Impact of growth temperature on InAs/GaInSb strained layer superlattices for very long wavelength infrared detection

Abstract

We explore the optimum growth space for a 47.0 Å InAs/21.5 Å Ga 0.75 In 0.25 Sb superlattices (SLs) designed for the maximum Auger suppression for a very long wavelength infrared gap. Our growth process produces a consistent gap of 50 6 5 meV. However, SL quality is sensitive to the growth temperature (T g ). For the SLs grown at 390À470 C, a photoresponse signal gradually increases as T g increases from 400 to 440 C. Outside this temperature window, the SL quality deteriorates very rapidly. All SLs were n-type with mobility of $10 000 V/cm 2 and 300 K recombination lifetime of $70 ns for an optimized SL. V C 2012 American Institute of Physics.[http://dx.doi.org/10.1063/1.4764015] An alternative infrared material system proposed by Smith and Mailhiot 1 uses the concept of broken-gap type-II band alignment of the InAs/GaInSb strained layer superlattices (SLs) to achieve narrow gaps for the very long wavelength infrared (VLWIR) detection (>14 lm). By alloying indium to the GaSb layer, the lattice constant of the GaInSb layer increases. The biaxial tension in the InAs layer lowers the conduction band, while the biaxial compression in the GaInAs layer raises the heavy-hole (HH) band. As a result, the very narrow gap can be achieved without sacrificing optical absorption. In addition, an intentionally introduced strain can create a large splitting between the HH and light-hole bands in the p-type SLs; this situation can prevent the holehole Auger recombination process, therefore the Auger limited minority carrier lifetime, detectivity, and the background limited operating temperature can be significantly improved. Grein et al.2 demonstrated how a small variation in the strain splitting between the uppermost valence bands can create an order of magnitude differences on their calculated detectivities. For the SLs designed for the same 80 meV gap at 40 K, the electronic band structure of either 49.7 Å InAs/57.0 Å Ga 0.9 In 0.1 Sb or 47.0 Å InAs/21. 5 Å Ga 0.75 In 0.25 Sb SL designs computed with the interface terms had the total lifetime of either 5 Â 10 À9 or 1.4 Â 10 À7 s. This difference leads the device detectivity to be either 5.2 Â 10 13 or 6.0 Â 10 14 cmHz 1 =2 /W. Although the greater strain splitting generates the better detectivity, one has to limit the indium alloy composition below 30% due to difficulties in strain balancing above this percentage. The authors also computed device detectivity for the more realistic case of 82 meV SL detectors with a 35 ns Shockley-Read-Hall (SRH) lifetime.3 Unexpectedly, Auger processes were fast enough to overrule SRH processes and the detectors still can operate at reasonably high temperatures, roughly exceeding 150 K.3 Therefore, strained InAs/GaInSb SL system appeared to be an excellent choice for the VLWIR detection. Unfortunately, the VLWIR InAs/GInSb material growth studies are still in immature stage of development. Many crystal growers tend to believe that the existence of an alloy layer can create a higher degree of disorder caused by alloy scattering, indium segre...

Optimum growth window for InAs/GaInSb superlattice materials tailored for very long wavelength infrared detection

“…Growth rates of 0.3 and 1.6 Å /s were used for InAs and GaInSb layers, respectively. Since monomeric Sb is the species critical to reducing densities of nonradiative recombination centers in the SL material, 19 we set the Sb cracking zone temperature at 950 C for the series, which is close to the suggested value by the EPI Model 200 cc Mark V Corrosive Series Valved Cracker, and the As cracking zone temperature at 900 C. 20 The interface type between layers was not intentionally controlled in order to compensate the residual strain in the ternary SLs. However, the net residual strain of our series was small and varied between þ0.2% and 0.0%.…”

Quantitative analysis of strain distribution in InAs/InAs1−xSbx superlattices

“…It is of interest to compare the above results with those reported in similar studies on InAs/In x Ga 1Àx Sb superlattices. 12,13 It is well known that strain localization at interfaces and the need to tailor interface composition plays an important role in Ga-based superlattices, due to the formation of Ga-As (tensile) and In-Sb (compressive) bonds at the interface. In the present study, however, the need for interface composition control is alleviated due to the absence of Ga in these superlattices.…”

“…Atomic scale strain distribution studies have been reported recently on InAs/In x Ga 1Àx Sb superlattices, which revealed the nature of strain localization at interfaces and surface segregation within GaSb layers. 12,13 There is, however, a lack of similar studies on InAs/InAs 1Àx Sb x superlattices. The objective of the present communication is to apply recent advancements in high-resolution transmission electron microscopy (HRTEM) to examine the nature of strain distribution in InAs/InAs 1Àx Sb x superlattices.…”

Control of anion incorporation in the molecular beam epitaxy of ternary antimonide superlattices for very long wavelength infrared detection

“…[3] proposed a strain balanced VLWIR ternary SL of 47.0 Å InAs/21.5 Å Ga 0.75 In 0.25 Sb, which is the design used in our studies. Haugan et al [4,5] have shown improvements in the quality of the ternary materials produced using molecular beam epitaxy (MBE) growth process for this design. Their optimized SL materials produced a strong photoresponse signal, a high mobility, and a long 300 K carrier lifetime.…”