Microbolometers offer high resolution at room temperature

17 Jun 2002

The importance of infrared imaging is leading to a booming detector market in low-cost hand-held imaging devices and focal plane arrays. Progress made in the last decade with uncooled detectors and two-dimensional arrays now makes them comparable with III-V and II-VI quantum detectors. Mohamed Henini reports.

From Opto & Laser Europe May 2001

Microbolometer arrays, which consist of detectors made from materials in which the electrical resistivity changes with temperature, produce high-resolution images but do not require cryogenic cooling. The increased resolution of focal plane arrays and microbolometer systems allows the detection of tiny temperature variations.

Jean-Pierre Chatard, technical and business director at Sofradir, a company based in Chatenay-Malabray, France, said: "The commercial market is already big (10 000 a year) and will rise in the next three years to half a million per year. Quantities could be several million a year before 2010."

These numbers refer to the detectors that use uncooled arrays. Chatard believes that the technology is important: "In the infrared domain we used to use cooled detectors based on special materials. The good point is the sensitivity of such detectors; the drawbacks are the cost of the detecting material and the cost and weight of the coolers. The uncooled detectors are simple, light, cheap and don't need maintenance. This technology is at the beginning of its life, but a lot of improvement will come."He is certain that this technology will open up new avenues that cannot be achieved with other technology. He added: "Many applications exist in the infrared domain, but the high price was until now a barrier. The market will be mainly commercial. The military market will be limited to 10 to 20% of production. The applications are still not well known because this market has to be created, but, for the moment, part of the market is identified as fire detection; surveillance; high-power electricity control; and process control and maintenance."

Single-pixel bolometric receivers have been around for some years. However, the production of arrays of bolometers similar to optical and infrared cameras has only come about with microfabrication methods.

Cooled focal plane arrays have been commercially available since the early 1990s. These systems operate in the 3 to 5 µm range and generally provide excellent sensitivity. The latest focal plane array imaging systems with uncooled detectors operate by sensing changes in electrical resistance across the detector (unlike previous infrared systems, which sensed photons). One of the basic types of uncooled detector that has emerged today is the microbolometer, developed by Honeywell in the US.

The Laboratoire d'Electronique de Technologie et d'Instrumentation (LETI) in Grenoble, France, has developed the original uncooled infrared detector microbolometer technology, now covered by a licence agreement with Sofradir. The shareholders are Thomson-CSF (40%), Sagem (40%) and the owner of LETI, the Commissariat à l'Energie Atomique (20 %).

The LETI microbolometer technology is based on an amorphous silicon (a:Si) resistive bolometer. It is 100% silicon compatible and monolithic, making it ideally suitable for low-cost production and highly reliable in high ambient temperatures (more than 100 °C). The main advantages of a-Si microbolometers are a low thermal constant that improves image quality and compatibility with standard CMOS for the read-out circuit. LETI and Sofradir have developed a 320 x 240 array with a 45 µm pitch based on a classical read-out circuit architecture in a 3.3 V, 0.5 µm CMOS silicon technology. First designed and tested in 1999, it is sensitive in the far-infrared (7 to 14 µm) and the power consumption is less than 200 mW at 30 °C.

Jean-Pierre Chatard said: "The Sofradir approach based on amorphous silicon is a low-cost approach for quantities. The sensitivity obtained up to now is better than 80 mK for an aperture of f1 and a temperature regulation at 30 °C. The major interest of this technology is also the time constant, which is lower than 4 ms - four times better than the other technologies."

LETI has also announced a collaboration with Sagem and Sofradir to produce prototype infrared binoculars, called Bolide. Night-vision infrared binoculars and cameras for the military include a cryogenic cooling system operating at about -160 °C that is both heavy and expensive. However, LETI's device operates at ambient temperatures. Each detector consists of a matrix of approximately 76 800 microbolometers, each measuring less than 50 µm. Each microbolometer is associated with a pixel on the final image.

The performance of the Bolide binoculars is expected to be less impressive than that of the existing models on the market in terms of range, but the new binoculars are both lighter and less expensive. Chatard believes that they could also be employed in vehicles to enhance the vision of their drivers. He told OLE: "Bolide is a military product that is a small part of the market. That is why we have developed something equivalent for the commercial market. The next step is to make a smaller array to reduce the price."

French researchers at Alcatel Space (Cannes-la-Bocca), Sofradir and Thomson-CSF Optronique (Guyancourt) have created a two-dimensional microbolometer camera working in the 8 to 12 µm infrared range, to be carried by the Fuego Earth observation mission, which aims to reduce damage caused by forest fires in the Mediterranean. The device will discriminate between clouds and smoke, detect false alarms and monitor the fires. The camera needs a large dynamic range of detectable radiances, a small volume, and low mass and power consumption. Alcatel developed the thermal infrared camera, Sofradir the two-dimensional microbolometer detectors and Thomson a hybrid lens.

In the UK, Mike Mansi of Infrared Integrated Systems coordinates a project called "the affordable uncooled technology with hermetic encapsulation for thermal imaging camera" (AUTHENTIC). Its aim is to produce an extremely low-cost, mid-performance infrared imaging camera quickly.

The hand-held camera will be sensitive in the 8 to 14 µm waveband and resolve temperature differences of 1 °C. It is intended for a range of commercial and domestic applications, including safety (fire detection and collision avoidance), preventative maintenance (hot-spot detection in cabling and bearings) and security (intruder alarms and perimeter surveillance).

The consortium, which includes Mitel Semiconductor (UK), National Microelectronics Research Centre, University College Cork (Ireland), Infrared Integrated Systems (UK), and Phoenix VLSI Consultants (UK), aims to restrict manufacturing costs to less than 71000, assuming volume production of 10 000 units per annum, permitting an approximate retail value of 71500 to 72000. The consortium will develop a new micromachined, detector array fabrication process that will be transferred to the commercial foundry at Mitel.

The process will integrate microbolometers with associated read-out electronics on a single CMOS wafer. The fabrication process is 100% "back-end" compatible and only uses standard foundry materials - the key to low-cost manufacturing. The focal plane array of microbolometers forms the thermally sensitive element of an infrared imaging camera. The project started on 1 January 2000 for a duration of 30 months, but, as yet, Mike Mansi is not at the point of releasing any technical data.

The major firms in Europe working on microbolometers and uncooled detectors are British Aerospace, Sofradir and a new start-up in Belgium. Other firms worldwide that are manufacturing uncooled detectors include Raytheon, Boeing and Lockheed Martin. Sofradir is considered by some to be the leader and already has a large market share predicted to be 1200 units in 2001 and 10 000 in 2002.Philippe Bensussan, managing director of Sofradir, summed up the state of the market: "Uncooled infrared detection represents a strategic diversification. Whereas until now Sofradir has based its growth on military applications, the uncooled infrared detectors developed by LETI will enable us to diversify into some very attractive civilian markets, such as remote surveillance, industrial control and, looking further ahead, the car industry.

"The price of top-grade infrared cameras will be halved, perhaps even quartered. Technology transfer is under way and we have orders already. Ultimately, uncooled detectors could represent as much as 30% of our activity."

The word bolometer comes from the Greek bole, which means ray. Invented in 1860 by US scientist Samuel Pierpont Langley, these thermal radiation detectors measure tiny amounts of radiant energy, ranging from light waves to microwaves.

The incident photons generate thermal phonons by transferring their energy to the absorber. This causes a change in the current flowing through a thermistor. The temperature of the absorber changes in response to changes in incident power and these temperature fluctuations are measured with a sensitive thermometer.

These thermal detectors are capable of responding to a range of wavelengths without appreciable variation in responsivity and they display high sensitivity at room temperature, which permits imaging applications.

In contrast, photon-type detectors, which operate on the principle of direct electron-photon interaction, have a better sensitivity and response speed. However, they typically require costly cryogenic cooling to minimize noise to achieve high sensitivity.

Most of the newer camera designs are based on a focal plane array device. Some of the most popular focal plane array technologies include both uncooled and cooled focal plane array systems. Detector arrays that operate at or above room temperature are known as "uncooled", the term distinguishing these detectors from standard ones, which only operate at cryogenic temperatures or below.

The advantages of uncooled systems are lifetime and cost. Early infrared sensors needed to be cooled to as low as 77 K by using liquid nitrogen or compressed gases. Thermal imaging systems have therefore evolved from cumbersome systems, often weighing more than 20 kg, to systems that fit in the palm of the hand.

At present there are four primary detector technologies: mercury cadmium telluride, indium antimonide, quantum-well infrared photodetectors, and microbolometer/pyrometer focal plane arrays.

Uncooled microbolometer technology is proposed for the following applications: surveillance and security; non-destructive testing; process control; driver vision enhancement; predictive maintenance; medical imaging; firefighting; quality assurance; and military applications, such as weapon and vehicle sights and goggles.

The spectral response for commercial imagers generally falls into two categories: 2 to 5 µm (near infrared) and 8 to14 µm (far infrared). Commercial infrared imagers are not manufactured in the 5 to 8 µm range owing to the atmospheric absorption of infrared energy at these wavelengths. Microbolometers are sensitive in the long wave (8—16 µm) range.