基于迁移学习的大尺寸铸锭晶体硅热场设计

不同尺寸的铸锭晶体硅生长过程具有相似性,小尺寸晶体的生长规律可以迁移至大尺寸。本文采用迁移学习(TL)对G8型铸锭炉进行热场设计,设计对象为侧、顶加热器位置及体积、侧隔热笼分区块高度,主要设计目标为减少晶体内部的位错缺陷、抑制硅锭边缘多晶且使晶体生长界面微凸。首先使用神经网络对已有的G7铸锭炉建立热场几何参数与热场评价参数间的映射模型,然后将该模型迁移至G8铸锭炉,对比不同模型结构对迁移过程的影响,采用Dropout分析模型是否存在过拟合,并使用遗传算法(GA)结合聚类算法(CA)对热场几何参数进行优化,以上为G8热场设计过程。最后对优化结果采用数值模拟方法研究其在晶体生长过程中的温度分布、固液界面形状等,最终选定的优化方案能够实现较高质量的长晶。将该方案同时应用于G7和G8热场并进行对比,结果表明G8在硅熔体和硅晶体中的轴向温度梯度均小于G7,在晶体生长界面沿径向的温度梯度也小于G7,这有利于减小晶体内部的热应力。     

运用数值模拟技术改进VGF法生长GaAs晶体

在VGF法生长GaAs晶体的过程中固液界面凹向晶体,很容易在坩埚圆锥面处生长多晶.本文采用专业晶体生长模拟软件CrysVUn对实验温场进行了计算机模拟并提出改进方案,把底加热器取消并在坩埚底部加入氦气冷源,底部结构类似于热交换法系统.这样改变热场结构,得到凸向熔体的固液界面.     

大晶粒多晶硅铸锭生长的热场设计与模拟

为了生长大晶粒的多晶硅铸锭,晶体从形核到后续生长的热场环境控制至关重要.本文首先在侧加热器与散热块之间加一可移动的隔热环.通过向上移动隔热环,并在底部喷射氩气冷却,对生长工艺进行优化控制.然后利用数值模拟,对改进后的生长界面形状、晶体和熔体中的等温线、晶体和熔体的轴向温度分布以及冷却量对生长环境的影响进行分析.模拟结果表明:冷却速率的最佳值在5 ～ 15 W/m2之间,且优化后的晶体和熔体中等温线更平坦,晶体轴向温度梯度增大约1.72 K/cm,从而可有效地避免侧壁形核,促进大晶粒的生长,同时提高了生长速率.     

热屏和后继加热器对生长φ300mm硅单晶热场影响的数值分析

本文采用有限体积元法软件CrysVUn对直拉法生长直径300mm硅单晶热场和热应力分布进行了模拟。后继加热器通过补充晶体径向的热散失,使得沿生长界面径向的由熔体向晶体的热输运实现平衡,使晶体生长界面更加平坦。随着热屏材料热辐射率的降低,晶体生长界面趋于平坦,生长界面上方热应力水平也随着热屏材料辐射率的减小而下降,使用内层高辐射率材料、外层低辐射率材料的复合式热屏结构进一步降低了晶体生长界面中心高度。     

热交换器中氦气流量对生长蓝宝石温场影响的研究

采用热交换法生长蓝宝石晶体,通过数值模拟与实验研究了晶体生长过程中热交换器中氦气流量对温场的影响.结果表明:晶体生长过程中固液界面为近弧面;随氦气流量增大,晶体与熔体的温度下降,温度梯度增大;加热器功率缓慢上升.     

用数学模型计算生长大尺寸光学晶体加热系统结构 第2部分—在φ4 00mmCaF_2晶体生长中计算辐射-传导热交换的方法和结果(英文)

大尺寸氟化物晶体的生长是基于对晶体炉热交换的实验研究和计算结果 ,在晶体生长过程的不同阶段解决了复杂结构生长容器的边界条件和温场的二维计算任务。我们在这里给出了晶体生长过程中温场设置和转变的具体数据。所有的计算都是根据晶体、熔体 ,容器材料的光学特性与光谱和温度的关系以及它们的热物理值与温度的关系做出的。这些结果包括了迄今有关氟化物晶体生长系统和过程的最精确的数据 ,可用于生长技术工艺的发展以及晶体生长炉和容器的设计。     

用数学模型计算生长大尺寸光学晶体加热系统结构第2部分-在Ф400mm CaF2晶体生长中计算辐射-传导热交换的方法和结果

大尺寸氟化物晶体的生长是基于对晶体炉热交换的实验研究和计算结果,在晶体生长过程的不同阶段解决了复杂结构生长容器的边界条件和温场的二维计算任务.我们在这里给出了晶体生长过程中温场设置和转变的具体数据.所有的计算都是根据晶体、熔体,容器材料的光学特性与光谱和温度的关系以及它们的热物理值与温度的关系做出的.这些结果包括了迄今有关氟化物晶体生长系统和过程的最精确的数据,可用于生长技术工艺的发展以及晶体生长炉和容器的设计.     

甩数学模型计算生长大尺寸光学晶体加热系统结构第2部分-在φ400mm CaF2晶体生长中计算辐射-传导热交换的方法和结果

大尺寸氟化物晶体的生长是基于对晶体炉热交换的实验研究和计算结果，在晶体生长过程的不同阶段解决了复杂结构生长容器的边界条件和温场的二维计算任务。我们在这里给出了晶体生长过程中温场设置和转变的具体数据。所有的计算都是根据晶体、熔体，容器材料的光学特性与光谱和温度的关系以及它们的热物理值与温度的关系做出的。这些结果包括了迄今有关氟化物晶体生长系统和过程的最精确的数据，可用于生长技术工艺的发展以及晶体生长炉和容器的设计。     

泡生法蓝宝石单晶不同生长阶段的数值模拟研究

针对泡生法蓝宝石单晶生长的不同生长阶段的温场、流场和固液界面形状进行数值模拟研究.并分析了加热器相对坩埚的轴向位置和不同生长速率对蓝宝石单晶生长的影响.结果表明:在蓝宝石单晶生长中,在靠近坩埚壁面和固液界面的熔体内,等温线密,温度梯度较大;在靠近坩埚底部的熔体内,等温线稀疏,温度梯度较小.随着晶体高度的增加,熔体对流由放肩阶段的两个涡胞变成等径阶段的一个涡胞,熔体平均温度有小幅度下降;加热器相对坩埚的轴向位置对晶体生长炉内温场和固液界面形状影响很大,随着加热器位置上移,晶体内平均温度升高,温度梯度减小;熔体内平均温度降低,温度梯度增大.同时固液界面凸度增大.随着晶体生长速率增大,固液界面凸度增大,界面更加凸向熔体.     

大直径直拉硅单晶炉热场的改造及数值模拟

为了降低大直径硅单晶生长过程中氧的引入,对常规的406mm(16英寸)热场进行了改造.设计了以矮加热器为核心的复合式加热器系统,使晶体生长过程中熔体热对流减小.通过对热场的数值模拟计算,分析了热场的温度分布,发现熔体的纵向温度梯度下降,熔体热对流减小,硅单晶中氧含量降低.     

准单晶硅铸锭过程中加热功率及硅料堆积孔隙率对籽晶熔化的影响

本文在考虑硅料的堆积孔隙率和熔化变形等因素的基础上,建立了基于多孔介质的堆积硅料简化模型,对光伏太阳能用准单晶硅铸锭系统的硅料熔化过程进行了数值模拟,研究了不同侧/顶加热器功率比、堆积孔隙率以及加热器总功率对籽晶熔化的影响。研究结果表明:硅料的熔化时间和籽晶的熔化比例取决于侧/顶加热器功率比,降低侧/顶加热器功率比和堆积孔隙率有助于籽晶的有效保留,但会导致籽晶的熔化界面形状发生变化,使杂质在籽晶熔化界面形状为“凹”的区域内聚集,进而影响后续晶体生长的质量;当加热器的总功率低于临界值之后,籽晶的熔化界面形状会在靠近坩埚壁面的边缘区域发生变化,导致不均匀成核的发生,不利于准单晶硅铸锭的生产。在实际工况条件下,可以根据由侧/顶加热器功率比、堆积孔隙率、加热器总功率、籽晶的熔化比例和状态绘制的等值线图对工艺参数进行合理配置。     

A geometric and analytic approach to the process of melt vitrification. I. Thermodynamic treatment

By using simple geometric concepts, general thermodynamic dependences governing the process of vitrification are derived. The thermodynamic driving forces of crystallization, ΔG_f,c(T), and stabilization, ΔG_st(T), melts and glasses are considered in the entire temperature interval from temperatures below the glass transition temperature, T_g, to temperatures high above the melting point, T_m. It is show that at large deviations from T_g the temperature dependence of the driving force for stabilization, ΔG_st(T), is non-symmetric and expressions derived by a Taylor expansion become a poor approximation. By using an appropriate thermodynamic model, a more general expression for ΔG_st(T) is obtained and the thermodynamic basis for the existence of superheated glasses is established. The thermodynamic instability of glasses below and above T_g, i.e., of supercooled and superheated glasses is discussed.     

Dendrite spacing in unidirectionally solidified Al-Cu alloy

Directional solidification experiments have been carried out to study the variation in primary and secondary arm spacings with solidification parameters in the Al-Cu system. It is found that the primary arm spacing Z₁ obeys the common correlation Z₁ = KG^-aV^-b in the high velocity regime or low temperature gradient regime, where a and b are constants, the value of K included the composition dependence, G is the temperature gradient in the liquid and V the growth rate; however, it does not obey this correlation for the low velocity regime or high temperature gradient regime but goes through a maximum or a catastrophe as a function of V or G at V = V_cs/k or G = kG_cs, where k is the equilibrium distribution coefficient, and V_cs and G_cs are the critical velocity and temperature gradient at the limit of constitutional undercooling respectively. The initial secondary arm spacings Z₂₀ are nearly independent of G and mainly depended on V, Z₂₀ = 0.016 V^-0.54 (mm). The secondary arm spacing Z₂ tends to c oarsen with time and thus is a function of coarsening time t_f, Z₂ = 0.016t^0.34_f (mm). Theoretical analyses of the primary arm spac ing and the initial secondary arm spacing have been proposed, and the derived relationships agree reasonably well with the above experimental results.     

自支撑氮化硅膜结构制备工艺优化

自支撑氮化硅膜结构一般是基于微纳加工技术来制备的。为了提高膜结构的品质,本文分别对自支撑氮化硅膜结构制备中的干法刻蚀参数和各向异性湿法腐蚀参数进行了研究和优化,其中干法刻蚀参数主要包括反应气体配比和刻蚀时间,各向异性湿法腐蚀参数主要包括腐蚀剂浓度和腐蚀温度。在不同的参数组合下进行实验,使用光学显微镜观察并比较不同样品的表面形貌,得到了较理想的参数组合。在干法刻蚀的反应气体中加入少量O₂可改善刻蚀效果,反应气体配比V(SF₆)∶V(CHF₃)∶V(O₂)=6∶37∶3,刻蚀时间2 min。湿法腐蚀中腐蚀剂在质量分数25%处达到最大的硅腐蚀速率,同时氮化硅表面形貌也较理想。     

On the theory of the weighing method for automatic crystal shape control in czochralski growth

We derive the exact analytical expression for the meniscus volume in terms of the crystal radius R, the melt contact angle θ_L, and the meniscus height h for the axisymmetric planar growth interface Czochralski configuration. This relation is applied together with motion and mass balance considerations to find the general rate of change in the force experience by the weighing cell in a standard diameter control arrangement. In contrast to the classical work of Bardsley et al. (1977), which is limited to the close viscinity of stationary right cylinder growth, our results are valid for growth developing any crystal profile. In addition, they also account for the fall of melt level due to finite crucible size. The magnitudes of the capillary contributions to the force, and their implications for diameter servo control, are discussed over a wide range of profile parameters using the Tsivinskii formula for h = h (R, θ_L).     

Growth of NaBi(WO4)2 crystal by modified-Bridgman method

NaBi(WO₄)₂ (NBW) crystals have been grown for the first time by modified-Bridgman method. Influences of some factors on the crystal growth process are discussed. X-ray powder diffraction experiments show that the unit cell parameters of NBW crystal are a=b=0.5284 nm, c=1.1517 nm, and V=0.3215 nm³. The differential thermal analysis shows that the NBW crystal melts at 923°C.     

硅单晶生长工艺参数建模及多目标优化

为了获得满足高品质硅单晶体生长的工艺参数,一种计算流体动力学(computational fluid dynamics,CFD)方法、数据处理组合法(group method of data handing,GMDH)型神经网络和第二代非支配排序遗传算法Ⅱ(nondominated sorting genetic algorithm Ⅱ,NSGA-Ⅱ)的混合策略被提出,并对Czochralski(Cz)法硅单晶生长进行建模和工艺参数的多目标优化.将包含了加热温度、晶转速度、埚转速度和提拉速度等四个设计变量的固液界面形变量h和缺陷评价准则V/G作为目标函数.通过CFD进行数值计算,获得GMDH训练所需要的样本,并建立目标函数多项式模型,最后利用NSGA-Ⅱ得到满足了Pareto最优解的工艺参数.通过实际工程验证,证明了所提出的混合策略为获取晶体生长工艺参数提供了一种新的数值计算方法.     

Glass-forming ability versus stability of silicate glasses. II. Theoretical demonstration

Possible relationships between measures of glass stability (GS) against devitrification on heating (evaluated by the Hrubÿ parameter K_H=(T_c^h−T_g)/(T_m−T_c^h), and the parameter K_w=(T_c^h−T_g)/T_m) and a criterion of glass-forming ability (GFA) – the critical cooling rate – were investigated by computing non-isothermal crystallization for typical values of the main quantities that control crystal nucleation and growth in silicate glasses. We limit these quantities to one thermodynamic parameter – the melting entropy (ΔS_m) and two kinetic parameters that control the viscosity (B and T₀ in the Vogel–Fulcher–Tamman equation or T_g and in Avramov’s equation). The effect of heterogeneous nucleation and, in particular, the possible role of the surface as active substrate is tested. The results presented herein demonstrate that GS and GFA are indeed related concepts.     

Impedance spectroscopy analysis of AlGaN/GaN HFET structures

For HFET application a series of samples with 30 nm Al_xGa_1−xN (x=0.02–0.4) layers deposited at 1040°C onto optimised 2 μm thick undoped GaN buffers were fabricated. The Al_xGa_1−xN/GaN heterostructures were grown on c-plane sapphire in an atmospheric pressure, single wafer, vertical flow MOVPE system. Electrical properties of the Al_xGa_1−xN/GaN heterostructures and thick undoped GaN layers were evaluated by impedance spectroscopy method performed in the range of 80 Hz–10 MHz with an HP 4192A impedance meter using a mercury probe. The carrier concentration distribution through the layer thickness and the sheet carrier concentration were evaluated. A non-destructive, characterisation technique for verification of device heterostucture quality from the measured C–V and G–V versus frequency characteristics of the heterostructure is proposed.