Theoretical and experimental investigation of MWIR HgCdTe nBn detectors

Abstract

We present in this study a theoretical and experimental investigation of the MWIR HgCdTe nBn device concept. Theoretical work has demonstrated that the HgCdTe nBn device is potentially capable of achieving performance equivalent to the ideal double layer planar heterostructure (DLPH) detector. Comparable responsivity, low current denisty ‫ܬ‬ ୢୟ୰୩ , and high detectivity ‫ܦ‬ ‫כ‬ values rival those of the DLPH device without requiring p-type doping. The theoretical results suggests that the HgCdTe nBn structure may be a promising solution for achieving a simplified MWIR device structure and addressing problems associated with reducing thermal generation in conventional p-on-n structures and processing technology limitations such as achieving low, controllable in-situ p-type doping with MBE growth techniques. Furthermore, the physical mechanisms for selective carrier conduction in the nBn structure may provide a basis to incorporate into future device structures to suppress intrinsic Auger carrier generation. Likewise, the experimental demonstration of the MWIR HgCdTe nBn devices introduces a promising potential alternative to conventional high performance p-n junction HgCdTe photodiodes. The experiments described in this study illustrate the successful implementation of a HgCdTe barrier-integrated structure. The measured current-voltage characteristics of planar-mesa and mesa HgCdTe nBn devices exhibit barrier-influenced behavior and follow temperature-dependent trends as predicted by numerical simulations. Optical measurements of the planar-mesa MWIR HgCdTe nBn device indicate a bias-dependent spectral response. Further changes to MWIR HgCdTe nBn layer structure has shown an over 10 5 A/cm 2 reduction in J dark as well as a shift to a lower turn-on operation bias. This experimental investigation highlights the potential for pursuing similar and related unipolar, type-I barrier devices for high performance infrared detector applications.

Theoretical Modeling of HOT HgCdTe Barrier Detectors for the Mid-Wave Infrared Range

“…14 and 15, the optimal composition to attain the highest D * for both nB n n and nB n p strictly increases with applied voltage (e.g., for V = 0.2 one has x = 0.54, while for V = 0.3 V one has x = 0.575). The simulated structures reach D * = 1.8 9 10 9 cmHz 1/2 /W (nB n n) and 1.2 9 10 9 cmHz 1/2 /W (nB n p) for V = 0.3 V and T = 200 K, comparable to results presented by Velicu et al 17 For nB n n(p)n + structures, the detectivity rises with the barrier composition, exhibiting values nearly two orders of magnitude higher (D * = 7 9 10 10 cmHz 1/2 /W for V = 0.1 V) compared with the nB n n(p) detectors, being comparable to the results presented by Itsuno et al 23 As mentioned, the highest D * is obtained for V = 0.1 V, which was assumed to be the turn-on voltage limit for J DARK [nB n n(p)n + structures].Figures 16 and 17 present D * versus absorber doping for selected voltages, simulated with and without the BTB/TAT mechanism at the n-n + and p-n + junction. The presented results indicate that correct optimization of both junctions allows D * to be increased to the level of 9.5 9 10 10 cmHz 1/2 /W and 1.5 9 10 11 cmHz 1/2 /W for the nB n nn + and nB n pn + structures, respectively.…”

“…The modeled nB n n and nB n nn + structures are shown in Fig. 1a-c. A similar BIRD design was reported by Velicu and Itsuno et al 17,23 The simulated BIRD structure consists of an n + -type contact layer (exclusion junction) and n-type HgCdTe absorber with thickness of 5 lm doped with In (n = 10 14 cm À3 ) with composition x = 0.275 for the MWIR range. After the absorber layer, an n-type HgCdTe barrier was incorporated with thickness of 0.15 lm, doped with In (n = 2 9 10 15 cm À3 ).…”

“…Figure 18 shows the simulated J DARK versus the operating temperature for the MWIR BIRD HgCdTe structures (k c = 5.2 lm, T = 200 K) in comparison with experimental dark current densities for the following detectors: InAs/GaSb with AlGaSb barrier T2SL nB p n (k c = 5.4 lm, T = 230 K), InAsSb with AlAsSb barrier (k c = 5.05 lm, T = 200 K), and the HOT HgCdTe nB n n detector (x = 0.3). 17,[28][29][30] The particular significance of the incorporation of the extra barrier for carriers (n + ) in the BIRD nB n nn + structures versus the single barrier (potential majority-carrier blocking) is clearly evident from the J DARK decrease from 3 A/cm 2 to 7 9 10 À3 A/cm 2 In addition, having taken the difference in the absorber's composition into consideration, we may assume that the presented results coincide with these published by Velicu et al 17,28 Comparing the nB n n and nB n p detectors, it is clearly evident that the SRH contribution is totally suppressed for the n-type absorber structures (diffusion limited), while the p-type absorber architecture exhibits a twoslope behavior with a crossover temperature estimated at the level of T c = 227 K. Both nB n nn + and nB n pn + detectors are diffusion limited (one-slope behavior). Figure 19 compares the R 0 A (k B T/q/J s ; V = 1 mV for BIRD structures) product for nB n n, nB n nn + , and nB n nN + (the N + layer, similarly to the barrier, consists of two sublayers-the very first one fitted with a graded composition to the absorber and the second with x = 0.4) versus the values given by ''Rule 07,'' being a simple means to compare mercury cadmium telluride (MCT) IR detectors.…”

“…Figure 18 shows the simulated J DARK versus the operating temperature for the MWIR BIRD HgCdTe structures (k c = 5.2 lm, T = 200 K) in comparison with experimental dark current densities for the following detectors: InAs/GaSb with AlGaSb barrier T2SL nB p n (k c = 5.4 lm, T = 230 K), InAsSb with AlAsSb barrier (k c = 5.05 lm, T = 200 K), and the HOT HgCdTe nB n n detector (x = 0.3). 17,[28][29][30] The particular significance of the incorporation of the extra barrier for carriers (n + ) in the BIRD nB n nn + structures versus the single barrier (potential majority-carrier blocking) is clearly evident from the J DARK decrease from 3 A/cm 2 to 7 9 10 À3 A/cm 2 and 40 A/cm 2 to 3 A/cm 2 for T = 200 K and T = 300 K, respectively. Proper contact layer arrangement increases the operating temperature by close to 75 K for V = 0.4 V. The absorber's p-type doping (assuming the same level of doping N D = N A = 10 14 cm À3 ) leads to a further decrease of J DARK to the level of 4 9 10 À3 A/cm 2 and 2 A/cm 2 for T = 200 K and T = 300 K, respectively.…”

“…15 Even though HgCdTe does not exhibit a zero valence band offset (''nested'' type I heterojunction-it is impossible to design an entirely majority-carrier blocking barrier device), the BIRD (nB n n) architecture was successfully adapted to the HgCdTe alloy, presenting technological advantages over the p-n HgCdTe homojunction [simplifying the fabrication process and circumventing the potential problems with p-type in situ doping in molecular beam epitaxy (MBE) growth]. [16][17][18] On the other hand, the p-type absorber offers a better absorption coefficient and lower Auger GR, which should improve the detector performance. 19 In addition, Klem et al 20 and Klipstein 5 reported on pB n n InAsSb/AlAsSb detector performance, pointing out their possible capabilities when a very large barrier in the conduction band is not available, allowing for operation at zero bias.…”