Premium
Microstructural Evolution of Mechanically Deformed Polycrystalline Silicon for Kerfless Photovoltaics
Author(s) -
Wu Min,
Murphy John D.,
Jiang Jun,
Wilshaw Peter R.,
Wilkinson Angus J.
Publication year - 2019
Publication title -
physica status solidi (a)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.532
H-Index - 104
eISSN - 1862-6319
pISSN - 1862-6300
DOI - 10.1002/pssa.201800578
Subject(s) - electron backscatter diffraction , materials science , recrystallization (geology) , annealing (glass) , dislocation , wafer , silicon , crystallite , grain boundary , composite material , crystallography , metallurgy , microstructure , optoelectronics , paleontology , chemistry , biology
Silicon wafers for photovoltaics could be produced without kerf loss by rolling, provided sufficient control of defects such as dislocations can be achieved. A study using mainly high resolution electron backscatter diffraction (HR‐EBSD) of the microstructural evolution of Siemens polycrystalline silicon feedstock during a series of processes designed to mimic high temperature rolling is reported here. The starting material is heavily textured and annealing at 1400 °C results in 90% recrystallization and a reduction in average geometrically necessary dislocation (GND) density from >10 14 to 10 13 m −2 . Subsequent compression at 1150 °C – analogous to rolling – produce sub‐grain boundaries seen as continuous curved high GND content linear features spanning grain interiors. Post‐deformation annealing at 1400 °C facilitates a secondary recrystallization process, resulting in large grains typically of 100 μm diameter. HR‐EBSD gives the final average GND density in as 3.2 × 10 12 m −2 . This value is considerably higher than the dislocation density of 5 × 10 10 m −2 from etch pit counting, so the discrepancy is investigated by direct comparison of GND maps and etch pit patterns. The GND map from HR‐EBSD gives erroneously high values at the method's noise floor (≈10 12 m −2 ) in regions with low dislocation densities.