Modelling of Solar Grade Silicon Crystallization Processes
Background and motivation
The conversion efficiency of multi-crystalline silicon (mc-Si) solar cells is strongly influenced by several types of impurities and defects that originate from the crystallization process. Crystallization front dynamics during casting and flow characteristics in the melt affect the structure and the quality of the crystals. In particular the dislocation density in crystal grains, the shape of the grains and intra-grain defects, such as impurities and small clusters of atoms or precipitates can significantly reduce the local solar cell efficiency.
Localized regions of high dislocation density have been identified to create particularly detrimental defect centers in mc-Si, and their occurrence should ideally be minimized. Dislocations are generated during ingot growth and cooling by mechanical stresses that are caused by temperature gradients inside the solidified ingot. Improved understanding of the complex crystallization process will enable reduction of the dislocation densities.
Numerical simulation has emerged in the last decades as an effective tool for the development and optimization of technical processes. This is especially true for processes with high demand on quality and high added value to the product. Near-perfect control of the temperature and flow distribution inside the crystallization furnace is essential for high-efficiency solar cells. The unique advantage of an advanced computer model, in addition to the ability to calculate macroscopic parameters such as temperatures, fluid flow, stresses, etc, is the possibility of predicting the distribution of even microscopically small crystal defects such as dislocations.
Recombination of charge carriers at various impurities and defects is an important physical mechanism that limits the efficiency of solar cells. While the effects of many impurities can be mitigated through various processes used during the solar cell fabrications, the dislocations in the crystal lattice are a particular source of problems in mc-Si solar cells, since the energy required for removing such defects is excessively high. The formation of dislocations can be attributed to plastic deformations at high temperatures. Different approaches have been applied in modelling of plasticity in casting of silicon. The zones with the highest dislocation density are often associated with a locally reduced efficiency. It is therefore evident that, in order to use modelling of the casting process to predict the efficiency of solar cells, modelling of the development of the dislocation density and its distribution must be brought to a higher level. It is not clear what causes these strong variations. However, impurities in the form of dissolved atoms or particles are believed to play an important role. These issues are addressed in the present project.
Measurements of the dislocation structure and density are possible, and the challenge will be to combine experimental data, fundamental theory and process simulations providing temperature and strain histories to build a model that can capture the key phenomena leading to the observed dislocation pattern.
Project leader is Dag Lindholm.