The trade-off between thick (~170 microns) silicon-based PV and thin (a few microns) film non-silicon and amorphous silicon PV is addressed by the development of single crystal silicon wafers of thicknesses of ~50 microns produced by epitaxy. This approach has the cost advantages of thin film technologies and the efficiency, reliability and non-toxicity of earth-abundant silicon PV.
Semiconductor technologies, of which photovoltaics is an increasingly large part, have had an obsession with dimensions and scaling over their history for increasing product functionality and reducing manufacturing costs. The most familiar and the most impactful is the ubiquitous scaling of gate dimensions in integrated circuits as described and predicted by Moore’s Law. This remarkable scaling of the critical dimensions in transistors in integrated circuits has been the driving force for the information, communication and entertainment industries. An additional important scaling in semiconductor manufacture has been silicon wafer size changes to improve the economics of manufacturing. Typically there have been wafer size (diameter) increases roughly every 10 years for about the last ~20 to 30 years, with the current leading-edge factories being based on 300 mm diameter wafers.
Scaling in photovoltaics has taken a different path. Clearly the analogy with semiconductor critical dimension scaling does not apply to photovoltaics where the device dimensions are equal to the wafer dimensions. In crystalline silicon photovoltaics, wafer size scaling is also not a very important factor, since very large wafers would lead to extremely unwieldy devices with large currents and low voltages limited by the band gap of silicon. Consequently, manufacturing cost reductions in photovoltaics are not dominated by wafer size changes.
The primary scaling factor in photovoltaics has been the thickness of the semiconductor or the absorber driven by manufacturing cost reasons, since materials constitute the major portion of manufacturing costs. There are, generally, two approaches to thickness scaling. The more fundamental one is based on the absorption coefficient of the semiconductor. Silicon, an indirect band gap semiconductor, has a low absorption coefficient for light, necessitating the use of fairly thick wafers (~180 microns today) to absorb a large portion of the sun’s spectrum that silicon is sensitive to and, more importantly, the difficulty of handling and processing silicon wafers of thicknesses much below current ~180 micron thickness. A response to the need to use thick silicon wafers has been the development of photovoltaic devices using direct band gap semiconductors such as amorphous silicon and various compound semiconductors with cadmium telluride (CdTe) and copper indium gallium selenide (CIGS) being the current favorites. These materials have superior absorption coefficients, enabling the use of very thin films of these materials and manufacturing technologies based on depositing thin films on rigid substrates such as glass or metal sheets.
With the current state of photovoltaic technology, there is a trade-off between reduced materials usage with thin film compound semiconductor (and amorphous silicon) photovoltaics and the relatively low energy conversion efficiency of these products. In contrast silicon wafer based devices, although consuming too much material are the undisputed leaders in achieved high energy conversion