University of Thessaly (UTH)
Physics-based electromigration analysis and mitigation for long-term reliability of VLSI interconnects
Abstract
dc:descriptionThe technology downscaling of Integrated Circuits (ICs), combined with the demand for increasingly powerful systems, is leading to highly complex circuits. This increased complexity has brought long-term reliability challenges to the forefront for the current and future generation of ICs, particularly for critical applications such as automotive and aerospace, where even minor failures can have serious consequences. Over time, long-term reliability issues can degrade IC performance, sometimes resulting in malfunctions or even destructive failures like open circuits. Among these long-term reliability issues, Electromigration (EM) in power grids and interconnects requires particular attention. These elements form critical subsystems responsible for ensuring stable power delivery across the IC, but they are especially vulnerable to EM due to the high current densities they sustain. EM occurs when positive copper ions migrate from the cathode to the anode along the current direction, causing material density changes that result in copper deficits and accumulations. Over time, this increases conductor resistance and can lead to open circuits at critical points. As a result, the assessment of this problem is a fundamental part of the final (sign-off) checks during IC production, even though the related methods are empirical, increasingly inaccurate, and often incorrect, as they were designed for older-generation circuits.This dissertation focuses on advancing the analysis of the EM problem by integrating modern physics-based models with efficient algorithms. It addresses limitations in existing techniques by introducing a novel methodology for comprehensive pre-void and post-void EM analysis, leveraging the coefficient matrix structure for accurate determination of void nucleation time and location. Additionally, it presents a numerical method that employs system-theoretic techniques to efficiently handle large-scale interconnect models. An advanced EM mitigation strategy is developed using computational intelligence algorithms to optimize wire sizing, thereby enhancing chip reliability while reducing overall area consumption. Finally, the research comes to an end with the development of PROTON, an open-source tool that integrates the analysis methodologies developed herein into industrial design workflows, providing a comprehensive solution to EM power grid assessment.
Degree
thesis:*- Grantor dc:publisher
- University of Thessaly (UTH)
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Axelou, Olympia-Kerasia
- Αξελού, Ολυμπία-Κερασία
Subjects
dc:subject × 26- Ηλεκτρομετανάστευση
- Μακροχρόνια Αξιοπιστία
- Αγωγοί ολοκληρωμένων κυκλωμάτων
- Ανάλυση ιδιοτίμων
- Μείωση τάξης μοντέλων
- Εξίσωση Korhonen
- Βελτιστοποίηση δικτύων τροφοδοσίας
- Electromigration
- Longterm reliability
- VLSI interconnects
- Eigenvalue decomposition
- Model order reduction
- Korhonen's PDE
- Power grid optimization
- Μαθηματικά
- Φυσικές Επιστήμες
- Μοντελοποίηση και Προσομοίωση
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Υπολογιστές, Υλικό (hardware) και Αρχιτεκτονική
- Mathematics
- Natural Sciences
- Modeling and Simulation
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Computer science, Hardware and Architecture
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/58053