| Summary: | <p>Approximately 50% of solar cells are based on multicrystalline silicon (mc-Si). A major limiting factor in their efficiency is recombination of electron-hole pairs at dislocations. Manufacturers of mc-Si wafers are focussing on reducing dislocation densities by modified casting techniques. However, an alternative approach is to remove the dislocations after growth and wafering. This thesis aims to explore the feasibility of using chemical processing to remove dislocations from mc-Si wafers.</p><p>For measuring the depth to which dislocation cores are removed an angle lapping technique is First developed. This allows the pit diameters and depth of bulk material removed to be measured. The parameter "aspect ratio" is defined as the ratio of dislocation etch pit depth to diameter; a high aspect ratio is desirable. The results gained allow the dislocation core removal process to be studied and optimised.</p><p>The first etching system explored is the Secco etch (1 K₂Cr₂O₇ (0.15M):2 HF (49%)), a commonly used defect-revealing etch. Samples are immersed in Secco etch for various durations at a range of temperatures, and pits are formed at the locations of dislocations and grain and twin boundaries. Secco etch is found to remove dislocations to depths of up to 40m. The activation energies for dislocation etching, etch pit widening and bulk etching are calculated as 0.20, 0.23 and 0.30eV respectively. Scanning electron microscopy is used to determine the geometry of the etch pits. Their structure is shown to be somewhat tubular, but the diameter of the pits is relatively large; too much bulk material is being removed. Further, at all temperatures the aspect ratio decreases with an increase in etch duration, so the aim of long, narrow etch tubes does not appear to be feasible using Secco etch.</p><p>Finding an alternative etching system with a lower bulk etch rate is the next aim. Anodic etching, where the silicon sample is biased whilst immersed in a solution, is therefore investigated, as control of the etch strength is provided by varying the electrical parameters. The overall etch rates for both dislocation etching and bulk etching are approximately two orders of magnitude lower; etch durations of over a day are needed to produce pits comparable to 20 minutes Secco etching. This, however, does not lead to improvements in aspect ratio, and the etching behaviour is hard to control. Anodic etching appears to be much more sensitive to small, local changes in condition, evidenced by varying results across different regions of a single sample.</p><p>Slower acting chemical etches are the final system that is investigated, and are chosen based on the strength of the oxidising agent present. The aim here is that the alternative etch solutions provide a very low bulk etch rate and are used at very long etch durations to produce long etch tubes with very high aspect ratio. An iodine-based etch (1 I₂ (0.01M):2 HF (49%)) is the only candidate found to produce good defect delineation, but the dislocation etch rate is very low and again the aspect ratio is unsuitable for removing dislocations deep into wafers.</p>
|
|---|