Two Wells Are Better than One for IR Detectors

Hank Hogan

Researchers at Princeton University in New Jersey have demonstrated that two wells are better than one as far as quantum-well infrared photodetectors are concerned. The key, as demonstrated in prototype devices, lies in making dual wells that are separate and unequal. This enables infrared detection over a wider spectral range than is otherwise possible, and it produces detectors that are more robust.

"Unlike all other schemes demonstrated so far, the detector can be biased at any voltage and can be operated at any temperature below the background-limited temperature," said Amlan Majumdar, who is now at Intel Corp. in Portland, Ore.

The binary superlattice structure has two minibands (M₁ and M₂, left), each with eight energy levels. The spectral response of the quantum-well IR detector is largely independent of temperature and voltage (right). Courtesy of Amlan Majumdar.

The Princeton group created the quantum-well detectors using a two-step process. The first consisted of growing the quantum wells on 3-in. GaAs wafers using molecular beam epitaxy systems from Sandia National Laboratories in Albuquerque, N.M., and from IQE Inc. of Bethlehem, Pa. This deposited 18 layers of alternating silicon and barrier strips. The silicon, which was uniformly doped by trace impurities, served as the active element.

The researchers completed the processing at the university, creating detectors of different widths sitting close to each other in what the scientists call a binary superlattice. Because of the interaction between the quantum wells, the detector has a spectral response width of 5.6 µm in the 8- to 14-µm range suitable for atmospheric transmission. This, Majumdar said, is a wider response than any other detector provides.

They characterized detectors from the two wafers, which revealed variations in their spectral response. The reason for this, he suggested, is the different background impurity levels in different molecular beam epitaxy systems. This suggests that manufacturing will have to be done on a single system to produce consistent results.

As for possible applications, IR detection has been used primarily by the military, but it could be used for process monitoring, detection of power cable faults and fire fighting. Majumdar said it could also spot cancer through differences in body heat. Clinical trials are under way to evaluate this, and research is ongoing to achieve the required detector technology. "For cancer detection, what one really needs is two-color detection, that is, detectors that respond in two different wavelength ranges," he said.

He added that the results have been promising and that prototypes may be demonstrated soon.