层叠微芯片封装翘曲行为优化分析  

Warpage optimization analysis of chip packaging process

作　　者：罗成 吴文云[1] 廖秋慧[1] 黄涛 LUO Cheng;WU Wenyun;LIAO Qiuhui;HUANG Tao(School of Material Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学材料工程学院,上海201620

出　　处：《智能计算机与应用》2021年第1期161-166,169,共7页Intelligent Computer and Applications

摘　　要：本文以某汽车用芯片为研究对象,研究芯片封装过程结构翘曲优化问题。首先采用Taguchi正交实验设计,结合M oldflow 2016微芯片封装模拟软件,分析各因素对芯片封装过程结构翘曲影响程度及影响规律。选择对芯片翘曲影响较大的因素为响应试验因素,芯片翘曲值为响应目标,进行Box-Behnken试验设计,建立响应面试验因素与目标的数学模型。利用Box-Behnken试验设计构建的数学模型,定义遗传算法优化适应度函数,基于Matlab 2016软件遗传算法工具箱(GUI),通过迭代寻优,获得芯片封装结构翘曲的最小值及最小值时的参数组合。按照芯片翘曲最小值,对芯片原始模型进行反变形补偿,通过实际生产验证,该优化方法具有较高的精度。In this paper,an automotive chip is taken as the research object,and the warpage optimization of chip packaging process is studied.Firstly,Taguchi orthogonal experiment design and Moldflow 2016 simulation software are used to analyze the influence of various factors on the warpage of chip packaging process.A box-Behnken experimental design is carried out to establish the mathematical model of response surface test factors and targets.Using the mathematical model constructed by box-Behnken experimental design,the fitness function of Genetic Algorithm Optimization is defined.Based on Matlab 2016 software genetic algorithm toolbox(GUI),the minimum value of chip package structure warpage and the parameter combination of minimum value are obtained through iterative optimization.According to the minimum warpage of the chip,the original model of the chip is compensated by anti deformation.The practical production shows that the optimization method has high accuracy.

关 键 词：芯片翘曲 Moldflow微芯片封装 Taguchi正交实验 Box-Behnken试验 遗传算法优化 

分 类 号：TQ320.66[化学工程—合成树脂塑料工业] TP391.7[自动化与计算机技术—计算机应用技术]