Cadmium telluride (CdTe) solar cells are traditionally made on glass in a superstrate configuration, taking advantage of the transparency and mechanical rigidity of glass, as well as the ability to form the back contact last. However, there are advantages to fabricating solar cells on flexible substrates such as plastic sheets. Advantages include high specific powers (power generated per mass of the device) for low weight applications, the possibility for continuous roll-to-roll manufacturing suitable for industry, and the opportunity to place solar devices on objects and structures that are not necessarily flat.
We have fabricated various flexible CdTe devices through sputter deposition, using aluminum doped zinc oxide (AZO) as a transparent conducting oxide (TCO) top contact. Early devices, made by naively applying fabrication techniques used with glass cells, had poor performance. This inspired several investigations.
Traditional simple heuristic analysis of solar cell current-voltage curves assumes the cell is reasonably performing and that series and shunt resistance losses are small. This was not true for our early cells, inspiring us to derive a simple model for very poor cells.
Early cells showed significant losses due to both series and shunt resistance. Series resistance losses in flexible cells inspired us to investigate more deeply series resistance losses in general, resulting in a fairly simple model that relates fill factor losses, in dot cells made in a typical geometry, to TCO sheet resistance. This model was confirmed by experiment.
Series resistance losses also led to futher investigation of AZO properties. We have found that AZO sheet resistance increases during cell fabrication, but a model that is supported by experiment indicates that the resistivity change is limited to a finite thickness of material. This allows fabrication of AZO which remains of good sheet resistance throughout the fabrication process.
Shunt resistance losses inspired a study of highly resistive and transparent (HRT) zinc oxide made by reactive sputtering with oxygen, which can be used as a buffer for shunts. Data on resisitivity as a function of oxygen in the sputter gas is given when using doped and undoped zinc oxygen sputter targets. Calculations on theoretical effects of such buffer layers are also presented.
A high temperature cadmium chloride (CdCl2) treatment is needed to create high efficiency CdTe cells. This can be a difficult step for flexible cells, as it can lead to CdTe film delamination. We have optimized a treatment for flexible cells and found that conditions, different from those which are optimal for cells on glass, allows high performance with no delamination.
Understanding of these issues allowed progress in the performance of our cells fabricated on flexible substrates. Ultimately, we have fabricated a flexible CdTe cell with efficiency of 10.5%. This efficiency is, to our knowledge, the best of any flexible CdTe cell which uses zinc oxide as a transparent conducting oxide. The performance of these cells also compares favorably to the best overall flexible CdTe cells.