New breakdown data generation and analytics methodology to address BEOL and mol dielectric TDDB process development and technology qualification challenges

Abstract

Both MOL PC-CA spacer dielectric and BEOL low-k dielectric breakdown data are commonly convoluted with multiple variables induced by process steps such as lithography, etch, CMP, cleaning, and thin film deposition. The traditional method of stressing one DUT per die or multiple DUTs per die, without careful data deconvolution, is incapable of addressing current complex MOL PC-CA and BEOL low-k dielectric breakdown modeling challenges. In this paper, a new big data generation method plus an analytics procedure method is proposed to soundly evaluate both MOL and BEOL dielectric time-dependent-dielectric breakdown data. A new diagnostic reliability concept is for the first time proposed for comprehensive process diagnostics and more accurate reliability failure rate determination.