SPECTRAL–ENVIRONMENTAL PERFORMANCE MODELING AND MULTI-SCENARIO DESIGN OF A SELF-POWERED ZNO/SI ULTRAVIOLET PHOTODETECTOR

Dust, humidity, and temperature act at the same time in service, yet their combined effect on self-powered ZnO/Si ultraviolet detectors is rarely evaluated. A reduced-order model was therefore assembled for an n-ZnO/p-Si device. It links wavelength-dependent ZnO absorption and thickness-dependent carrier collection with an Ångström soiling term, humidity-dependent surface transport, thermally activated dark current, shot and Johnson noise, and adsorption-controlled transients. Calibration used one quantity only: the clean zero-bias responsivity at 365 nm. With the adopted idealized noise expression, D* was 4.10 × 10¹³ Jones, compared with 3.5 × 10¹³ Jones in the reference device (17.2% higher). This comparison was used as a plausibility check and was not a fitting target. Performance retention was summarized by a weighted geometric Reliability Index (RI). The design was then tested by changing the RI weights and passivation penalty, refining the numerical grid, and checking Monte Carlo convergence. The practical grid selected 225 nm and the refined grid 226 nm, giving a conditional thickness region of 220–235 nm. By contrast, the preferred passivation index changed with the assumed coating penalty. For the harsh case (0.35 mg cm⁻² dust, 85% RH, and 325 K), optimization raised RI from 0.255 to 0.483. In 12,000 conditional uncertainty realizations, the median was 0.738 and the 5th–95th percentile range was 0.671–0.806. The model can be used to screen designs before fabrication, but it does not qualify device reliability. Measurements are still needed to identify the coating and material parameters.