Apparatus for Optimizing Photovoltaic Solar Manufacturing Efficiency through Real-Time Process Feedback and Spectral Binning of Cells

Novel, inline measurement techniques are needed to improve the quality and performance of new solar cells. Manufacturers today measure cell performance inline under white light conditions, but are unable to obtain the cell response as a function of wavelength without extensive offline testing. They are also unable to extract detailed spatial maps of full-spectrum spectral response. This information is fundamental to the device performance and, as device complexity increases it becomes even more important for successful process control, yield management, and module power optimization. This project helps to support the nations long-term energy goal of building higher performance, consistently robust solar devices at a lower cost to the end user. Our current phase II project involves developing the individual components needed for a new class of non-contact solar cell metrology, and will integrate them, by the end of phase II, into a fully functioning prototype at a major U.S. manufacturer. In phases I and II, we developed prototype non-contact sensors and integrated them with an advanced broadband light source. The light source has 64 colors, individually modulated, to simultaneously stimulate cell response from the ultraviolet to infrared. The response is detected via Fourier analysis in one of three ways: 1) capacitive, non-contact sensors, 2) inductive non-contact sensors and 3) conventional metal contacts. The full-spectrum measurement is completed in one second, and thus is fast enough to be used as an inline process monitor. The non-contact methods show good correlation with the contacting method, and are able to detect subtle shifts in photoresponse long before the device is completed. The resulting system has been tested on both conventional and thin film cells, as well as high performance interdigitated back contact cells. Various sensor configurations were developed and used to measure as early as the emitter formation step, and as late as a fully encapsulated module. In phase IIB, we will develop a new luminescence imaging technique and combine it with the Fourier illumination system developed in phase II to achieve another >1,000x speed improvement. The technique is based on PLE- Photoluminescence Excitation- and will allow high speed imaging of module spectral response both indoors and out. The expected commercial applications are: 1) outdoor module scanner for on-site failure analysis, 2) inline cell spectral sorting and process control, and 3) laboratory failure analysis to determine root cause power loss.