Challenges in multiphysics modeling of dual-band HgCdTe infrared detectors.

We present three-dimensional simulations of HgCdTe-based focal plane arrays (FPAs) with two-color and dual-band sequential infrared pixels having realistic truncated-pyramid shape, taking into account also the presence of compositionally graded transition layers. After a validation against the spectral responsivity of two-color, mid-wavelength infrared detectors from the literature, the method is employed for a simulation campaign on dual-band, mid-, and long-wavelength infrared FPAs illuminated by a Gaussian beam. Simulation results underscore the importance of a full-wave approach to the electromagnetic problem, since multiple internal reflections due to metallizations and slanted sidewalls produce non-negligible features in the quantum efficiency spectra, especially in the long-wavelength band. Evaluations of the optical and diffusive contribution to inter-pixel crosstalk indicate the effectiveness of deep trenches to prevent diffusive crosstalk in both wavebands. In its present form, the detector seems to be subject to significant optical crosstalk in the long-wavelength infrared band, which could be addressed through pixel shape optimization.

Constraints and performance trade-offs in Auger-suppressed HgCdTe focal plane arrays.

Majority carrier depletion has been proposed as a method to suppress the dark current originating from quasi-neutral regions in HgCdTe infrared focal plane array detectors. However, a very low doping level is usually required for the absorber layer, a task quite difficult to achieve in realizations. In order to address this point, we performed combined electromagnetic and electric simulations of a planar $ 5 \times 5 $5×5 pixel miniarray with 5 µm wide square pixels, assessing the effect of the absorber thickness, its doping level in the interval $ {N_D}{ = [10^{14}}{,10^{15}}] \;{{\rm cm}^{ - 3}} $ND=[1014,1015]cm-3, and temperature in the interval 140 K-230 K, both in the dark and under illumination. Looking for a trade-off, we found that the path towards high-temperature operation has quite stringent requirements on the residual doping, whereas a reduction of the absorber thickness helps only moderately to reduce the dark current. Under illumination, interpixel cross talk is only slightly cut down by a decrease of temperature or absorber doping in the considered intervals, whereas it gets more effectively reduced by thinning the absorber.