MoS 2 Interface Engineering for Lead-Free RbGeI 3 Perovskite Photodetectors: SCAPS-1D and Machine Learning Optimization

Interface engineering represents an effective strategy for passivating interfacial defects and suppressing dark current in photodetectors. Numerical simulations of an ITO/TiO 2 /MoS 2 /RbGeI 3 /Cu 2 O/Au n-i-p heterojunction demonstrate the efficacy of a MoS 2 interlayer in enhancing device performance. Systematic optimization of the MoS 2 intrinsic parameters and associated interfacial defect properties is performed. This optimization leads to a substantial reduction in dark current density at −0.6 V, from approximately 3 × 10 −7 to 1 × 10 −9 mA cm −2 . Consequently, the specific detectivity at 600 nm increases from 4.5 × 10 13 Jones to 9.1 × 10 14 Jones, while external quantum efficiency and responsivity remain stable. Furthermore, machine learning models reveal that the TiO 2 /MoS 2 interfacial defect density is the primary contributor to variations in dark current, whereas the MoS 2 /RbGeI 3 interfacial defect density most strongly affects the external quantum efficiency. The impact of other parameters is comparatively minor. These findings provide a clear guideline for optimizing lead-free perovskite photodetectors incorporating functional interlayers.