Characterization and fundamental investigation of laser doping for silicon solar cells

Advanced structures with localized contacts can enable high efficiency crystalline silicon solar cells, but they are complex and costly to fabricate using conventional techniques. Hence, the development of new, cheaper processes to produce localized contact regions is of great interest. One of the most promising processes is laser doping. This thesis explores three main areas related to the challenges of laser doping, in particular the characterization of the laser-doped regions, the degradation of the silicon/dielectric interface in the vicinity of the laser-doped regions, and the degradation of the laser processed silicon. A technique named secondary electron microscopy dopant contrast imaging (SEMDCI) is firstly introduced for the characterisation of the cross-section/top-surface of laser-doped samples. This technique can be used for a large range of different dopant sources and different laser doping methods. The technique can be employed to obtain quantitative dopant density images for p-type laser-doped regions, albeit currently over a limited range of dopant densities and with relatively large error. Using this technique, the risk of metallization shunts near the edges of dielectric film windows opened by the laser can be evaluated. Furthermore, this technique is useful for understanding the interaction between different materials during the laser doping process. The impact of the silicon nitride layer thickness in silicon dioxide / silicon nitride stacks on the properties of laser-doped lines is investigated through measurement of the doping profile near the edge of the dielectric window which is an important factor in determining the likelihood of high recombination or even shunting from the subsequent metallization process. Fundamentally, a problem of exposed and undoped silicon near the dielectric window is identified for most of the investigated parameter range. However, optimization of the laser parameters and dielectric film conditions is shown to be capable of preventing or at least minimizing this problem. Beside this problem, the residual dielectrics close to the edge might be degraded after laser process. Therefore, experimental methods are proposed to detect this possible degradation. In the last part of the thesis, boron diffused samples capped with different dielectric films (including bare surfaces) are processed using laser pulses and characterized by photoluminescence imaging (PL) to study the degradation of the electronic properties of the processed regions. In this way, without the interference of a dopant precursor, the thermal and dielectric effects are separately investigated. It is found that the thermal effects do not lead to significant damage and additional recombination, provided no severe silicon evaporation occurs. However, when a dielectric film is present, a considerable increase in recombination is observed irrespective of laser parameters. The magnitude of the increase in recombination varies substantially depending on the dielectric used. It is demonstrated that both sufficient silicon melting and low recombination can in principle be achieved, particularly using long pulse durations and small pulse distances. In addition, nitrogen gas annealing after laser process was proved to be an effective way to reduce the defects induced by both the thermal effects and dielectrics