QQ客服 官方微博 反馈留言

题名	基于粒子群算法的光刻机光源掩模投影物镜优化技术
作者	王磊
文献子类	博士
导师	王向朝
关键词	光刻 optical lithography 分辨率增强技术 resolution enhancement technique 光源掩模投影物镜优化 source mask projector optimization 粒子群算法 particle swarm optimization
其他题名	Source Mask Projector Optimization using Particle Swarm Optimization Algorithm in Optical Lithography
英文摘要	光刻技术是极大规模集成电路制造的关键技术之一。光刻分辨率决定集成电路的特征尺寸。为了制造具有更小特征尺寸的集成电路，业内提出了一系列的光刻分辨率增强技术。传统的分辨率增强技术通过单独地优化光源或掩模，有效提高了光刻分辨率。光源掩模优化技术相比于传统分辨率增强技术具有更强的分辨率增强效果，是28nm及以下技术节点集成电路制造的主要分辨率增强技术之一。相比于光源掩模优化技术，光源掩模投影物镜优化技术同时优化照明光源、掩模图形和投影物镜，具有更高的优化自由度，可以进一步提高光刻分辨率。基于遗传算法的光源掩模投影物镜优化技术不需要掌握光刻先验知识，对光刻成像模型和优化目标适应性强。然而，基于遗传算法的光源掩模投影物镜优化技术通过模仿自然进化过程进行优化搜索，优化时所需的大量光刻仿真计算导致其收敛速度较慢，从而限制了该技术的实际应用。针对此问题，本文提出了一种基于粒子群算法的光源掩模投影物镜优化技术，围绕本技术开展了光源优化技术、掩模优化技术及光源掩模投影物镜优化技术方面的研究，主要内容包括： 1. 提出了一种基于粒子群算法的光源优化技术。以接触孔阵列掩模图形、密集线掩模图形和含有交叉门的复杂掩模图形为例进行了仿真验证，三种图形的图形误差分别降低了69.6%，51.5%和40.3%。结果表明本技术有效提高了光刻成像质量。与基于遗传算法的光源优化技术相比，本技术在得到相近成像质量的前提下，具有更快的收敛速度。与基于梯度法的光源优化技术相比，本技术不需要计算梯度值，对光刻成像模型及优化目标具有较强的适应性。研究了初始照明光源形状对基于粒子群算法的光源优化技术的影响，结果表明本技术对初始照明光源形状具有较强的鲁棒性。 2. 提出了一种基于粒子群算法的掩模优化技术。采用离散余弦变换将掩模图形从空域转换到频域，采用粒子群算法在频域对掩模图形进行优化。以密集线掩模图形和含有交叉门的复杂掩模图形为例进行了仿真验证，两者的图形误差分别降低了94.8%和84.2%。结果表明本技术有效提高了光刻成像质量。通过选取掩模图形的低频部分作为变量进行优化，在提高光刻成像质量的同时，有效降低了优化后掩模图形的复杂度，增强了优化后掩模的可制造性。 3. 提出了一种基于粒子群算法的光源掩模投影物镜优化技术。采用粒子群算法对由像素表征的光源、由离散余弦变换基表征的掩模及由泽尼克系数表征的投影物镜进行联合优化。在标称条件和工艺条件下，以含有交叉门的复杂掩模图形为例进行了仿真验证，图形误差分别降低了94.2%和93.8%。结果表明本技术有效提高了光刻成像质量。本技术具有优化自由度高、收敛速度快和优化后掩模可制造性强的优点，适用于光刻机光源参数、掩模参数和投影物镜参数的联合优化。同时，本技术为实现全芯片多参数联合优化奠定了技术基础。; Optical lithography is one of the key technologies in the manufacturing of ultra- large-scale integrated circuits. The resolution of lithography determines the critical dimension of the integrated circuits. Resolution enhancement techniques are used to fabricate integrated circuits with smaller critical dimension. Traditional resolution enhancement techniques optimize the illumination source or the mask pattern separately to improve the resolution of lithography. In comparison with the traditional resolution enhancement techniques, the source mask optimization technique has a better performance to scale down the resolution of lithography. The source mask optimization technique is considered as one of the key techniques in 28nm technological node and beyond. Compared with the source mask optimization technique, source mask projector optimization technique optimizes the illumination source, the mask pattern and the projector simultaneously and has higher degree of freedom for optimization. The source mask projector optimization technique can further improve the resolution of lithography. The source mask projector optimization technique using genetic algorithm requires little prior knowledge of the lithography process, and it has a strong adaptation to various lithographic imaging models and optimization objectives. However, the source mask projector optimization technique using genetic algorithm imitates the principle of natural evolution and it requires the lithography simulation to be performed many times. It slows down the optimization process and restricts the practical application of this technique. In this dissertation, a source mask projector optimization technique using particle swarm optimization algorithm is proposed. Centering on the proposed technique, the source optimization technique, the mask optimization technique and the source mask projector optimization technique are studied. The main contents are as follows: 1. A source optimization technique using particle swarm optimization algorithm is proposed. The proposed technique is verified by using three typical mask patterns, including a periodic array of contact holes, a vertical line/space pattern and a complicated pattern with cross gate design. The pattern errors are reduced by 69.6%, 51.5% and 40.3%, respectively. Compared with the source optimization technique using genetic algorithm, the proposed technique obtains comparable optimization results with faster convergence rates. Compared with the source optimization technique using gradient descent algorithm, the proposed technique does not need the calculation of gradients, and it has a strong adaptation to various lithographic imaging models and optimization objectives. The robustness of the proposed technique to initial illumination sources is also verified. 2. A mask optimization technique using particle swarm optimization algorithm is proposed. The discrete cosine transform is applied to transform the mask patterns from the spatial domain to the frequency domain. Then the frequency components of the mask patterns are optimized by using particle swarm optimization algorithm. The proposed technique is demonstrated by using two typical mask patterns, including a vertical line/space pattern and a complicated pattern with cross gate design. Simulation results verify the validity of the proposed technique. The pattern errors are reduced by 94.8% and 84.2%, respectively. Besides, the manufacturability of the optimized mask patterns is enhanced when only the low frequency components are optimized. 3. A source mask projector optimization technique using particle swarm optimization algorithm is proposed. The discrete cosine transform basis functions are used to represent the mask patterns. The coefficients of Zernike polynomials are used to represent the projectors. The pixelated illumination sources, the mask patterns and the projectors are encoded into particles. These particles are optimized by using particle swarm optimization algorithm. This technique is demonstrated by using a complicated pattern with cross gate design in nominal condition and process condition. The pattern errors are reduced by 94.2% and 93.8%, respectively. The proposed technique has advantages of high degree of freedom for optimization, fast convergence rate and enhanced manufacturability of the optimized mask patterns. The proposed technique is suitable to the joint optimization of illumination sources, mask patterns and projectors in lithography tools. At the same time, this technique establishes a foundation for full-chip multi-parameter optimization.
学科主题	光学工程
内容类型	学位论文
源URL	[http://ir.siom.ac.cn/handle/181231/31025]
专题	中国科学院上海光学精密机械研究所
作者单位	中国科学院上海光学精密机械研究所
推荐引用方式 GB/T 7714	王磊. 基于粒子群算法的光刻机光源掩模投影物镜优化技术[D].

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。