UV + Damp Heat Induced Power Losses in Fielded Utility N-Type Si PV Modules

Abstract

A recent trend in commercial PV modules is a transition to n-type silicon cells, including passivated emitter rear totally diffused (n-PERT), tunnel oxide passivated contact (TOPCon), and silicon heterojunction (SHJ). There is evidence via lab studies that some of these cells are more susceptible to UV induced degradation (UVID), yet there is a lack of confirmation that such degradation occurs in the field. Current IEC standards designed to screen for early module failures require only minimal UV exposure (15 kWh/m² 280–400 nm, ~2–3 months equivalent outdoor exposure). Here, we investigate fielded n-PERT silicon (Si) modules from a commercial utility that show power losses of ~2%/year. We present a comprehensive picture of the physics and chemistry of degradation supported by both module and cell electronic characterization (EL, PL, IV, EQE, and DLIT) and materials-level morphological and chemical analysis (SEM, EDS, XPS, FTIR, and HPLC). All sampled site modules show short circuit current (I_sc) and open circuit voltage (V_oc) losses when compared to unfielded spares, with the most severely degraded also having losses in fill factor (FF). We identify two different degradation modes contributing to overall power loss: (1) external quantum efficiency (EQE) measurements show losses in the blue range of the spectra, indicative of cell surface recombination losses, and (2) variations in high series resistance (R_s) at the cell level that are correlated with compositional differences in cell metallization. Using unfielded spares, we were able to reproduce V_oc, I_sc, and EQE losses via a minimum UV stress of 67.5 kWh/m² (280–400 nm), 4.5× the exposure currently required in IEC 61215-2 (MQT 10). Degradation continued with additional UV dosage equivalent to the fielded modules (405 kWh/m² total), with power loss leveling out at an average of 6.1%. Subsequent 1000 h of 85% RH/85°C damp heat testing showed that cells exposed to UV underwent additional severe series resistance degradation, even those without the susceptible paste composition seen in the field, whereas non-UV exposed cells saw little change. We attribute this to higher concentrations of acetic acid generated on the UV exposed area of the module, leading to degradation of the gridline/cell interface and high R_s. This study is unique in that it reproduces field observed utility scale UVID with an accelerated test and supports the need for standards development for longer UV exposure combined with other stress factors to catch materials interplay within a module package.