Firing stability of tube furnace‐annealed n‐type poly‐Si on oxide junctions

Stability of the passivation quality of poly‐Si on oxide junctions against the conventional mainstream high‐temperature screen‐print firing processes is highly desirable and also expected since the poly‐Si on oxide preparation occurs at higher temperatures and for longer durations than firing. We measure recombination current densities (J0) and interface state densities (Dit) of symmetrical samples with n‐type poly‐Si contacts before and after firing. Samples without a capping dielectric layer show a significant deterioration of the passivation quality during firing. The Dit values are (3 ± 0.2) × 1011 and (8 ± 2) × 1011 eV/cm2 when fired at 620°C and 900°C, respectively. The activation energy in an Arrhenius fit of Dit versus the firing temperature is 0.30 ± 0.03 eV. This indicates that thermally induced desorption of hydrogen from SiH bonds at the poly‐Si/SiOx interface is not the root cause of depassivation. Postfiring annealing at 425°C can improve the passivation again. Samples with SiNx capping layers show an increase in J0 up to about 100 fA/cm2 by firing, which can be attributed to blistering and is not reversed by annealing at 425°C. On the other hand, blistering does not occur in poly‐Si samples capped with AlOx layers or AlOx/SiNy stacks, and J0 values of 2–5 fA/cm2 can be achieved after firing. Those findings suggest that a combination of two effects might be the root cause of the increase in J0 and Dit: thermal stress at the SiOz interface during firing and blistering. Blistering is presumed to occur when the hydrogen concentration in the capping layers exceeds a certain level. Stable passivation quality of poly‐Si on oxide junctions against the mainstream high‐temperature screen‐print firing processes is highly desirable. By measuring recombination current densities and interface state densities of symmetrical samples with n‐type poly‐Si contacts before and after firing with varying capping layer stacks, firing conditions, and postfiring annealing steps, we hypothesize that a combination of two effects might be the cause for the observed decrease in passivation quality: Thermal stress at the SiOz interface during firing and blistering.