Three-dimensional numerical simulation of planar P⁺n heterojunction In_0.53Ga_0.47As photodiodes in dense arrays part II: modulation transfer function modeling

Abstract

Processing improvements have facilitated manufacturing reduced pixel dimensions for lattice-matched InGaAs on InP short-wave infrared detectors. Due to its technological maturity, this material system continues to garner attention for low-light level imaging applications. With pixel dimensions smaller than minority carrier diffusion lengths, optimizing array performance by reducing crosstalk from lateral carrier diffusion remains an important design issue. Analytical models, however, have provided limited insight on underlying mechanisms limiting device performance in the conventional planar double heterointerface device. Quantitative modeling provides tools to investigate performance sensitivities and their underlying mechanisms. In this work we develop a three-dimensional numerical simulation for dense P + n In 0.53 Ga 0.47 As on InP photo detector focal plane arrays using a conventional planar, back-illuminated structure.We evaluate optical generation with finite-difference time-domain analysis, and model carrier transport in a driftdiffusion analysis simultaneously solving the carrier continuity and Poisson equations. Using this model we investigate modulation transfer function variations with pixel pitch and diffused junction geometries for small dimension arrays. By accounting for carrier diffusion effects, these results should provide a benchmark against which to evaluate modulation transfer function contributions from other effects, such as crosstalk attributable to photon recycling.

Dense Array Effects in SWIR HgCdTe Photodetecting Arrays

“…Figure 8 shows the dark current I-V behavior for the central diode in p-type 5 9 5 miniarrays, and a single epitaxial diode in 50-lm 2 bulk epilayers, with N A = 5 9 10 15 cm À3 at 300 K. As expected, the dark current falls with pitch in both n-and p-type arrays, a result attributable to lateral diffusion current suppression under the dense array boundary condition (no net lateral diffusion at symmetry planes between adjacent diodes). 2 As shown in Fig. 7, the effect increases in strength as the lateral diode separation falls relative to the minority-carrier diffusion length.…”

“…7, the effect increases in strength as the lateral diode separation falls relative to the minority-carrier diffusion length. 2 In the p-type arrays, however, all three pitch cases are well below ) 7.66 9 10 13 5.11 9 10 12 l n (cm 2 …”

“…2 The array is dense when both the absorber thickness, as well as the radial distance from the lateral edge of a depletion region around a reverse-biased diode to the midpoint between that diode and a neighbor, are smaller than the minority-carrier diffusion length. 2 To simulate dense arrays, junctions were arrayed in rectangular fashion with center-to-center junction placement (pitch) of 10 lm, 15 lm, and 20 lm. The center pixel in 3 9 3 miniarrays is taken as representative of a dense array pixel.…”

“…2,8 Because the Shockley-ReadHall recombination is not fundamentally limiting, in these simulations we consider only radiative and Auger recombination.…”

Dark Currents in a Fully-Depleted LWIR HgCdTe P-on-n Heterojunction: Analytical and Numerical Simulations

“…3 The model can also accommodate larger arrays at any pixel pitch or bias configuration completely capturing dense array effects. 18 Since the focus of this work is on studying effects along the growth direction (vertical QNR and SCR currents) and not lateral/3D effects, the simulations in this manuscript were conducted using a 2D rather than a 3D numerical simulation. Reducing the dimensionality from 3D to 2D immensely reduces the computational complexity, making the process of mapping the parameter space of pertinent sensitivity parameters far more tractable.…”