant intensity depends upon the object’s temperature, emissivity and area. At a specific wavelength, a hot, low-emissivity target can have the same intensity as a cooler, high-emissivity target. How do you tell the difference?
Distinguishing hot from cool
If the IR emission from an object is measured at two, three or more wavelengths, all the relevant characteristics of the object can be identified. But with airborne, space-borne or even field-deployable systems, weight is at a premium. That’s why defense organizations such as the US Army’s Night Vision and Electronic Sensors Directorate (NVESD) Missile Defense Agency (MDA) have funded the development of multicolor focal plane arrays.
A multicolor array is a monolithic two-dimensional array of IR-sensitive pixels where each pixel is actually composed of two or more photosensitive elements built on top of or next to one another. Ideally, such arrays provide simultaneous, identical images of a scene, but at different wavelengths. According to Meimei Tidrow, the agency’s program manager for the Passive EO/IR Technology Program, “For ballistic missile midcourse applications, two-color arrays can measure the cooling rate of imaged objects to separate decoys and other closely spaced objects from a real target. Two-color detection could be a critical element in the plume to hard-body hand-over [the transition from tracking the hot plume to tracking the cooler missile body].”
Two multicolor focal plane array technologies are under development. First is mercury-cadmium-telluride (MCT), a material that has been used for IR detection since the late 1950s. Second is the quantum-well infrared photodetector (QWIP). Introduced only about 10 years ago, this relative newcomer is constructed from layers of III-V materials such as GaAs/AlGaAs. Tidrow said that MCT currently has better performance for single-color but that QWIP has much higher uniformity and a more direct road to low-cost production.
Quantum-well infrared photodetector arrays can be manufactured in many arbitrary configurations. This 640 × 512 array consists of four separate 640 × 128 regions, each sensitive to a different range of IR wavelengths. Courtesy of Jet Propulsion Laboratory.
Traditional material, new design
Jeff Johnson, the manager for dual-band programs at Rockwell Scientific Co. in Thousand Oaks, Calif., has been working with two-color MCT arrays since 1995 and is quick to dispel one of the myths surrounding this material. “There is a perception that MCT production is a ‘black art,’ ” he said. “It’s difficult to work with, but people who work exclusively with MCT have developed the expertise to exert a surprising amount of control over processing.” That control is necessary because several manufacturing steps are required to build two-color arrays.
MCT detects photons directly. When a photon’s energy matches or exceeds the bandgap of the material, it transfers energy to an electron in the valence band and bumps it into the conduction band. The bandgap is adjusted by changing the relative percentage of mercury and cadmium, which results in a peak absorption in the range of 1 to 12 μm. In a two-color device, each pixel can be constructed by depositing higher-bandgap MCT on top of a layer composed of smaller-bandgap material. Shorter-wavelength radiation is absorbed in the top region, while longer-wavelength light passes through the top to be absorbed in the lower material. The focal plane array is then mated with a readout integrated circuit that collects and formats the signal to interface with other circuit elements.
Rockwell Scientific, in work sponsored by NVESD, produces a dual-band 256 × 256 30-μm-pixel MCT focal plane array with quantum efficiency as high as 65 percent. “A big application for these devices is for the US Army,” Johnson said. “They would like to detect and identify enemy targets before being detected themselves.”
The long-wavelength band detects the presence of a warm object, while the shorter-wavelength pixels have high enough resolution to identify it. Because images at both wavelengths are simultaneously acquired, the long-wavelength imager can direct automated software routines to examine specific regions in the short-wavelength image. Customers are interested in very large format arrays with small pixels, so Rockwell is developing focal plane formats of 640 × 480 and 1280 × 720 arrays of 20-μm pixels.
Johnson sees the development of large-format, small-pixel arrays as a challenge but feels it is manageable through straightforward engineering practices. The more difficult challenge is inherent to the nature of the IR focal plane array business. “We are a low-volume industry,” he said. “We produce as many devices in a year as a silicon foundry might produce in an hour.” That requires the processing to be flexible, but high reproducibility from one production run to the next is critical to keep the product affordable.
The situation with QWIP production is, ironically, a near-mirror-image of that for MCT. QWIP focal plane arrays are typically constructed from AlGaAs/GaAs, a highly developed and mature manufacturing technology. NASA’s Jet Propulsion Laboratory in Pasadena, Calif., has been developing these arrays for several years under the guidance of Sarath Gunapala, supervisor of the Infrared Focal Planes & Photonics Technology Group. “Our unique design parameters for producing FPAs,” he said, “don’t require changes in the process parameters.”
This dual-band QWIP device senses both long- (left) and midwave infrared radiation. Two-color imaging allows precise determination of the temperature and emissivity of resolved objects. This device, developed by BAE Systems of Nashua, N.H., is used in advanced forward-looking infrared systems.
High-yield, high-uniformity, high-pixel operability and low cost are part and parcel of the GaAs manufacturing technology. But QWIP array development has its own challenges.
As evidenced by their name, QWIP focal plane arrays are constructed from quantum wells. A lower-potential GaAs region is sandwiched between higher-potential barriers of AlGaAs. The energy difference between successive accessible electron levels is a function of only the height and width of the well — both of which are easily adjusted in manufacturing. When the detector is biased, the absorption of photons of a specific energy will put the electrons in a mobile state, allowing them to be collected as a measured photocurrent. A single-color QWIP photodetector will typically consist of a stack of up to 50 identically designed quantum wells, with the total photocurrent being the sum of all the electrons released from their confinement.
A two-color device is produced by stacking one set of quantum wells on top of another set. Different design parameters result in two spatially colocated pixels sensitive to different wavelengths. Conductive layers on the base, on the top and between the two pixels allow the pixels to be biased and the focal plane array to be interfaced to a readout integrated circuit.
The Jet Propulsion Lab has developed a 640 × 512-pixel four-band device for earth science applications covering the regions from 3 to 5, 8.5 to 10, 10 to 12 and 13 to 15 μm. The first band is ideal for detecting forest fires; 8.5 to 12 μm are good wavelengths for detecting natural and industrial pollutants and mineral resource signatures; and the final band can produce a temperature profile of the atmosphere good to 1 °C.
QWIP technology can produce a nearly arbitrary range of both broad and narrow wavebands within a single device, and can do it all affordably. So what’s the problem? Two words: quantum efficiency, which is the probability that an incident photon will interact with an electron within the detector.
The electrons within the quantum wells are quantized in the vertical dimension, but the electric fields of the incident photons are horizontal. That means there is no overlap between the photon and the electron to make absorption possible. The solution is to manufacture coupling gratings on top of QWIP arrays that change the orientation of the incident electric field. In practice, QWIP arrays can reach quantum efficiencies of about 15 percent, which can be increased to 30 percent when operated at extremely cold temperatures.
Does low quantum efficiency make the arrays’ performance unacceptable? Not according to Gunapala. “In terrestrial applications or in a boost-phase missile detection system, where there is a lot of photon flux, QWIPs are very competitive,” he said. He also said he believes that QWIP technology holds advantages for low- background, long-wavelength applications, from 7 μm up to about 16 μm. When identifying challenges for the future, he brushed off the issues associated with making large-format, high-uniformity arrays, stating simply, “We need to increase QE.”
Tidrow of the MDA?is pleased with the progress in both of these multicolor focal plane array technologies. Although she would like to see better quantum efficiency from QWIP detectors, for boost phase, “when there’s lots of signal,” QWIP focal plane arrays are a viable option. MCT focal plane arrays perform well and have been improving their process control, but she’s still looking for better quality and uniformity.
Driving down costs
“They’re improving,” she said, “like art.” Two-color detectors provide an obvious weight reduction, which has a direct impact on the cost for air- and space-borne defense.
Gunapala agrees that cost is a significant driver. “For many commercial applications, and even NASA and [US Department of Energy] uses, cost is a critical parameter. If you want to price a two-color infrared camera in the $50k range, you need to get 640 × 512-pixel FPAs priced around $10k.” He has had people approach him to investigate the feasibility of multicolor focal plane arrays for not only defense and civil space applications, but also commercial and medical applications.
Rockwell Scientific’s Johnson pointed out that dual-color IR focal plane arrays can have significant ramifications for a full system design. “You’re taking what would have been two complex optical trains, requiring precise, stable alignment, and reducing the load and the complexity.”
In the evolution of detector technology, the next step is to provide multiple functions within a single detector, and “dual-color is the first step.” He summarized, “The more functionality on a single chip, the cheaper it is at a system level.”
Customers crave the depth of information provided by multicolor IR imaging; MCT and QWIP technologies are poised to make it realistic and affordable.
Missile Defense: Multicolor Infrared Takes on a Hot Job
In 1998, the Rumsfeld Commission Report was delivered to the US Congress. It stated that more countries have the means to develop ballistic missiles and that “the newer ballistic-missile-equipped nations’ capabilities will not match those of US systems for accuracy or reliability. However, they would be able to inflict major destruction on the US within about five years of a decision to acquire such a capability.” The report was one of the factors behind a renewed political interest in developing an antiballistic missile defense system.
These systems have two basic challenges: Missiles are launched thousands of miles from their targets, and they travel at several hundred miles per hour; therefore, any system that plans to defend against them will have to detect them from a great distance and acquire accurate and detailed information about potential targets very rapidly.
In theory, this might not seem too difficult; after all, there are not too many other objects to confuse the detection process. But how can the missiles be detected? It is nearly impossible for space- or airborne visible-wavelength surveillance systems to pick out a 10-meter-long missile from a visible scene tens of kilometers on a side. But because every object has a characteristic thermal emission, IR detection avoids those problems; that is, if a warm object arcing across the background of planetary space is headed toward a defended target, it’s probably a missile. And it’s easy for infrared systems to pick out warm objects from a cold background. Unfortunately, there are several additional layers of complexity.
First of all, the geometry of detection usually means that the background is not cold, empty interstellar space, but the warm, cluttered background of Earth. The infrared signature of the target is now camouflaged within the natural scene. There are specific wavebands where thermal emission from Earth is minimized by atmospheric absorption bands, but some residual background is always present. So the natural background makes the task more difficult, but the adversary is about to make things even tougher.
Any state with technology sophisticated enough to deploy nuclear weapons on a ballistic missile is probably sophisticated enough to deploy a range of countermeasures — mechanisms designed specifically to defeat an antiballistic missile system. In April 2000, the Union of Concerned Scientists and Massachusetts Institute of Technology’s Security Studies Program jointly published a report titled “Countermeasures.” In it they state, “An attacker could overwhelm the system by using ‘antisimulation balloon decoys’; that is, by deploying its nuclear weapons inside balloons and releasing numerous empty balloons along with them.
“An attacker also could cover its nuclear warheads with cooled shrouds, which would prevent the kill vehicles from detecting and therefore from homing in on the warhead.” Suddenly, what seemed at least a straightforward task — identifying warm missiles against a cold background — has become a daunting, even formidable prospect.
That’s where multicolor infrared detectors can help. Multiple wavelengths allow a precise determination of emissivity and temperature. The rate of temperature change — or the lack thereof — can potentially separate uncooled balloons from rotating painted aluminum decoys from cryogenically cooled Mylar envelopes. The task is still a challenge, but within the missile-defense community it is believed to be a surmountable one.
In testimony given before the US Congress in November 2000, Lt. Gen. Ronald Kadish, director of what was then the Ballistic Missile Defense Office, stated, “We do, however, have great confidence, based on the testing and analysis we have done so far, that we will be effective against the countermeasures we expect.” And multicolor infrared focal plane arrays will be critical to achieve that goal.