Motivation
The Czochralski (Cz) method is the main technology used for obtaining Silicon crystals for electronics and photovoltaic applications. Initially the Silicon crystals obtained with the Cz method have been used in the electronics industry that requires high purity crystals, at least 99.999999 (9 Ns) pure. The scraps from the electronic industry have been, for a long period of time, the main source for the feedstock used in the photovoltaic (PV) industry that, in order to reduce production costs, doesn't require Silicon crystals as pure as the ones used in the electronic industry (99.9999 - 6 Ns pure). The process of obtaining monocrystalline Silicon crystals with the Cz method is a hot topic and has been extensively studied. The solar cell efficiencies is correlated directly with the material quality, i.e. higher efficiency is achieved the less crystal defects and impurities are present in the material. Therefore it is important to study the transport of impurities in the melt as it has an influence on the species incorporation in the melt. Since silicon at high temperature reacts easily with almost everything, and due to purity requirements, manufactures are restricted to very few crucible materials. In addition, requirements for mechanical strength are also limiting factors. Therefore, silica remains the material of choice for crucible and mould applications because it is readily available in high purity form and because the reaction product of Si and SiO2 is a gaseous phase, which can be removed from the heat zone, thereby minimizing the contamination that can occur during the production of silicon ingots. The three largest drawbacks of silica as crucible material are: being a thermal insulator, oxygen contamination of silicon, and cost due to single use (contributes to ~30% to the conversion from poly-Si to the as-grown ingot). Due to the difference in thermal expansion and phase transformation the crucible cracks during the cooling process. After usage the crucible is discarded and landfilled. Therefore, research into alternative crucible materials is currently on-going. Silicon nitride (Si3N4) crucibles are an attractive alternative because of the absence of oxygen and, in contrast to silica crucibles, they have the potential of being reused. Reusable crucibles made entirely of silicon nitride have been proposed and patented, but no commercial application of these types of crucibles is known. Incorporation of nitrogen in the melt occurs because of the interaction between the silicon melt and the Silicon nitride crucible. The first goal of this present proposal is to study the nitrogen transport from a Si3N4 crucible into the melt. However, the silicon melt contains phosphorus and boron particles that are added to the feedstock in order to achieve an either n-type or a p-type conductivity depending on the technical applications. There particles diffuse into the ingot during growth, altering the resistivity profile and jeopardizing the solar quality of the silicon.
In addition, molten silicon wets silicon nitride, which causes the melt to creep up on the inside wall of the crucibles and might even leak out of the crucible. Wetting tendency can be reduced by increasing the furnace chamber pressure. Several attempts have been made to develop reusable crucibles for silicon ingot crystallization, but all have failed to be introduced into the market. However, the Steuler company has developed a production process with proprietary production technology for reaction bonded silicon nitride materials where physical and mechanical properties are substantially improved. Through these achievements the manufactured silicon nitride is classified as "Advanced Ceramic". Initial tests have shown that it is possible to grow mc-Si ingots by DS and sc-Si by the Cz process using crucibles of this material. In the present proposal a study of the fluctuations in dopant concentration near the solid-liquid interface will be done in order to correlate the results with the crystal striations observed in the solidified Silicon ingot at the SINTEF Materials and Chemistry Institute in Norway that our group has a long collaboration.
Another important issue in crystal growth is the control of melt convection in order to increase the quality of the grown silicon ingots. The convection control is done to either discard the perturbations produced by the natural convection which can influence the shape of the solid-liquid interface and, thus, the crystal quality or to achieve a homogenization of the impurities in doped melts. The magnetic fields have proven to be an efficacious method to control the melt flow in electrically conducting melts. For the Cz grown Silicon crystals the use of magnetic fields as an external force to control the melt convection is well established and the research in this direction is ongoing. Depending on the applications, steady-state magnetic fields (e.g., vertical, transverse, cusp-shaped) or time-dependent magnetic fields (e.g., rotating, alternating, traveling) can be used. A review of the applications of magnetic fields in the growth process of semiconducting crystals is done in [D. Vizman - Flow Control by Magnetic Fields during Crystal Growth from Melt, Book Chapter, Handbook of Crystal Growth: Bulk Crystal Growth, 2014, 909-950, Elsevier]. The influence of the forced convection, generated in the melt by applying various configurations of magnetic field, on the concentration fluctuations for a Cz configuration hasn't been studied yet. This project proposes to investigate, through numerical modeling, the influence of different type of magnetic fields on the fluctuations of impurity concentration in a Cz Silicon melt.
Because the real crystal growth experiments are very expensive, numerical modeling is a necessary tool in the optimization of the crystal growth process. Therefore, the present project will provide an understanding of the impurity transport process in a Czochralski configuration for the growth of solar Silicon crystals. It will be expected to prove a correlation between crystals striations observed experimentally at SINTEF and the impurity concentration fluctuations near the solid-liquid interface obtained numerically. Using numerical simulation, an analysis of the way the complex forced melt convection generated by external magnetic fields in the Silicon melt will influence the fluctuations of the impurity concentration in the melt, will also be performed.