To overcome the limitations of CdZnTe substrates for large-format, low-cost HgCdTe infrared focal plane arrays (IRFPAs), the epitaxial growth of HgCdTe films on alternative substrates (e.g., GaAs and Si) has become an important research focus. The lattice mismatch of approximately 14% between the GaAs alternative substrate and the HgCdTe material generates a high density of interfacial defects, such as dislocations and twins. These defects induce a high density of interface states within the near-interface bandgap, resulting in interfacial recombination and consequently limiting device performance. This paper proposes an optimization method for the HgCdTe/GaAs interface that involves substrate removal and surface passivation after the fabrication of GaAs-based HgCdTe infrared (IR) detectors. The GaAs substrate was removed without damage through chemical mechanical polishing (CMP) and selective wet chemical etching. A bromine-based solution (Br2–HBr) was employed to eliminate the surface damage layer for interfacial optimization, and a composite dielectric film was deposited to achieve simultaneous surface passivation and optical antireflection. Experimental results on n-on-p devices operating at 80 K demonstrate that after interfacial optimization, the average quantum efficiency across the 3.5–6.1 μm wavelength range increased from 58% to 84% and the blackbody responsivity improved from 8.7 × 106 V/W to 1.6 × 107 V/W. Both quantum efficiency and blackbody responsivity reached levels comparable to those of CdZnTe-based detectors. Numerical fitting based on the carrier diffusion model indicated that interfacial optimization reduced the surface potential by approximately two orders of magnitude, effectively suppressing interfacial recombination.
Li et al. (Tue,) studied this question.