Abstract
Reliability modeling and analysis today lacks a robust method to directly validate the lifetime of devices and circuits. As aging mechanisms are usually gradual, i.e., a slow process, conventional aging analysis relies on the extrapolation from a short-term measurement, resulting in unreliable prediction of End of Life (EOL). Such situations are exacerbated at scaled technology nodes at high temperatures where Bias Temperature Instability (BTI), a more gradual and stochastic mechanism, dominates aging compared to Hot Carrier Injection (HCI). To improve the robustness of aging modeling, this work proposes a new approach to adaptively stress the device to EOL in an accelerated and controllable manner. It enables us to monitor the entire process of degradation and validate related analysis tools. The contributions of this paper include: (1) development of a closed-loop test methodology, Adaptive Accelerated Aging (AAA), that effectively accelerates the degradation, (2) application of AAA on 28nm high-k/metal gate (HKMG) devices, proving the feasibility of controllable stress to EOL within 1h, and (3) demonstration of AAA at the circuit level using ring oscillators (ROs).
Accepted Version (
Free)
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call