2020
In this study, we investigated how the glass beneath an aluminium-doped zinc oxide film affects its behaviour in heat and humidity. ZnO:Al is a transparent conducting material used in thin-film solar cells and other optoelectronic devices. It is attractive because it is abundant, non-toxic and can be deposited at relatively low temperatures. But if the transparent front contact loses conductivity, the performance and lifetime of the whole device are at risk.
We compared ZnO:Al films sputtered onto two kinds of glass: sodium-poor borosilicate glass and sodium-rich soda-lime glass. Both sets were exposed for up to 1,032 hours at 85 °C and 85 percent relative humidity, the demanding conditions used in the IEC 61215 photovoltaic reliability test. Along the way, we used electrical, optical, chemical and microscopic measurements to follow what was changing.
Both types of film degraded. Charge-carrier mobility fell and sheet resistance rose, while carrier concentration and optical behaviour changed less. The resistance increase followed a square-root relationship with exposure time and was mainly caused by lower mobility. In the middle of the films, there was no large difference in conductivity loss between sodium-rich and sodium-poor substrates. That challenged the simple assumption that sodium content alone determines damp-heat stability.
At the edges, however, we saw a clear difference. Both film types developed small surface spots. On soda-lime glass, broad whitened edges also appeared as fractal-like structures rising above the surface. These areas were rich in carbon, while the edge structures contained particularly high sodium levels. Our chemical observations point to the role of water and migrating species, and they show that film edges deserve separate attention as a reliability issue.
For us, the practical lesson is that durable solar devices depend on the entire material stack, its interfaces and the sealing strategy—not only on the active solar layer. A water barrier may help limit mechanisms like those we observed. By separating broad changes in conductivity from local edge effects, we provide a more precise basis for designing stable thin-film photovoltaic modules and other transparent-electrode devices.