Signal processing and analysis for copper layer thickness measurement within a large variation range in the CMP process

Hongkai Li, Qian Zhao, Xinchun Lu, Jianbin Luo

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)

Abstract

In the copper (Cu) chemical mechanical planarization (CMP) process, accurate determination of a process reaching the end point is of great importance. Based on the eddy current technology, the in situ thickness measurement of the Cu layer is feasible. Previous research studies focus on the application of the eddy current method to the metal layer thickness measurement or endpoint detection. In this paper, an in situ measurement system, which is independently developed by using the eddy current method, is applied to the actual Cu CMP process. A series of experiments are done for further analyzing the dynamic response characteristic of the output signal within different thickness variation ranges. In this study, the voltage difference of the output signal is used to represent the thickness of the Cu layer, and we can extract the voltage difference variations from the output signal fast by using the proposed data processing algorithm. The results show that the voltage difference decreases as thickness decreases in the conventional measurement range and the sensitivity increases at the same time. However, it is also found that there exists a thickness threshold, and the correlation is negative, when the thickness is more than the threshold. Furthermore, it is possible that the in situ measurement system can be used within a larger Cu layer thickness variation range by creating two calibration tables.

Original languageEnglish
Article number115103
JournalReview of Scientific Instruments
Volume88
Issue number11
DOIs
Publication statusPublished - 1 Nov 2017
Externally publishedYes

Fingerprint

Dive into the research topics of 'Signal processing and analysis for copper layer thickness measurement within a large variation range in the CMP process'. Together they form a unique fingerprint.

Cite this