论文部分内容阅读
随着集成电路特征尺寸进入2Xnm及以下节点,光源与掩模联合优化(SMO)成为了拓展193nm ArF浸没式光刻工艺窗口、减小工艺因子的重要分辨率增强技术(RET)之一。提出了一种基于随机并行梯度速降(SPGD)算法的SMO方法,通过随机扰动进行梯度估计,利用估计梯度来迭代更新光源与掩模,避免了求解梯度解析表达式的过程,降低了优化复杂度。对周期接触孔阵列及十字线、密集线三种掩模图形的仿真验证表明,三种掩模图形误差(PE)值分别降低了75%、80%与70%,该方法较大程度地提高了光刻成像质量。
As integrated circuit feature sizes reach 2Xnm and below nodes, light source-mask co-optimization (SMO) becomes one of the key resolution enhancement technologies (RETs) that extend the 193nm ArF immersion lithography process window and reduce process factors. An SMO method based on Stochastic Parallel Gradient Descent (SPGD) algorithm is proposed. By using stochastic perturbations to estimate the gradient and updating the light source and mask iteratively using the estimated gradient, the process of solving the gradient analytic expression is avoided and the optimization complexity degree. Simulation results of the periodic pattern of contact holes and cross lines and dense lines show that the mask pattern errors (PE) of the three masks are reduced by 75%, 80% and 70%, respectively, and this method is greatly improved Lithography imaging quality.