Experiments involving correlations between CdTe solar cell fabrication history and intrinsic device stability

An orthogonal full-factorial design was used to study the effect of CdS and CdTe layer thickness, oxygen ambient during vapor CdCl₂ (VCC) and the use of nitric-phosphoric (NP) acid as a pre-contact etch on the initial and stressed performance of CdS/CdTe small-area devices. The best initial device efficiency (using thinner CdS, thicker CdTe, no oxygen during VCC, and NP etch) also showed poor stability. Increasing the CdS thickness significantly improved stability with only a slight decrease in resulting initial performance. All devices used a thin margin of CdTe around the perimeter of the backcontact that was shown to significantly reduce catastrophic degradation and improve overall test statistics. The latter degradation is modeled by the formation of a weak-diode/low shunt resistance localized near the edge of finished devices. This shunting is believed to occur through the CdS/CdTe interface, rather than along the device edge, and is exacerbated by thinner CdS films and fabrication defects resulting in direct metal-CdTe contact.