The authors report on an investigation of the dependence of the minority carrier lifetime in midwave infrared InAs/GaSb strained layer superlattices on a number of varied parameters: layer placement of two dopants (either Be or Te), and interface treatment between InAs and GaSb layers. In samples where the dopant and doping location was varied, it was found that the nonintentionally doped control sample exhibited the longest lifetimes (∼49 ns at 77 K under low injection), followed by the Be-doped and the Te-doped samples. Regardless of the type of doping, samples with dopants in only the InAs layer appeared to have longer lifetimes [low injection: 15 ns (Be), <3 ns (Te); high injection: 38 ns (Be), 16.2 ns (Te) at 77 K] compared to samples with dopants in the GaSb layer or all layers. However, because trap saturation behavior was observed in the transient photoluminescence (PL) decay, the intensity-dependent PL lifetime is a function of both the minority and majority carrier lifetimes, complicating the interpretation of the data. In samples where the treatment of the InAs/GaSb interface was varied, the sample that demonstrated the longest lifetime had a one-period growth sequence of InAs, an Sb soak, GaSb, and an InSb strain compensation layer. Of the three interface samples investigated, the sample (with a growth sequence of InAs, an Sb soak, GaSb, and a growth interrupt) that demonstrated the shortest lifetime also exhibited a fast initial decay for all injection levels, at only 110 and 150 K. This fast initial decay has been attributed to the appearance of another Shockley–Read–Hall trap level, contributing to a shorter carrier lifetime.
Read full abstract