Back to search

University of Waterloo

DFM Techniques for the Detection and Mitigation of Hotspots in Nanometer Technology

Abstract

dc:description.abstract

With the continuous scaling down of dimensions in advanced technology nodes, process variations are getting worse for each new node. Process variations have a large influence on the quality and yield of the designed and manufactured circuits. There is a growing need for fast and efficient techniques to characterize and mitigate the effects of different sources of process variations on the design's performance and yield. In this thesis we have studied the various sources of systematic process variations and their effects on the circuit, and the various methodologies to combat systematic process variation in the design space. We developed abstract and accurate process variability models, that would model systematic intra-die variations. The models convert the variation in process into variation in electrical parameters of devices and hence variation in circuit performance (timing and leakage) without the need for circuit simulation. And as the analysis and mitigation techniques are studied in different levels of the design ow, we proposed a flow for combating the systematic process variation in nano-meter CMOS technology. By calculating the effects of variability on the electrical performance of circuits we can gauge the importance of the accurate analysis and model-driven corrections. We presented an automated framework that allows the integration of circuit analysis with process variability modeling to optimize the computer intense process simulation steps and optimize the usage of variation mitigation techniques. And we used the results obtained from using this framework to develop a relation between layout regularity and resilience of the devices to process variation. We used these findings to develop a novel technique for fast detection of critical failures (hotspots) resulting from process variation. We showed that our approach is superior to other published techniques in both accuracy and predictability. Finally, we presented an automated method for fixing the lithography hotspots. Our method showed success rate of 99% in fixing hotspots.

Degree

thesis:*
Grantor dc:publisher
University of Waterloo
Year dc:date.issued
2015

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Madkour, Kareem

Subjects

dc:subject × 3

Rights

Language dc:language.iso
en

Identifiers

dc:identifier.*
Handle dc:identifier.uri
http://hdl.handle.net/10012/9213

Chain of custody

source
Harvested from
University of Waterloo
Base URL
dspacemainprd01.lib.uwaterloo.ca/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Madkour, Kareem. DFM Techniques for the Detection and Mitigation of Hotspots in Nanometer Technology. University of Waterloo, 2015. http://hdl.handle.net/10012/9213