【摘要】 针对传统频域方法用于分析金属铝互连中的噪声信号的局限性,本文采用多尺度熵方法分析电迁移噪声时间序列.结果表明在电迁移早期,噪声信号较不规律,复杂度较大;随空洞成核的发生,噪声信号规律性增强,复杂度明显减小,反映出随电迁移过程的进行系统的混乱度不断减小.通过与传统表征参量的对比,证明多尺度熵能够对电迁移失效过程进行表征.更多还原

【Abstract】 In order to solve the limitation of traditional frequency-domain analysis in Al interconnection electromigration noise,the multiscale entropy(MSE) method is proposed to analyze the electromigration noise time series.The result shows that at the early stage,electromigration noise is irregular,the noise complexity is large,and with the occurrence of nucleation,the electromigration noise becomes more regular,and the complexity decreases obviously.It indicates that the disorder degree of electromigration system decreases continually.Compared with traditional parameters,MSE can be used to successfully characterize electromigration process.更多还原