IR IMAGING: MCT to see an avalanche of applications

Mercury cadmium telluride (MCT) has long been one of the materials of choice for detection of infrared radiation, and recent years have seen focal-plane arrays (FPAs) based on MCT grow larger, with individual detectors spaced ever closer, to bring MCT into the imaging arena.

Th Ir 01

Mercury cadmium telluride (MCT) has long been one of the materials of choice for detection of infrared radiation, and recent years have seen focal-plane arrays (FPAs) based on MCT grow larger, with individual detectors spaced ever closer, to bring MCT into the imaging arena.

One recent advance was the use of MCT as an avalanche photodiode, first demonstrated by Jeffrey Beck at DRS Infrared Technologies (Dallas, Texas) in 2001. Since then, several groups have pursued the potential of MCT avalanche photodiodes, incorporating them into FPAs. But a group at CEA-Leti (a laboratory of the Minatec Innovation Center in Grenoble, France) has been pushing the limits of what the detectors can do–as a spate of recent and forthcoming publications bears out.1-4

The results have been filled with surprises, not least of which is the finding that MCT avalanche photodiodes have an “excess noise factor”–the noise incurred by the avalanche gain process–equal to one. That is, the avalanche photodiodes act as perfect amplifiers. Uniquely to MCT, the avalanche process is electron-dominated, meaning that the bandwidth is independent of the gain. This is in sharp contrast to III-V materials or silicon, in which the gain-bandwidth product remains fixed.

“We started to realize how much this exceptional performance would change the field of IR photodetection,” said Johan Rothman, who heads the group at CEA-Leti working on the detectors. “We found that for a slight change of our standard process we could not only reproduce the results obtained by Beck, but we were also able to push the limits of the APD performance, establishing a record high gain of 5300, record low dark current, and a first demonstration of the independence of the response time on the gain.”

The group has pursued a modified “n on p” manufacturing process, and the team’s record gain was reached with a bias of just 12.5 V, with no change to the detectors’ spectral or spatial response.

The ability of MCT arrays to operate with such high detectivity and high dynamic range makes them ideal candidates for a number of applications, the researchers say. Passive imaging would be improved in inclement weather conditions, or could be used to simplify existing systems. The team points also to hyperspectral imaging, which samples a number of discrete, individually low-flux bands–an application that is seeing increased use in satellite technology.

The team has already demonstrated the benefit of using gain in passive imaging (see figure), and is now putting its efforts into active imaging, aiming to demonstrate a 3-D laser detection and ranging (LADAR) system in spring 2009. The approach could easily be adapted for dual-mode passive and active imaging.

Rothman admits that will take some work, not only on the detectors themselves, but also on the readout electronics that lie beneath the array. They are aiming toward a sensitivity below one photon per second, and a response time below 100 ps. ¿¿

Long in coming

What may have stopped the MCT avalanche photodiode idea from coming to fruition earlier is the theory of R. J. McIntyre, who in 1966 proposed that the excess noise factor of avalanche detectors could not go below a limit of two. The result, which would have precluded the use of MCTs in the most interesting applications, has been taken as a fundamental limit that researchers have only recently beaten.

“I think the observation of an excess noise factor lower than the well-established theoretical McIntyre limit is the result which has been central in the development of the MCT avalanche photodiodes,” said Rothman. “It has been a true adventure to try to convince people for whom the simplified theory has become a true quantum-mechanical limit. It is particularly surprising given the strong simplifications that were used in the McIntyre theory.”

Th Ir 01
Without gain, the noise from readout electronics dominates a low-flux IR image (left). A gain of seven makes the image significantly clearer (right). (Courtesy of CEA-Leti)
Click here to enlarge image

The group is awaiting publication of several papers, including its first results on passive amplified imaging. In another, the group demonstrates a new MCT alloy that will maintain the high gain at low bias but be operable at higher temperatures. It is a result to keep an eye out for; applications would snowball if the detectors can get away from cryogenic cooling.

With the benefit of the DEFIR (Design and Future of the IR) initiative, which brings together detector manufacturer Sofradir’s (Chatenay-Malabry, France) research and development team with CEA-Leti’s researchers, the progress made on MCT avalanche photodiodes will be furthered with industrialization in mind. As a result, FPAs based on MCT avalanche photodiodes will make it into commercial cameras that much sooner. –Jason Palmer

REFERENCES

  1. G. Perrais et al., J. Electron. Mater. 37, 1261 (2008).
  2. J. Rothman et al., J. Electron. Mater. 37, 1303 (2008).
  3. J. Rothman et al., Proc. SPIE 6940, 69402N (2008).
  4. F. Guellec et al., Proc. SPIE 6940, 69402M (2008).

More in Detectors & Imaging