SBIR Phase I: Disruptive Semiconductor Software Tool for Recipe Optimization for Deposition and Etching Processes

这个小企业创新研究（SBIR）第一阶段项目的更广泛的影响/商业潜力是，它将使自旋扭矩转移随机存取存储器（STT-RAM）和高效太阳能电池等尖端技术渗透到市场。目前，像这些技术的半导体应用面临着重大的纳米制造挑战。事实上，制造半导体器件是如此具有挑战性，以至于55%的新半导体产品无法满足其最初的发布日期，超过40%的开发项目超出了计划预算。通过虚拟配方开发环境，SBIR项目将使半导体芯片和设备制造商节省高达66%的工艺开发成本，通过三倍的开发周期获得市场份额，并实现下一代高性能，节能电子器件的工艺开发。拟议的项目将使用统计自学习推理算法、复杂的工艺模型和大量可用的晶圆厂数据来进行高精度的工艺预测，并实现创新纳米技术的大批量制造。将开发三个关键组件：（1）能够执行多个工艺目标的工艺预测和工艺设计的配方分析引擎，（2）用于纳米特征轮廓预测的形貌模拟器，以及（3）用于实时工艺开发决策的商业平台接口。这些组件将被设计为准确地捕获大型过程系统和多个过程目标，降低计算费用，并促进商业采用。这些创新将通过以下方式实现：（a）使用并行计算来减少计算费用，（B）开发降阶等离子体和表面动力学模型，（c）采用自适应算法来加速配方优化，及（d）使用不确定性分析技术，该奖项反映了NSF的法定使命，并通过使用基金会的智力价值进行评估，被认为值得支持和更广泛的影响审查标准。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is that it will enable technologies on the cutting-edge like Spin Torque Transfer-Random Access Memory (STT-RAM) and high efficiency solar cells to permeate the marketplace. Currently, semiconductor applications like these technologies face significant nanomanufacturing challenges. In fact, fabricating semiconductor devices is so challenging that 55% of new semiconductor products fail to meet their original launch date and over 40% of development projects exceed the planned budgets. With its proposed virtual recipe development environment, this SBIR project will allow semiconductor chip and equipment manufacturers to save up to 66% of process development costs, gain market share through three times faster development cycles, and enable the process development of next-generation of high performance, energy efficient electronic devices. The proposed project will use statistical self-learning inference algorithms, sophisticated process models, and the vast amount of available fab data to make high accuracy process predictions and enable high volume manufacturing of innovative nanotechnologies. Three key components will be developed: (1) A recipe analytics engine capable of performing process predictions and process design for multiple process objectives, (2) A topography simulator for profile prediction of nanofeatures, and (3) A commercial platform interface for real-time process development decision making. The components will be devised to accurately capture large process systems and multiple process objectives, reduce computational expense, and facilitate commercial adoption. These innovations will be accomplished by: (a) Using parallel computing to reduce computational expense, (b) Developing reduced-order plasma and surface kinetic models, (c) Employing adaptive algorithms to accelerate recipe optimization, and (d) Using uncertainty analysis techniques for recipe co-optimization.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.