Analysis of the auger recombination rate in P+N−n−N−N HgCdTe detectors for HOT applications

Abstract

Infrared (IR) photon detectors must be cryogenically cooled to provide the highest possible performance, usually to temperatures at or below ~ 150K. Such low operating temperatures (Top) impose very stringent requirements on cryogenic coolers. As such, there is a constant push in the industry to engineer new detector architectures that operate at higher temperatures, so called higher operating temperature (HOT) detectors. The ultimate goal for HOT detectors is room temperature operation. While this is not currently possibly for photon detectors, significant increases in Top are nonetheless beneficial in terms of reduced size, weight, power and cost (SWAP-C). The most common HgCdTe IR detector architecture is the P+n heterostructure photodiode (where a capital letter indicates a wide band gap relative to the active layer or “AL”). A variant of this architecture, the P+N−n−N−N heterostructure photodiode, should have a near identical photo-response to the P+n heterostructure, but with significantly lower dark diffusion current. The P+N−n−N−N heterostructure utilizes a very low doped AL, surrounded on both sides by wide-gap layers. The low doping in the AL, allows the AL to be fully depleted, which drastically reduces the Auger recombination rate in that layer. Minimizing the Auger recombination rate reduces the intrinsic dark diffusion current, thereby increasing Top. Note when we use the term “recombination rate” for photodiodes, we are actually referring to the net generation and recombination of minority carriers (and corresponding dark currents) by the Auger process. For these benefits to be realized, these devices must be intrinsically limited and well passivated. The focus of this proceeding is on studying the fundamental physics of the intrinsic dark currents in ideal P+N−n−N−N heterostructures, namely Auger recombination. Due to the complexity of these devices, specifically the presence of multiple heterojunctions, numerical device modeling techniques must be utilized to predict and understand the device operation, as analytical models do not exist for heterojunction devices.