Infrared sensor designers move beyond their goal of simply owning the night

New design trends in IR detectors involve staring arrays, ever-shrinking cooled sensor subsystems, uncooled devices that are within the size and budget constraints of many new applications, and multispectral sensors that provide more information while reducing the signal-processing load

Th Mae72435 55

New design trends in IR detectors involve staring arrays, ever-shrinking cooled sensor subsystems, uncooled devices that are within the size and budget constraints of many new applications, and multispectral sensors that provide more information while reducing the signal-processing load

By John Keller

The infrared sensor took center stage seven years ago during the Persian Gulf War, when these devices turned the desert night into virtual daylight for aircraft pilots and tank crewmen. These devices most often were large, expensive, cooled devices suitable for mounting only on large platforms such as aircraft, ships, and missiles.

In today`s landscape of infrared sensors, the cryogenically cooled device - albeit improved - still plays a big role, particularly in military applications whose users demand the highest possible sensor range, resolution, and sensitivity. Yet cooled infrared (IR) sensors are only part of the picture. New design trends involve infrared sensors that lack a cooling mechanism, which dramatically reduces the sensor`s size, weight, power consumption, and price. On the horizon is a new design approach called quantum well infrared photoconductor - QWIP, for short - that is increasing the attractiveness of a detector material for IR sensors that engineers typically associate with radar - gallium arsenide (GaAs).

The relatively low price of uncooled IR sensors, coupled with skyrocketing commercial demand, is fueling an environment that offers the military and aerospace systems designer an ever-less-expensive, and more reliable IR sensor option than he has ever had before.

Also playing an increasing role in cooled and uncooled IR sensor systems are the emerging multispectral optical sensors that are able to detect electromagnetic energy in the longwave IR, midwave IR, and sometimes even in the ultraviolet (UV) spectra. Multispectral sensors offer users far greater flexibility than do mono-spectral sensors, and with the addition of digital signal processing offer the user a multicolor overlay-type IR image that displays a wealth of knowledge at a single glance.

Cooled sensors

The growing popularity of uncooled infrared sensors is not to say that the era of cooled sensors is at an end - far from it. Fitting IR sensors with cryogenic coolers, despite the relatively large size, heavy weight, and risk of failure of these devices, is still essential for making the most of the IR sensor`s range, sensitivity, and resolution, experts say.

The new breed of cooled sensors involves a broadening number of detector materials, ever-more-efficient cooling systems, and staring arrays that enhance image clarity, reduce the image "bloom" associated with wide temperature variations, and reduce the need for image processing to filter out image-degrading optical "noise." Detector materials of choice for cooled IR sensors typically are mercury cadmium telluride, indium antimonide, or platinum silicide.

Typical applications of high-end cooled IR sensors include missile guidance, missile warning receivers, and ballistic missile defense. Cryogenically cooled IR sensors operate at -196 degrees Celsius, which is the temperature of liquid nitrogen. "Those arrays act as if you put 78,000 rain gauges out to store energy, measure it, and convert it into visible light for images," says Bill Martin, vice president of FLIR Systems Inc. in Portland, Ore. "They can detect extremely minute changes in temperature."

Military officials most often are looking for the highest performance possible, which inevitably drives them to cooled IR sensors. "Most cooled sensors applications will stay cooled," Martin says. "The military is never satisfied with detection ranges. Targeting FLIRs default to highest-performance sensors. Cooled applications will remain nichy, and mostly in the military. For most commercial requirements, uncooled sensors will fill that need."

For the military, performance, not cost, continues to be the chief driver. "Our customers put a premium on sensitivity - on finding small targets in clutter. In large-format, high-resolution, TV-quality sensors, we use the latest advances we can in infrared focal plane array technology, for longwave in the 8-to-12-micron band or midwave in the 3-to-5-micron band," explains Scott Porter, director of combat electronic systems marketing at the Northrop Grumman electronic sensors and systems division in Baltimore. "Our enabling technologies are high-resolution focal plane arrays and wideband digital processing. These allow us to take the large amounts of information that these sensors are seeing, process it to a meaningful picture or format, and make data that an operator can use."

Engineers at FLIR Systems manufacture cooled and uncooled IR sensor subsystems - most notably an enhanced version of the U.S. military AN/AAQ-22 airborne thermal imaging system called Star SAFIRE. U.S. military officials have certified Star SAFIRE for more than 20 fixed-wing aircraft and helicopters, including the U.S. Navy Lockheed Martin P-3 Orion maritime patrol turboprop, the U.S. Air Force Lockheed Martin C-130 cargo turboprop, the U.S. Army/National Guard Bell UH-1N, Sikorsky HH-60, S-61, and S-76 helicopters.

"The SAFIRE family is at the top of our list in terms of hot products," FLIR`s Martin says. "Star SAFIRE offers flexibility that you couldn`t get a few years ago in size. It is our most successful military product ever."

The Star SAFIRE, which will be available by this spring as a replacement for the AN/AAQ-22, features a high-performance 320-by-240-pixel IR focal plane array, and has a more-efficient Stirling cooler than the existing system. It is smaller and uses less power than the AN/AAQ-22, and has 4-quadrant dithering that electrically and mechanically moves the array in two directions to improve resolution. "We get resolution that is closer to 640-by-480 resolution," Martin says. The system also can package as many as four sensor payloads such as the IR detector, laser rangefinder, visible-light camera, and spotter scopes.

"Ruggedizing sensors for helicopters is at the core of our business," he explains. "We use extremely robust components for stabilization, and slip rings, and all is designed to withstand aggressive shock, vibration, and temperature extremes."

Breakthrough enabling technologies Martin cites in the Star SAFIRE include microprocessor control "that has dramatically improved our image processing and system management. We use a significant number of DSP and Intel Pentium processors."

Leaders of Sanders, a Lockheed Martin Company in Nashua, N.H., are leveraging their GaAs experience, their ability to use molecular beam Epitaxy to grow integrated circuit material, and the manufacturing base they established for radar and electronic warfare applications into QWIP development. "Mercury cadmium telluride has been the material of choice for uncooled IR sensors, but QWIP is based on GaAs," says John Ahearn, director of integrated optoelectronics in the Sanders Microwave Electronics Division.

Sanders IR experts are using QWIP technology to build a cooled infrared focal plane array (IRFPA) that operates in either the 8-to-12-micron or 3-to-5-micron wavelengths. The target application is an upgrade to the Lockheed Martin Low-Altitude Navigation and Targeting Infrared for Night (LANTIRN), which engineers at the Lockheed Martin Electronics & Missiles Division in Orlando, Fla., are integrating. The upgrade program, LANTIRN 2000, is for jet fighter-bombers such as the U.S. Air Force F-16 and F-15E, as well as the U.S. Navy F-14 and F/A-18. In addition to a QWIP forward-looking infrared (FLIR) system, the LANTIRN 2000 upgrade will incorporate a 40,000-foot altitude diode-pumped laser, and a compact, powerful computer system.

QWIP technology "is more producible, lower cost, and represents a paradigm shift" in sensors over competing materials and design approaches, Ahearn explains. "The mercury cadmium telluride material is so limited that people use scanning arrays to move the image across the sensor. QWIP allows us to go to a staring array."

Of all its advantages over mercury cadmium telluride and even over indium antimonide or platinum silicide, the producibility of GaAs IR sensors is perhaps most profound. This approach piggybacks on advanced semiconductor manufacturing techniques to leapfrog over competing materials in terms of device density and reliability.

GaAs for infrared detectors

GaAs is a common substrate for solid-state transceivers that go into advanced phased-array radar sensors, as well as for some of the highest-speed semiconductors in niche commercial applications such as mobile communications, optoelectronics, photonics, high-speed data transfer, direct broadcast satellite, and cable TV. Electrons move five times faster through GaAs than through silicon, which not only makes them attractive for demanding applications, but also provides an economy of scale advantage to GaAs manufacturing over mercury cadmium telluride and other IR detector materials. By some estimates, the market for gallium arsenide, in fact, is expected to grow into a $1 billion market this year and reach $1.8 billion by 2000.

"Mercury cadmium telluride has not had the large enough substrate or uniform enough material" to compete on an even footing with GaAs-based QWIP technology, Ahearn explains. For example, a common-module scanning mercury cadmium telluride detector has a resolution of 60-, 120-, or 180-by-1 pixels. The existing version of LANTIRN, in fact, uses a scanning mercury cadmium telluride IR detector with 180-by-1-pixel resolution, he says.

It is necessary for the scanning-type sensor to move over the image perhaps hundreds of times to craft a high-resolution image line-by-line. Moreover, the scanning processing makes the image much more susceptible to the thermal "blooming," or whiteout problem from vastly different temperatures. A staring QWIP IR detector from Sanders, by contrast, yields a 640-by-480-pixel image in an instant, with little, if any, blooming, and TV-quality resolution, Ahearn says.

GaAs, by itself, does not detect IR energy in the wavelengths of most interest to IR detector designers - 3 to 5 microns, and 8 to 12 microns, Ahearn explains. GaAs, instead, detects electromagnetic energy in the 85-micron range. "So we create a quantum well structure that allows you to have absorption in the proper wavelength," he says. "Layered structures artificially create an absorbing layer in the 8-to-12-micron region."

In essence, the QWIP approach places wells, or "dimples," in a flat detector surface to collect electrons, Ahearn explains. Then IR photons pull the electrons up out of the wells and onto the detector plane, where the photons spread out and collect as photocurrent, he says. "We are probably the only industry laboratory doing this," Ahearn says. Scientists in a West Coast government lab, the NASA Jet Propulsion Laboratory in Pasadena, Calif., also are pursuing QWIP technology for space-based IR sensors.

Although he cannot cite numbers, Ahearn says QWIP is substantially less expensive to produce than mercury cadmium telluride, and in fact says U.S. Air Force officials considered and rejected a second-generation mercury cadmium telluride IR detector for LANTIRN 2000. Even though it requires a cryogenic cooler, Ahearn says QWIP technology should also be attractive for civil applications such as surveillance, aircraft cockpit situational awareness and synthetic vision, and medical imaging.

Even though QWIP`s advantages make it sound like it represents the future of cooled IR sensors, there are applications where it probably is not the best choice. One example is missile guidance, Ahearn says. Compared with many surveillance sensors, "missile seekers run a higher temperature and operate with fewer pixels per inch," Ahearn says. Typical missile seekers must cool down to operating temperature very quickly, and operate at between 64-by 64 pixels and 256-by-256 pixels, he says.

Sanders engineers are producing 3-inch GaAs wafers for QWIP sensors today, and are preparing to install equipment to produce 6-inch GaAs wafers. Company experts could be ready to produce 6-inch GaAs wafers in quantity within the next two years, he says.

Even though the rule for cooled IR sensors hitherto has involved large, platform-based systems, these sensitive detectors also are beginning to make their way into hand-held devices.

Engineers at the Cincinnati Electronics Corp. Detector and Microcircuit Devices Laboratories in Mason, Ohio, have developed the hand-held, battery-powered NightMaster ruggedized miniature infrared imager, which uses a 256-by-256-pixel indium antimonide array. This IR surveillance camera provides continuous operation for periods as long as two hours. The 0.25-watt Stirling cycle cooler is the Model K508A from Ricor Ltd. of En-Harod, Israel.

Several other companies have strong reputations for producing cooled and uncooled infrared sensors, such as the Hughes Santa Barbara Research Center in Goleta, Calif., Raytheon Amber in Goleta, Calif., the Texas Instruments Defense Electronics Group in Dallas, and the Hughes Sensors & Electronic Systems division in El Segundo, Calif. These companies, however, are consolidating into new owner Raytheon Co., and were not available for comment.

Uncooled sensors

Despite their superior sensitivity and range, cooled IR sensors continue to be relatively problematic in terms of cost and reliability - chiefly because of the coolers that are necessary to operate them.

"The costliest item, and most likely failure, is its cryogenic cooler," explains Northrop Grumman`s Porter. "It is still the pacing item. If you could find applications where you could do away with that, it takes out a costly component and removes mechanical components subject to failure and makes your packaging easier. You can see the allure of maturing the technology."

Even advocates of cooled IR sensors point out the benefits of uncooled devices. "Coolers are complex and susceptible to failure," says Martin at FLIR Systems. "They take more space and use effectively 10 times the power, will weigh three to four times as much, and will last one-eighth the time as uncooled sensors - or worse."

At FLIR Systems, engineers typically find their mean time between failures of cooled sensor systems to be 3,000 to 4,000 hours, Martin says. Contrast that with the mean time between failures at FLIR for uncooled IR sensor systems - 40,000 to 60,000 hours. "Cooled will remain nichy, and mostly in military," Martin says." For most commercial requirements, uncooled will fill that need."

The emergence of uncooled room-temperature IR sensors is not only a boon for military forces whose soldiers, sailors, and airmen fight at night, but also for civil uses such as law enforcement, lights-out nighttime driving, forest fire hot spot monitoring, finding power line leakage, and even for improving home insulation by detecting where heat leaks out.

Surprisingly, the rapidly growing popularity of uncooled devices only rarely leads to the replacement of military cooled IR sensors with their uncooled counterparts, experts say. What is happening, however, is the appearance of IR sensors where they have never been before, such as aids to civil aircraft pilots, vehicular safety, and perimeter security. Uncooled IR sensors also are becoming attractive for hand-held cameras, unmanned aerial vehicles (UAVs), and even for infantry rifle night sights.

Uncooled IR sensors essentially come in two technologies - microbolometers and ferroelectric bolometers. Unlike a cooled IR sensor, which detects photons in the infrared wavelengths, the microbolometer senses thermal energy and converts it into electrical energy. It actually heats up when IR energy hits it, and its electrical resistance changes with different amounts of IR energy. The sensor measures the electrical resistance by applying a bias current to the detector. The ferroelectric bolometer also senses thermal energy and converts it to electrical energy, but by using the change in the dielectric constant of a ferroelectric material when it is heated beyond its Curie temperature.

The departure of scanning sensors is perhaps the most promising design trend in uncooled detectors, experts say. "What is new is uncooled staring focal plane arrays," says FLIR`s Martin. "The benefits are in the reliability over scanned, and in sensitivity. Every pixel stares at the target."

Engineers at Microcam Corp. in Waltham, Mass., for example, are using uncooled IR detector technology to create a sensor small enough for a hand-launched pilotless aircraft called Pointer that about the size of a hobbyist`s model airplane. Microcam`s project, which company engineers are performing for the U.S. Naval Air Systems Command at Patuxent River Naval Air Station, Md., will be a 12-ounce IR camera that senses IR energy in the 8-to-14-micron wave band, runs on 4.5 watts, and receives power from the aircraft`s main batteries.

The Honeywell-designed 320-by-240-pixel IR focal plane array detector head weighs eight ounces, and with the lens fits into a 3-inch cube. The imaging sensor is to be a plug-in replacement for the currently used charged-coupled-device visible-light camera. "The IR imagery is indistinguishable from black-and-white TV imagery, Bodkin says.

"It is the smallest UAV they are working on now, and it has about a 4-foot wingspan," says Microcam President Andrew Bodkin. "It now has a TV transponder so you can fly by sight, but they now need an IR camera so we can fly it at night. It has no motor noise, can be backpacked out into the field, and can go a couple of kilometers." Flight tests were to be in last March at Patuxent River NAS.

Multispectral sensors

Perhaps the most advanced trend in IR sensors is multispectral imaging. This approach seeks to operate different kinds of sensors from the same aperture - sometimes from the same semiconductor device. These sensors may detect different IR bandwidths, different light spectra, or even blend optical and RF detectors, as well as laser rangefinders.

"Today in IR we have pushed the state of the art in imaging for target recognition and precision aim point tracking," explains Bill McGuinness, manager for advanced weapons business development, and Northrop Grumman Electronic Sensors and Systems Division, in Baltimore. "We own the night. The next issue in the weapons community is owning the weather, owning the obscurants, and owning the jamming environment - whether that be GPS [Global Positioning System], or IR countermeasures.

"We need to process entire IR scene to look for target," McGuinness continues. "If you run automatic target recognition (ATR) in the imagery, you need to process the entire scene. Issue then is processor size and cost. So the idea is to beat down the overall cost of the seeker subsystem by minimizing processing power. How do you do that given that you rely heavily on IR system? In our view you go multispectral."

Northrop Grumman engineers, for example, are developing a dual-mode missile seeker called the MMW/I2R Common Aperture Seeker that blends IR and millimeter wave radar sensors for autonomous acquisition and tracking of targets.

Combining different types of sensors fine-tunes IR imaging sensor data by enriching the amount of information available, and reduces the need to digitally filter out unknown or unwanted information as "noise."

"If you can give that IR sensor range information, that helps with the automatic-target-recognition processing because it can size the IR templates to the IR scene," McGuinness explains. "If I operate multispectral with a 3.5-micron imaging IR, co-boresighted with imaging millimeter wave radar, I can look at the same target at the same time at the same angle. It allows you to eliminate a large area of the IR scene as non-target, and minimizes your processing load. Now with range information I can size template with ATR. That`s the way we are going."

Lockheed Sanders experts likewise are looking to their GaAs QWIP technology for future multispectral sensor applications. QWIP has the ability to detect more than one light bandwidth. Sanders`s Ahearn suggests the future possibility of a multi-color display showing IR energy detected in the 3-5-micron band, in the 8-12-micron band. This approach, he says would suit well to demanding applications such as missile warning detectors. It would provide better target discrimination than missile warning systems deliver today, and could substantially reduce the number of false alarms.

In addition, Ahearn suggests, designers also might be able to stack a GaAs QWIP sensor on top of a nitride semiconductor to combine for multispectral IR sensing, as well as ultraviolet sensing. "You could grow the material on top of each other, all with GaAs," he says. This choice might be superior for missile warning receivers, since ultraviolet detectors are particularly adept at sensing electromagnetic energy from missile plumes.

Th Mae72435 55
Click here to enlarge image

Unmanned aerial vehicles such as that pictured above are prime platforms for sophisticated imaging infrared sensors.

Th Mae72435 56
Click here to enlarge image

Th Mae72435 57
Click here to enlarge image

The Star SAFIRE imaging infrared sensor from FLIR Systems is a multi-payload package that consists of a narrow- and medium-field-of-view IR sensor, visible-light camera, laser range finder, and laser illuminator. It is fitted to several military and aviation platforms, including the U.S. Army HH-60 Black Hawk helicopter, at right.

Th Mae72435 58
Click here to enlarge image

The Dual-Mode Seeker from Northrop Grumman Corp., pictured above, consists of an infrared sensor and millimeter wave radar to detect targets at night and in smoke and dust.

Th Mae72435 59
Click here to enlarge image

The IR Microcam sensor from Microcam Corp., pictured above, is an uncooled system intended for extremely small unmanned aerial vehicles that must navigate during the day and at night.

Th Mae72435 60
Click here to enlarge image

Ever-smaller infrared sensors are giving rise to a new class of hand-held IR cameras for military and civil surveillance applications. Pictured above is the NightMaster IR camera from Cincinnati Electronics Corp.

Infrared sensor market expected to grow through 2004

The market for imaging infrared focal plane array sensors and systems is expected to grow by a compound annual growth rate of 29 percent through at least 2004, according to a study by analysts at Frost & Sullivan of Mountain View, Calif.

The study, entitled U.S. Markets for Commercial and Military IR Imaging FPA Sensors and Systems, places the greatest market growth in emerging IR technologies such as microbolometer, ferroelectric, and quantum well infrared photoconductors.

Other market opportunities will be in very-high-performance IR technologies as military programs begin production and require large numbers of sensor systems, the study says.

Market drivers include military and paramilitary IR imaging system upgrades, and a quickly growing commercial market, primarily for uncooled IR sensors such as microbolometers and ferroelectric bolometers.

The study, released in January, segments the infrared sensor market in two different ways. First, the study breaks the market into sensors and integrated systems.

Second, the study divides the market into five major segments - mercury cadmium telluride, indium antimonide, platinum silicide, extrinsic silicon, and emerging focal plane arrays and imaging systems such as microbolometers, ferroelectric bolometers, and quantum well infrared photoconductors.

In 1996, total revenues for imaging infrared focal plane array sensors were $283.7 million. Revenues for IR sensor integrated systems, meanwhile, were $1 billion, according to Frost & Sullivan figures.

The other infrared focal plane array sensor markets in 1996, according to Frost & Sullivan figures, were:

- mercury cadmium telluride arrays - $190 million for sensors, and $535.7 million for systems;

- indium antimonide arrays - $53 million for sensors, and $264 million for systems;

- platinum silicide arrays - $16.4 million for sensors, and $82 million for systems;

- extrinsic silicon arrays - $5 million for sensors, and $25 million for systems;

- ferroelectric arrays - $9.6 million for sensors, and $48 million for systems;

- microbolometer arrays - $9.7 million for sensors, and $46.2 million for systems; and

- quantum well infrared photoconductors - no revenue in 1996 because the technology was only reaching the marketplace that year. In 1997, however, QWIP technology was worth $3.5 million for sensors, and $17.5 million for systems. - J.K.

More in Computers