Continuous Czochralski Growth of Silicon Single Crystals

9414606 PRASAD高产量、高性能微电子设备要求大直径硅(Si)晶片具有高度的晶体完整性、均匀且低的轴向和径向电阻率梯度、低氧杂质以及高度均匀的电气和机械性能。用Czochralski(CZ)方法生长的晶体中的许多不均匀性可以归因于生长动力学的非稳定性质，这主要是由于从开始到结束熔体高度的连续变化。当CZ工艺被放大以生长大直径晶体时，产生流动不稳定性和振荡的力变得更强。为了克服传统CZ工艺的许多缺点，将研究一种多晶硅颗粒进料连续Czochralski(CCZ)生长工艺。通过降低熔体高度和保持熔体高度不变，这种新工艺可以抑制多种不稳定动力学和不均匀。将进行建模、模拟、设计和实验的综合计划，以开发商业上可行的CCZ工艺。为了模拟具有自由和/或移动边界和界面的不规则区域中的三维输运过程，提出了一种基于多区域自适应网格生成和曲线有限体积离散的高分辨率计算机模型。这样就可以在一系列适用于工业过程的参数中准确地检查熔体流动再循环和振荡、熔体和晶体的热传递、晶体/熔体界面形状及其动力学、杂质传输和颗粒熔化的影响。在模拟CZ系统的装置中，将研究温度(使用液晶)和流场的非侵入性可视化、数字图像处理和传热学实验，以获得关于过程物理的基本信息。将尝试从二维水平和垂直图片进行3D图像的计算机重建。数值计算和实验室实验的结果将有助于设计CCZ生长实验，该实验将在工业研究设施的商业拉拔机上进行，以确定最佳工艺条件。该项目建立在先前研究的基础上，证明了使用小颗粒而不是熔化大块硅来连续生长硅单晶的可行性。工业合作伙伴承诺为该项目指派一名研究人员，进一步证明了预期的高成功概率。该项目的成功完成将打破目前限制可生产硅单晶尺寸的技术障碍，并为将目前的间歇工艺转变为真正意义上的连续工艺奠定基础，有可能提高产量和质量。这项技术的成功商业化将使美国能够在这个重要的电子市场保持其竞争优势。

9414606 Prasad High yield, high performance microelectronic devices require large diameter silicon (Si) wafers with a high degree of crystallographic perfection, uniform and low axial and radial resistivity gradients, low oxygen impurity and highly uniform electrical and mechanical properties. Many of the inhomogenities in crystals grown by the Czochralski (CZ) method (which is used to manufacture almost all Si crystals for microelectronics applications) can be attributed to the non-steady nature of the growth kinetics due primarily to the continuous change in melt height from start to finish. As the CZ process is scaled up to grow large diameter crystals, the forces which produce flow instabilities and oscillations become much stronger. To overcome many of the shortcomings of the conventional CZ process, a polysilicon pellets-feed continuous Czochralski (CCZ) growth process will be investigated. By reducing the melt height and keeping it fixed, this novel process can suppress many kinds of unsteady kinetics and inhomogenities. A comprehensive program of modeling, simulation, design and experiments will be performed to develop a commercially viable CCZ process. To simulate three- dimensional (3D) transport processes in an irregular domain with free and/or moving boundaries and interfaces, a high resolution computer model based on multizone adaptive grid generation and curvilinear finite volume discretization will be developed. It will then be possible to examine accurately the effects of melt flow recirculation and oscillations, heat transfer from the melt and crystal, crystal/melt interface shape and its dynamics, impurity transport, and pellets melting in a range of parameters suitable for industrial processes. Non-invasive visualization of temperature (using liquid crystals) and flow fields, digital image processing and heat transfer experiments in an apparatus simulating the CZ system to obtain basic information on the physics of the process will be investigated. Computer reconstruction of 3D images from two-dimensional horizontal and vertical pictures will be attempted. The results from numerical computations and laboratory experiments will help in designing the CCZ growth experiments which will be conduced in a commercial puller at an industrial research facility to determine the optimal process conditions. This project builds on prior research demonstrating the feasibility of continuous Czochralski growth of silicon single crystals using small pellets instead of melting a large block of silicon. Anticipated high probability of success is further demonstrated by the industrial partner's commitment to allocate a researcher to this project. Successful completion of this project will break down the current technological barrier limiting the size of silicon single crystal that can be manufactured and lays the foundation for converting the current batch process to truly continuous process with the potential to increase the yield as well as quality. Successful commercialization of this technology will enable the U.S. to maintain its competitive edge in this important electronics market.