A concept for optimizing the designs of two-wavelength GaAs-based quantum-well infrared photodetectors (QWIPs) could make it practical to use focal-plane arrays of QWIPs as image sensors in two-wavelength infrared cameras. Potential applications for such cameras include surveillance, tracking of military targets, night vision, and thermal mapping. One important advantage of a two-wavelength over a one-wavelength camera is the possibility of using Planck's radiation law to calculate the temperature of an imaged object from the ratio between the brightnesses of the object at the two wavelengths.

Concepts for using focal-plane arrays of GaAs-based QWIPs in different configurations as two-wavelength infrared image sensors have been described in previous articles in NASA Tech Briefs. The present concept applies to a configuration in which each pixel in a focal-plane array would contain two stacked multiple-quantum-well (MQW) photodetectors. The energy depths of the wells, the geometric thicknesses of the wells, and the geometric thicknesses of the barriers between the wells in one MQW structure would be chosen to obtain peak response in the desired long-wavelength infrared (LWIR) band; the corresponding parameters for the other MQW structure would be chosen to obtain peak response in the desired medium-wavelength infrared (MWIR) band (see figure). Both MQW structures would be biased and read out simultaneously, but independently of each other, through separate indium bump contacts connected to a readout multiplexer.

This Configuration and Its Conduction-Band Electron-Energy Diagrams are typical for a device like that described in the text. The device would be grown by molecular-beam epitaxy onto a semi-insulating GaAs substrate. The contact layers would be made of heavily doped GaAs. Indium bumps would serve as contacts to external circuitry.

Research on previous designs based on this configuration has revealed two problems that until now have made it impractical to make a two-wavelength infrared camera. The first problem is that for simultaneous detection two-QWIP stack in each pixel must be supplied with two different bias potentials, which can not be obtained from any currently available readout multiplexers. The second problem is that in voltage-tunable design a high bias potential (>8 V) must be supplied to the LWIR QWIP to switch on LWIR detection.

The present concept would make it possible to operate the QWIPs with equal and lower bias potentials. The QWIPs designed according to this concept would be at least as responsive as are those of older designs that require higher and unequal bias potentials. The concept involves the following reasoning: The need for different and higher bias voltages in older designs arises, in part, from fundamental physical mechanisms that make it necessary to use larger bias electric fields to detect photons of shorter wavelengths. One could obtain a given bias electric field at a lower bias potential by applying that potential across a lesser thickness of material; that is, across fewer quantum-well periods. For fundamental physical reasons, it turns out that the responsivity of an MQW device is independent of the number of MQW periods (provided that the electric field remains the same), so that one has some design flexibility to decrease the number of MQW periods.

The relevant design parameters must satisfy the following equations:


where FL and FM are the electric field needed for detection of LWIR and MWIR photons, respectively; VL and VM are the bias potential applied to the LWIR and MWIR structures, respectively; NL and NM are the number of spatial periods in the LWIR and MWIR MQW structures, respectively; and LL and LM are the depths of a quantum-well period of the LWIR and an MWIR MQW structures, respectively. Heretofore, one would typically choose LL = LM and NL=NM, and choose two different bias potentials to obtain the required FL and FM. One can still choose LL = LM while choosing different values of NL and NM to make it possible to useVL = VM to obtain the required FL and FM. The required values of NL and NM would then be related by


To obtain reasonably high reponsivity from each stack, FM >FL is required. Thus, the obvious choice would be NL> NM. It is also worth noting that the total number of periods in the structure (NL+NM) is limited by the molecular-beam-epitaxy (MBE) growth time because it tends to increase the number of defects with the increasing growth time. Therefore, one has to increase NL and decrease NM. from their values being equal. Although, this choice would result in a decrease in LWIR noise current and increase in MWIR noise current, the noise current of the LWIR QWIP is still higher than that of the MWIR QWIP. This is because dark current of the LWIR QWIP is few orders larger than that of MWIR QWIP. Therefore, overall performance of the two-wavelength QWIP will be enhanced.

This work was done by Sumith Bandara, Sarath Gunapala, and John K. Liu of Caltech for NASA's Jet Propulsion Laboratory. In accordance with Public Law 96-517, the contractor has elected to retain title to this invention. Inquiries concerning rights for its commercial use should be addressed to

Technology Reporting Office
Mail Stop 122-116
4800 Oak Grove Drive
Pasadena, CA 91109
(818) 354-2240

Refer to NPO-20279

This Brief includes a Technical Support Package (TSP).
Optimized two-wavelength focal-plane array of QWIPs

(reference NPO20279) is currently available for download from the TSP library.

Don't have an account? Sign up here.