

A simple iterative calibration process using in-situ InP growth rate measurements combined with ex-situ x-ray and photoluminescence measurements was used to reproducibly grow the InPAsSb. Compositions of AlInAsSb were calibrated by in-situ growth rate measurements of AlAs and InP to obtain group-III arrival rates, with the As/Sb flux ratio adjusted to achieve lattice matching.

Growth temperatures were 440-490 ☌ for the absorbers, 500 ☌ for the AlAsSb barriers, and 440 ☌ for other InAs layers. All layers were nominally lattice matched to the InAs substrate and typically had lattice mismatch of less than 10 -3. Starting with n+, S-doped, (100)-oriented InAs substrates, the epitaxial structures consisted of 100 nm of an n+ (Si-doped 2×10 18 cm -3) InAs buffer, 4000 nm n (1×10 16 cm -3) quaternary absorber, 100 nm undoped AlAs 0.16Sb 0.84 barrier, and 150 nm of n-InAs with doping graded to 1× 10 18 cm -3 at the surface. Anticipating the development of a viable substrate removal technique, we have investigated the potential of two quaternary alloys, AlInAsSb and InPAsSb, as absorber materials on InAs substrates for nBn-architecture 17 detectors having cutoff wavelengths shorter than 3.0 μm.ĭevices with a simple nBn detector epitaxial structure were grown by molecular beam epitaxy using solid sources except for P, which was supplied using a cracking PH 3 injector. The primary drawback to this approach is that the InAs substrate itself is highly absorbing in the e-SWIR region and therefore must be removed in the typical hybridized focal plane array configuration.

Starting with high-quality InAs, the addition of relatively small amounts of other III-V elements is enough to increase the bandgap of the resulting quaternary into the e-SWIR spectrum shorter than 3.0 μm in two alloy systems, and along with that comes an expectation for a corresponding reduction in dark current. Similar e-SWIR alloy materials may be grown lattice matched to InAs substrates, with the advantage that their compositions may lie further from the regions of immiscibility than alloys of identical bandgap grown lattice matched to GaSb, thereby potentially exhibiting superior minority carrier transport. Among the potential difficulties with these materials are relatively high background carrier concentration related to high GaSb content and the tendency to phase separate, 16 which will both act to reduce minority carrier diffusion lengths and thereby detector quantum efficiencies. Devices with GaInAsSb 11- 13 and InPSb 14, 15 absorbers have been demonstrated but not fully characterized or optimized for the highest temperature operation. These have the potential advantages of improved transport and minority carrier lifetime in comparison to type-II superlattice materials. 3- 6 In the extended short-wave infrared (e-SWIR) region between 1.7 and 3.0 μm, recent demonstrations of antimonide-based detectors have been dominated by GaSb/InAs, 7 InAs/AlSb, 8 and more complex superlattice-absorber devices, 9, 10 which often exhibit dark currents substantially higher than described for the highest-quality traditional HgCdTe devices.Īlternatively, lattice-matched ternary or quaternary alloys with relatively wide bandgaps have been considered for e-SWIR absorbers grown on GaSb substrates. 1 In the short-wavelength infrared, materials grown on InP demonstrate excellent performance at wavelengths as long as 1.7 μm, 2 but have reduced performance at longer wavelengths due the relatively poor quality of the unconventional or lattice-mismatched materials required. High-performance mid- and long-wavelength infrared detectors have been demonstrated over the last decade using a variety of antimonide absorber materials, including InAs/InAsSb and GaSb/InAs type-II superlattices and metamorphic InAsSb alloys. For unoptimized InPAsSb devices with 2.55 μm cutoff, 200 K areal and perimeter dark current densities at -0.2 V bias in devices of various sizes were approximately 1x10 -7 A/cm 2 and 1.4x10 -8 A/cm, respectively. Dark currents for InPAsSb devices were lower than AlInAsSb devices with similar cutoff wavelengths. Despite the shallow-etch mesa nBn design, perimeter currents contributed significantly to the 200 K dark current. Top-illuminated devices with n+ InAs window/contact layers had external quantum efficiencies of 40-50% without anti-reflection coating at 50 mV reverse bias and wavelengths slightly shorter than cutoff. Devices having 4 μm thick absorbers exhibited sharp cutoff at wavelengths of 2.9 μm or longer and softer cutoff at shorter wavelengths. Minority-carrier lifetimes determined by microwave reflectance measurements were 0.2-1.0 μs in doped n-type absorber materials. We have fabricated and characterized AlInAsSb- and InPAsSb-absorber nBn infrared detectors with 200 K cutoff wavelengths from 2.55 to 3.25 μm.
