The objective of this work was to lay the groundwork for the development of a new tunable II-VI infrared (IR) material system using mature III-V semiconductors as lattice-matched substrates. Mercury cadmium selenide (HgCdSe) was studied as an alternative to mercury cadmium telluride (HgCdTe) as an IRdetecting material.
The advantage of the HgCdSe system is twofold. First, it is tunable to any IR wavelength of interest by controlling the cadmium (Cd) composition, and second, it is lattice-matched to mature III-V semiconductor systems, such as gallium antimonide (GaSb) and indium arsenide (InAs). By using a lattice-matched mature substrate technology, in principle, HgCdSe can be grown with limited generation of dislocations, resulting in extremely high-quality and uniform IR sensing material, which would greatly improve operability of the focal plane array (FPA). This is completely analogous to HgCdTe grown on lattice-matched bulk cadmium zinc telluride (CdZnTe) substrates, which result in two orders of magnitude reduction in dislocation density over HgCdTe grown on silicon (Si) substrates.
Previous experience in growing the IIVI IR detecting compound, HgCdTe, by molecular beam epitaxy (MBE) was leveraged to develop a completely new IR sensing II-VI system, HgCdSe. Currently, HgCdTe is the material used in the majority of fielded Army IR systems, and much effort has been expended to push the technology to both large-format and low-cost systems while still maintaining superior performance. However, this technology has been sitting at a roadblock for several years that, to date, has not been overcome, and no solution appears imminent. Namely, in order to achieve large-format sizes as required for third generation, a new Sibased composite substrate technology needed to be developed since current lattice-matched substrate technology (bulk grown CdZnTe) is severely limited in size and scalability (a maximum of 6 × 6 cm2 is commercially available).
Due to the huge lattice mismatch between Si and HgCdTe (19%), and associated strain energy, misfit dislocations need to be generated somewhere in the thin film stack to alleviate this energy, which ultimately propagates into the IR-absorbing layer. Generally, two orders of magnitude higher dislocation density is present in scalable HgCdTe/Si material with respect to non-scalable HgCdTe/CdZnTe. It has been demonstrated that this higher dislocation level results in lower device performance. There is ongoing effort to either reduce dislocation density in HgCdTe/Si or render the dislocations electrically inactive, but, to date, no single approach appears to offer a clear and immediately successful option.
This work proposes that HgCdSe will act very similarly to HgCdTe in terms of its IR material properties, but will have the great advantage of having commercially available large-area (and scalable) substrates readily available. Additionally, previous work on MBE growth of CdSeTe indicated that the Cd-selenide (Se) bond strength is stronger than the Cd-telluride (Te) bond strength. It is expected that this will be another advantage of HgCdSe over HgCdTe in that the material itself may be more tolerant of dislocations due to this fact.
Mercury selenide (HgSe) is a semimetal, and cadmium selenide (CdSe) is a wide-bandgap semiconductor, both with nearly identical lattice constants (HgSe = 6.08 A, CdSe = 6.05 A). By forming an alloy of Hg1-xCdxSe, the bandgap of the material can be tuned to absorb any wavelength of IR light resulting in an ideal material to cover the entire IR spectral range. In addition, two III-V binary semiconductors are available to use as substrates that are nearly lattice-matched to HgCdSe, specifically InAs and GaSb. These substrates are developed and readily available from commercial suppliers.
By using III-V bulk substrates, a starting template for HgCdSe growth with a very low dislocation density will be available. Presently, GaSb is quoted as having a dislocation density of less than 104 cm-2. Even with the small lattice mismatch present between GaSb and HgCdSe (f ~0.70% depending on the exact HgCdSe composition), the final HgCdSe dislocation density is expected to measure in the 104 cm-2 to low 105 cm-2 range. This type of dislocation density is exactly what is achieved for HgCdTe material grown on bulk CdZnTe substrates, and which produces state-of-the-art infrared focal plane arrays (IRFPAs) for numerous applications. There is no reason to assume that achieving HgCdSe material with the same dislocation density value will not yield similar IRFPA performance.
This work was done by Gregory Brill and Yuanping Chen of the Army Research Laboratory. ARL-0125
This Brief includes a Technical Support Package (TSP).
New Material System for 3rd Generation IR Applications
(reference ARL-0125) is currently available for download from the TSP library.
Don't have an account? Sign up here.