CVD Co film was investigated as an alternative barrier layer to the conventional PVD TaN\Ta in V1\M2 structure for 32nm node. We improved via filling performance and upstream (V1→M2) electromigration (EM) lifetime by more than three times. Excellent step coverage of CVD barrier makes it possible to reduce the thickness of the barrier metal by 30% and to increase the volume of Cu in metal lines. RC delay also reduced with decrease in resistance. Since adhesion at the interface between the barrier-Co and Cu also is strong, migration of Cu atoms is dramatically slowed down. EM in the via is finally deterred due to absence of pre-existing voids, consequently lifetime increases. This CVD Co process is expected to be beneficial for the next technology generation beyond 20nm node.
Pre-existing void effect during electromigration in a sub-30nm wide Cu interconnect was observed. Two types of void are intentionally produced in a single damascene interconnect: 1) A void between Cu and capping dielectric layer (center void) is mainly produced from an excessive overhang by depositing a thick seed layer. 2) A void between Cu and barrier metal (side void) is produced from depositing a thin, discontinuous seed layer. Bi-modality was observed in center voided samples. 44% of lines with center voids show stiff resistance rises at high current density and most of them failed shortly after the resistance rise. No stiff resistance rise was observed at lower current density up to 3000 A.U. In side voided samples, no early failures was observed and the failure show no bimodal trend. Change in local current density around the void is expected to be the major factor for the electromigration performance difference between lines with center and side voids. We were able to show that shape and location of the pre-existing void have a significant effect on the reliability of Cu interconnect, and also the void behavior is highly sensitive to current density. [
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.