Optimization of thin film flexible solar cell production process

Abstract

The quest for a cleaner and greener fuel has necessitated the development of renewable energy sources. Sun has been a source of energy for human civilization since agrarian times. With the recent developments in science and technology, Solar energy has evolved as the front runner to power the future energy needs. The plentiful, eco-friendly, and abundant solar energy resource has been evolved from the first generation solar comprising of crystalline silicon based solar panels to second generation photovoltaic cells.

This work is based on the scope of the FlamingoPV project which is the collaboration between TU Delft and HyET Solar, Netherlands, to realize large scale commercialization of second generation thin film flexible solar cells. The second generation solar cells offer flexibility, economic viability, and ease of manufacturing. To reach the desired goals of roll-to-roll production of a-Si:H/nc-Si:H and a-Si:H/nc-Si:H/nc-Si:H with stabilized efficiencies of 13\% and 14\% respectively, various production processes need to be optimized.

The current production line at HyET Solar faces losses in performance due to high open circuit resistance and low parallel (shunt) resistance. Moreover, the lab processing route needs to be optimized to improve the yield of the samples produced. The sputtering process of depositing the back contact is considered as one of the potential causes of low shunt resistances in the lab samples. Hence, the sputtering process is optimized for electrical and optical properties. The effect of sputtering pressure and RF power is understood to develop dense and compact films for the Aluminium doped Zinc Oxide (AZO) films used in the back contact of the solar cell stack. It is found that a lower sputtering pressure provides better optical and electrical properties. Furthermore, optical simulation from GenPro4 suggest an increase in intrinsic layer absorption under specific conditions of AZO deposition.

Furthermore, the monolithic series interconnection process at HyET Solar is a bottleneck in the lab processing route. Additionally, the laser power used for scribing is not optimized for tandem and triple junction solar cells. To ensure faster learning cycles for understanding the quality of the silicon layers deposited a novel way of processing solar cells is developed. This approach helps to distinguish various losses generated in the solar cell and ensures quality control in the production line. Finally, a series of triple junction solar cells deposited on the aluminum substrate is processed without laser scribing to demonstrate the possibilities of the process.

The PECVD tool at HyET Solar needs to be optimized to deposit good quality bottom cell for the micromorph solar cells. The continuous variable thickness series have been used to understand Lambert Beer relation for short circuit current density and deposition thickness. The ASA software is used to simulate case studies for troubleshooting deposition of the bottom cell. The results from the electrical simulation are used to understand the problems with the silane dilution ratios in the previous PECVD runs. Furthermore, the results of the simulation suggest the sensitivity of the process to ensure good quality nanocrystalline silicon deposition. It is observed that the first layer of growth is very critical for the nanocrystalline silicon deposition while after a certain thickness of nanocrystalline silicon inclusion of amorphous material does not affect the properties of the absorber layer.

Overall, the production process has been optimized from the lab scale to the industrial scale. One of the causes of low performance in solar cells has been addressed. The bottleneck in the production of lab samples has been addressed ensuring good quality material for faster learning cycles. The deposition conditions for the growth of nanocrystalline silicon have been analyzed. This can help in troubleshooting the production process and maintaining a high-quality deposition process.