Electromagnetic modeling of n-on-p HgCdTe back-illuminated infrared photodiode response

The mercury cadmium telluride (MCT) photodiode is a well-known detector for infrared (IR) sensing. Its growth (mainly liquid phase epitaxy (LPE)) and photovoltaic technology (ion implantation planar technology for instance) for second-generation IR detectors (linear and 2D monospectral arrays) now appear to be mature, well mastered, and understood, and allow optimal detection in a wide range of spectral bands. However, the next generation of IR detectors is supposed to use more sophisticated structures and technologies (such as mesa technology for dual-band detection or advanced heterostructures for high-operating-temperature detectors). Such structures are usually grown by molecular beam epitaxy (MBE) and consist of a layered stack of different thicknesses, HgCdTe (MCT) compositions, and doping levels. Moreover, pitches accessible today with advanced hybridization techniques (20 mu m or less) tend to approach the diffraction limit, especially for long-wave (LWIR) and very long-wave (VLWIR) devices. Hence, the physical understanding of these third-generation pixels from an electromagnetic (EM) point of view is not straightforward as it will have to take into account diffraction effects in the pixels. This paper will focus on EM simulation of advanced MCT detectors, using finite element modeling (FEM) to solve Maxwell's equations in a two-dimensional (2D) configuration and calculate absorption in the pixel. The corresponding collected current is then estimated by introducing a simple diffusion modeled diode and is compared to spot-scan experiments and/or experimental spectral responses to validate the method.