Abstract: Wafer yield is an important indicator to measure the quality of wafer products,and achieving stable optimization can effectively control production costs.In view of the numerous factors affecting wafer yield and the huge amount of data,according to the wafer acceptance testing,a wafer yield optimization method based on improved particle swarm algorithm was designed.Based on the preprocessing of wafer acceptance test data and the extraction of key parameters,a wafer yield prediction model was established for evaluating particle fitness,an iterative adaptive particle swarm inertia factor and acceleration factor were designed,and the local search mechanism based on simulated annealing was designed,which achieved the goal of maximizing wafer yield with minimum adjustment cost.In the example test of a wafer manufacturing workshop,the optimization results of improved particle swarm algorithm,genetic algorithm and traditional particle swarm algorithm were compared and analyzed,and the effectiveness of the proposed method was verified.

Key words: wafer yield improvement, particle swarm optimization, wafer acceptance test, prediction model, local search

摘要： 晶圆良率是衡量晶圆产品质量的重要指标,实现其稳定优化能够有效控制生产成本。针对晶圆良率影响因素众多、数据量庞大等特点,以晶圆允收测试为依据,设计了基于改进粒子群算法的晶圆良率优化方法。该方法在晶圆允收测试数据预处理与关键参数提取基础上,建立评价粒子适应度的晶圆良率预测模型,设计迭代自适应的粒子群惯性因子与加速因子,以及基于模拟退火的局部搜索机制,实现最小调整成本下的晶圆良率最大化目标。在某晶圆制造车间算例实验中,通过对比分析验证了所提方法的有效性。

关键词: 晶圆良率优化, 粒子群算法, 晶圆允收测试, 预测模型, 局部搜索