| Polycrystalline silicon is one of semiconductor materials with superior performance widely used in electronics and photovoltaic (PV) industry. At present, polysilicon is mainly industrially produced by the modified Siemens process. The chemical vapor deposition (CVD) of the trichlorosilane and hydrogen in Siemens reactor is the key process of the Siemens method. However, the low silicon deposition rate and high energy consumption are the main challenges in present. In order to reduce the energy consumption and save the production cost, hydrogen reduction process of SiHCl3 has become the subject of intense attention. With the rapid expansion in world production capacity and the change of international and domestic situation in recent years, the price of polysilicon sustains to decline. Under this situation, every proposal for reducing energy consumption and saving production cost becomes the.urgent needs of polysilicon enterprise. The numerical simulation technology is widely adopted to solve the problem in engineering, which could simulate the operate conditions easily which is hard to improved by experiment.Apparently, with the improvement of inspection level for Siemens reactor and intensive study of mechanism of complex phenomena, the excessively experience-dependent traditional mode to guide the actual polysilicon production with Siemens reactor should be abandoned, and the simulation system for CVD process, which could possess off-line predicting function, should be proposed. The novel research method can be used to analyze the working state of furnace in detail, and provide rich production information of reactor. Therefore, the aim of this doctor thesis is to investigate the CVD process of SiHCl3-H2 system in Siemens reactor, which is based on the industrial scale CVD reactor of a polysilicon production enterprise in Yunnan province and present fundanmental research of polysilicon deposition regarding multi-physics field coupled momentum, heat and mass transfer. Numerical simulation studied on the process from several points of view and levels aims to predict the information of the process and virtually reappear the state of furnace on a computer. This doctor thesis applies the theory of computational fluid dynamics (CFD) and numerical heat transfer (NHT) to analyze the CVD process in Siemens reactor, and the following results are carried out:Firstly, the convective heat transfer model for Siemens reactor has been developed. And the model is applied to predict the transport phenomena in a laboratory prototype to confirm the model valid. The relative error is within 1% by the comparison of calculated results and experimental data in open literature, which shows the model valid. The model is then employed to investigate the effect factors influencing the convective heat loss of silicon rod in the laboratory prototype and the specific energy consumption. By application of the convective heat transfer model to Siemens industrial scale reactor, the convection heat loss of silicon rod located in different rings have been researched in Siemens reactor with 12 pairs and 24 pairs of rods, respectively.The radiation heat transfer model for Siemens reactor has been developed, with which the thermal radiation heat transfer within the industrial scale CVD reactor is analyzed in detail for two state-of-the-art configurations:24 rods arranged in 2 rings,48 rods arranged in 3 rings. According to the principles of rods arranged in concentric rings in Siemens reactor, the arrangement of rods in several configurations is studied in detail. Based on radiation heat transfer model, the radiation exchange between the polysilicon rods and the reactor wall is analyzed in detail for the Siemens reactor with 12 pairs of rods. Furthermore, the power radiated by the characteristic rods of each ring to the wall and influence of the wall emissivity on radiation heat loss are also studied. The interesting ways to decrease radiation heat loss can be obtained: increasing radius of the rod at the end of the process and decreasing the wall emissivity by wall surface treatment or/and selecting more appropriate materials. Application of the radiation heat transfer model to a larger Siemens industrial scale reactor with 24 pairs of rods, the radiative behaviors of silicon rod in CVD process have been analyzed. The results show that increasing the total number of rods in the Siemens reactor meanwhile reducing the proportion of rods located in the outer ring has a very obvious effect on decreasing the radiation heat loss per unit area of rods. Moreover, the arrangement of present 48-rod Siemens reactor 1# is optimized to be 2# and the average radiation heat loss of rods in the optimized Siemens reactor 2# could be saved nearly 5% comparing to 1#.Based on the study of the convective and radiative heat transfer, the electric heating model of silicon rod for the Siemens reactor has been developed with the purpose of understanding the thermal and electrical behaviors of silicon rods in electric heating process, which is crucial for an optimal operation of the Siemens reactor. The proposed model is reliable for describing the behaviors of the rods since the relative error of the simulated results are within 10% by comparing to industrial data. Based on the developed model, the influence of the location of silicon rods within Siemens reactor and reactor wall emissivity during electric heating process has been investigated and the optimized current-voltage curves for the silicon rods located in different rings of 24-rod and 48-rod Siemens reactor. Interesting results show that the temperature gradient of silicon rods located in the outer ring is larger than the other ones when the rods are heated up by direct current (DC). The temperature gradient within the rod becomes smaller and the required voltage and current decreases when the emissivity of the reactor wall surface decreases. For the configurations with rods arranged in concentric rings, increasing the total number of rods in the Siemens reactor meanwhile reducing the proportion of rods located in the outer ring has a very obvious effect on decreasing the temperature gradient within the rod located in the inner ring, and the required voltage and current could also be decreased, simultaneously.At last, the transport-kinetics model of Siemens reactor in SiHCl3-H2 system has been developed by considering the momentum, heat and mass transfer. The model is employed to investigate the effect of parameters, including rod temperature, inlet gas temperature, velocity, composition and pressure, on the silicon growth rate in CVD process. Exactly understanding of flow field and temperature field in Siemens CVD reactor, the theoretical relationship between the initial condition and the react species in polysilicon CVD process can be established by application of fluid mechanics and reaction kinetics model of Siemens reactor.
|
PDF Full Text Request