Measurement of band offsets and shunt resistance in CdTe solar cells through temperature and intensity dependence of open circuit voltage and photoluminescence. (1st September 2019)

• Record Type: Journal Article
• Authors: Swartz, Craig H.; Rab, Sadia R.; Paul, Sanjoy; van Hest, Maikel F.A.M.; Dou, Benjia; Luther, Joseph M.; Pach, Gregory F.; Grice, Corey R.; Li, Dengbing; Bista, Sandip S.; LeBlanc, Elizabeth G.; Reese, Matthew O.; Holtz, Mark W.; Myers, Thomas H.; Yan, Yanfa; Li, Jian V.
• Is Part Of: Solar energy. Volume 189(2019)
• Page Start: 389
• Page End: 397
• Publication Date: 2019-09-01
• Subjects: Photovoltaic device characterization -- Loss analysis -- Model fitting -- Contact barrier -- Recombination -- Injection dependence; Solar energy -- Periodicals; Solar engines -- Periodicals; 621.47
• DOI: 10.1016/j.solener.2019.07.057
• Languages: English
• ISSNs: 0038-092X

Abstract

Highlights: An analysis of temperature & injection dependent Voc and PL intensity is developed. CdTe-based solar cells are grown with an assortment of windows and back contacts. Band offset, shunt, and recombination effects are separated. Standard J-V shunt resistances must be well below a kΩ cm 2 to be meaningful. Abstract: Band offsets at the back contact and front window layer in CdTe-based solar cells affect photovoltaic performance and challenge standard characterization methods. By analyzing the temperature and excitation dependence of both open circuit voltage and absolute photoluminescence intensity, we show that the effects band offsets can be separated from the effects of recombination and shunting. Solar cells were grown with MgZnO window layers and compared to cells with CdS window layers containing varying amounts of oxygen. It was demonstrated that band alignment rather than reduced recombination velocity is the reason for the success of MgZnO as a front interface contact. An assortment of thin back contact interlayers were also deposited, and a PbTe interlayer showed some promise as an Ohmic contact to the CdTe, though it appears to induce a photoconductive shunt. Finally, we show that the shunting resistance given by a standard current-voltage curve technique generally does not represent a physically meaningful quantity unless it is well below one kiloOhm square cm.