Indium arsenide/gallium antimonide type-II strained-layer superlattice infrared photodetectors.

Abstract

InAs/GaSb type-II strained-layer superlattice (SLS) infrared detectors are currently under intensive research. The technique of reducing detector dark current by inserting resonant tunneling barriers into a normal InAs/GaSb SLS was investigated. A tunneling double barrier (25 monolayer (ML) GaSb/11ML InAs/25ML GaSb) was designed to be inserted into a normal 8ML 1nAs/8ML GaSb SLS to block thermally generated electrons, while permitting photo-generated electrons to pass through. Through fabrication and characterization of the tunneling and normal InAs/GaSb SLS detectors, this work demonstrates that the tunneling barriers suppress dark current more effectively than photocurrent: the current responsivity of the tunneling InAs/GaSb SLS detector measured at 84 K was 27% lower than that of the normal SLS detector, however, the measured dark current density of the tunneling InAs/GaSb SLS detector was reduced by a factor of 3.8 at 77K. Both types of InAs/GaSb SLS detectors demonstrated room-temperature operation. The Johnson-noise-limited detectivity (measured at 4 mum) of the tunneling SLS detector was 18% higher than that of the normal SLS detector. Improving carrier lifetimes is key in the development of InAs/GaSb SLS detectors. To measure InAs/GaSb SLS carrier lifetimes, photo-induced open-circuit voltage decay (PVD) and picosecond excitation correlation (PEC) measurements were performed at 77 K. We developed a general theoretical model that is capable of simulating the PEC signals obtained from any bulk semiconductor. The radiative and non-radiative recombination carrier lifetimes (&tgr;<italic>_r</italic> and &tgr;<italic>_nr</italic>) of as-grown 8ML InAs/8ML GaSb SLSs extracted using this model were in the range of 18--175 ns and 2.3--2.4 ns, respectively, indicating the dominance of nonradiative carrier recombinations. From PEC experiments, we extracted &tgr;<italic>_nr</italic> of InAs/GaSb SLSs (with etched mesas) passivated in ammonium sulfide solutions under various conditions. The best passivation condition found in our study was dipping the sample in (NH4)₂S 21%:H₂0=1:4 for two hours, which increased the &tgr;<italic>_nr</italic> from 1.3--1.4 ns to 3.5--3.8 ns. The 77 K surface recombination velocity of unpassivated and ammonium sulfide passivated ((NH₄)₂S 21 %:H₂0=1:4 for 30 minutes) InAs/GaSb SLSs were determined by variable-area diode array experiments to be 9.5x10<super>5</super> cm/s and 4.0x10<super>5</super> cm/s, respectively. At 77 K the typical measured values of zero-bias dynamic resistance-area product (<italic>R₀A</italic>) in the passivated devices were 2--2000 O cm<super> 2</super> versus 0.05--0.2 O cm<super>2</super> for the unpassivated diodes. There was no change of the <italic>R₀A</italic> values of ammonium sulfide passivated InAs/GaSb SLS diodes measured 24 hours, 45 days and 120 days after the sulfidation, indicating long-term stability of the aqueous ammonium sulfide passivation.