西门子反应器中传热和多晶硅化学气相沉积的研究

发布时间：2018-03-04 09:00

  本文选题：多场耦合　切入点：西门子反应器　出处：《昆明理工大学》2016年博士论文　论文类型：学位论文

【摘要】：作为一种性能优越的半导体材料,多晶硅被广泛地应用于电子及光伏产业。改良西门子法是目前生产多晶硅的主流技术,其中三氯氢硅和氢气在西门子反应器中发生化学气相沉积过程是该工艺的关键工段,然而就如何提高硅沉积速率、降低还原能耗是目前工业生产遇到的最大挑战和难题。为了更有效地降低还原过程能耗、减少多晶硅生产成本,国内外科技工作者针对还原工艺进行了大量深入的研究。尤其是近年来,随着多晶硅产能的快速释放及国际国内形势的变化,多晶硅价格一直下降,逐步降低到成本线以下。在这种情况下,节能降耗、降低成本更加成为多晶硅企业的当务之急和迫切需要。目前,计算机数值模拟技术已广泛应用于各种工程实际中,尤其在实验条件或实验技术存在困难的情况下,计算机模拟方法往往是最为有效的方法。随着西门子反应器监测水平的提高和复杂现象机理研究的深入,应该逐渐抛弃原有靠经验指导西门子反应器实际生产的传统方式,建立起多晶硅化学气相沉积过程仿真体系,使其具备模型预测功能,从而能够细致分析炉况、提供丰富信息。为此,本文针对SiHCl3-H2西门子体系化学气相沉积过程,开展多场耦合下西门子反应器传热及多晶硅化学气相沉积基础理论研究。以云南某多晶硅生产企业实际生产用西门子反应器为研究对象,耦合流动、传热以及传质过程,建立多物理场耦合数学模型,从多角度和层次进行仿真,拟实地预见和再现西门子反应器炉内状态。论文应用计算流体力学(CFD)和数值传热学(NHT)理论剖析了西门子反应器还原过程,开展了以下几方面的研究工作：首先建立了西门子反应器中对流传热模型,并将模型预测得到的实验室规模西门子反应器的总能耗与公开文献实验数据进行对比,其相对误差都在1%以内,说明所建立的对流传热模型有效。分析了实验室规模西门子反应器中影响对流热损失主要因素,获得了单位产品多晶硅对流能耗的影响规律。同时应用该对流传热模型到工业规模西门子反应器中,预测获得了12对棒以及24对棒西门子反应器中各环硅棒表面单位面积对流热损失的变化趋势。建立了工业规模西门子反应器的辐射传热模型,重点分析了目前生产实践中常用的12对棒和24对棒两种炉型。根据目前工业应用的西门子反应器中硅棒圆周排布原则,分析了不同炉型反应器硅棒的排布方式,揭示了硅棒圆周排布的规律。分析了12对棒西门子反应器中硅棒的辐射行为,探究了硅棒半径以及反应器壁发射率对内、外环硅棒表面单位面积平均辐射热损失的影响规律,结果表明,增大硅棒最终沉积半径、降低反应器内壁发射率对降低硅棒表面辐射热损失具有明显效果。在12对棒反应器中硅棒辐射行为分析的基础上,对工业规模24对棒西门子反应器中硅棒的辐射行为也进行了相应理论分析,结果表明,增大反应器中硅棒数目同时降低外环硅棒数比例,对降低硅棒表面单位面积平均辐射热损失也具有明显效果。同时对现有的24对棒西门子反应器1#中硅棒排布方式进行优化,优化后的西门子反应器2#中硅棒的平均辐射热损失更低,理论节能近5%。在对流和辐射传热研究的基础上,为了进一步了解西门子反应器电加热过程中硅棒的热电行为,建立了硅棒的直流电加热模型,并将预测的12对棒西门子反应器各环硅棒的单位电压与工业实际测量数据进行对比,其相对误差都在10%以内,说明所建立的直流电加热模型有效。研究了工业规模12对棒和24对棒西门子反应器中硅棒辐射位置和反应器壁发射率对电加热过程的影响,获得了符合工业实际不同条件下各环硅棒电流-电压操作曲线。通过研究发现：随着硅棒半径逐渐增大,最外环硅棒内部径向温度梯度要明显高于其它环硅棒；降低反应器壁发射率,可以明显降低硅棒内部温度梯度、硅棒两端电压以及相应电流；对于圆周排布的西门子反应器而言,增大硅棒总数目同时降低最外环硅棒数比例,可以明显降低内环硅棒径向温度梯度,且能相应降低通过硅棒的电流和硅棒两端的电压,从而达到节能降耗的目标。最后,结合动量、热量、质量传递模型,耦合建立的SiHCl3-H2体系的反应动力学模型,构成了完整的传递-动力学模型。应用该模型分析了沉积温度、进气速度、进气组分、压强等因素对CVD过程硅沉积速率的影响。在准确获悉西门子反应器流场、温度场的基础上,将流体力学和反应动力学模型应用于多晶硅生长过程模拟,建立生长初始条件和反应过程各组分的理论关系。
[Abstract]:As a kind of semiconductor material with excellent performance, polysilicon is widely used in electronic and photovoltaic industry. The improved SIEMENS method is the mainstream technology of polysilicon production, including trichlorosilane and hydrogen in the SIEMENS reactor in a chemical vapor deposition process is the key section of the process, however, how to improve the deposition rate of silicon, reducing energy consumption is one of the biggest challenges and problems encountered in current industrial production. In order to effectively reduce the energy consumption reduction process, reduce the production cost of polysilicon, the domestic and foreign research in-depth for large amount of reduction process. Especially in recent years, with the change of the rapid release of polysilicon production capacity and the international and domestic situation, polysilicon prices have decreased gradually to reduce the cost of energy saving in the following lines. In this case, reduce the cost of polysilicon enterprises become more and more urgent matter and Urgent need. At present, simulation technology has been widely used in practical engineering in computer numerical, especially in the presence of experimental conditions and experimental techniques under difficult circumstances, the computer simulation method is often the most effective method. With the research on the mechanism of reaction to improve the monitoring level for SIEMENS and complex phenomena in depth, we should abandon the traditional mode of the original rely on the experience of guiding SIEMENS reactor of actual production, establish a simulation system of polysilicon chemical vapor deposition process, which has a model prediction function, so as to provide a detailed analysis of the furnace condition, rich information. Therefore, according to the SIEMENS SiHCl3-H2 system of chemical vapor deposition process in this paper, carry out theoretical research of SIEMENS reactor heat transfer and polysilicon chemical gas sedimentary facies based multi field coupling. In Yunnan a polysilicon production enterprise actual production by SIEMENS reactor as the research object, Coupled flow, heat transfer and mass transfer process, the establishment of multi physics field coupling mathematical model, from multi angle and level simulation, to anticipate and reproduce the SIEMENS field reactor furnace. The application of computational fluid dynamics (CFD) and numerical heat transfer (NHT) theory to analyze the reduction process of SIEMENS reactor, the research work has been carried out the following aspects: first, SIEMENS convection heat transfer reactor model was established, and compared the total energy consumption prediction model for the SIEMENS laboratory scale reaction and published experimental data, the relative error is less than 1%, indicating the convective heat transfer model is established. Analyzed the main factors affecting the convective heat loss reaction SIEMENS is in a laboratory scale, influence obtained unit product energy consumption. At the same time the application of polysilicon convection of the convective heat transfer model to industrial scale against SIEMENS Is, was predicted the trends of 12 units on the surface of silicon rod rod and 24 ring rod of SIEMENS in the reactor area of convective heat loss. A radiation heat transfer model for the SIEMENS reaction on an industrial scale, with emphasis on the production practice of the commonly used 12 sticks and 24 of two rod furnace according to the SIEMENS reactor silicon rod circular distribution principle of present industrial application, analyzes the arrangement of different type reactor silicon rods, silicon rods reveal the circular distribution law. Analysis of 12 pairs of SIEMENS stick reactor silicon rod radiation behavior, explores the silicon rod radius and reactor wall emissivity on the inside that rule, the outer surface of the silicon rod unit area average radiation heat loss. The results showed that the increase of silicon rod final deposition radius, reduce the reactor wall emissivity to reduce the heat loss of silicon rod surface radiation has obvious effect in 12. Based on analysis of the reactor rod silicon rod radiation behavior, the industrial scale of 24 bar SIEMENS reactor silicon rod radiation behavior has also carried on the corresponding theoretical analysis, the results show that the increase of reactor and reduce the number of silicon rod outer ring silicon rod number proportion, to reduce the silicon rod surface radiation heat loss per unit area also has obvious effect. At the same time, the existing 24 bar of SIEMENS reactor 1# silicon rod arrangement is optimized, the average heat loss is lower after the optimization of SIEMENS reactor 2# silicon rod, saving nearly 5%. based on the theory of heat convection and radiation research, in order to further understand the SIEMENS reactor electric heating process of thermoelectric silicon rod the behavior, established the silicon rod DC heating model, and the prediction of the 12 great SIEMENS reactor each ring of silicon rod unit voltage and industrial actual measurement data are compared, the The errors are less than 10%, indicating the DC heating model is established. On the 12 and 24 industrial scale to stick on SIEMENS silicon rod radiation reactor rod position and reactor wall emissivity effects on electric heating process, was consistent with the actual conditions of the different industrial silicon rod ring current voltage operation curve. It is found that with the silicon rod radius increases gradually, the outer ring of the silicon rod internal radial temperature gradient was higher than that of other ring silicon rod; reduce the reactor wall emissivity, can significantly reduce the temperature gradient inside the silicon rods, silicon rod ends voltage and current; for SIEMENS reactor circle configuration. Increase the total number of silicon rods and reduce the proportion of the number of silicon rod outer ring, inner silicon rod can significantly reduce the radial temperature gradient, and can decrease the silicon rod through the voltage current and the silicon rod ends, so as to achieve the festival The energy saving targets. Finally, combined with the momentum, heat and mass transfer model, kinetic model of coupled SiHCl3-H2 system established, constitute a complete transfer kinetics model. The model is used to analyze the deposition temperature, air velocity, air composition, the effect of pressure on the deposition rate of silicon in accurate CVD process. SIEMENS was informed that the reactor flow, based on the temperature field, fluid mechanics and reaction dynamics model is applied to polycrystalline silicon growth process simulation, establish the theoretical relationship between the growth of each initial condition and the reaction process.

【学位授予单位】：昆明理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN304.12
，

本文编号：1564970