TY - GEN
T1 - BTI-induced aging under random stress waveforms
T2 - 51st Annual Design Automation Conference, DAC 2014
AU - Sutaria, Ketul
AU - Ramkumar, Athul
AU - Zhu, Rongjun
AU - Rajveev, Renju
AU - Ma, Yao
AU - Cao, Yu
PY - 2014
Y1 - 2014
N2 - The BTI effect, which consists of both stress and recovery phases, poses a unique challenge to long-term aging prediction, because the degradation rate strongly depends on the stress pattern. Previous approaches usually resort to an average, constant stress waveform to simplify the situation. They are efficient, but fail to capture the reality, especially under dynamic voltage scaling (DVS) or in ana-log/mixed signal designs where the stress waveform is much more random. This paper presents a suite of solutions that enable aging simulation under all possible stress conditions. Key contributions include: (1) Compact modeling of BTI when the stress voltage is varying. The results to both reaction-diffusion (RD) and trap-ping/detrapping (TD) mechanisms are derived. (2) Efficient simula-tion under DVS, leveraging the new BTI models; (3) Silicon vali-dation at 45nm and 65nm, at both device and circuit levels. As the results illustrate, it is necessary to combine both RD and TD mech-anisms to accurately predict aging under changing stress voltages. Our proposed work provides a general and comprehensive solution to aging analysis under random stress patterns.
AB - The BTI effect, which consists of both stress and recovery phases, poses a unique challenge to long-term aging prediction, because the degradation rate strongly depends on the stress pattern. Previous approaches usually resort to an average, constant stress waveform to simplify the situation. They are efficient, but fail to capture the reality, especially under dynamic voltage scaling (DVS) or in ana-log/mixed signal designs where the stress waveform is much more random. This paper presents a suite of solutions that enable aging simulation under all possible stress conditions. Key contributions include: (1) Compact modeling of BTI when the stress voltage is varying. The results to both reaction-diffusion (RD) and trap-ping/detrapping (TD) mechanisms are derived. (2) Efficient simula-tion under DVS, leveraging the new BTI models; (3) Silicon vali-dation at 45nm and 65nm, at both device and circuit levels. As the results illustrate, it is necessary to combine both RD and TD mech-anisms to accurately predict aging under changing stress voltages. Our proposed work provides a general and comprehensive solution to aging analysis under random stress patterns.
KW - Bias Temperature In-stability
KW - DVS
KW - Random Stress Waveform
KW - Reliability Prediction
UR - http://www.scopus.com/inward/record.url?scp=84903124430&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84903124430&partnerID=8YFLogxK
U2 - 10.1145/2593069.2593101
DO - 10.1145/2593069.2593101
M3 - Conference contribution
AN - SCOPUS:84903124430
SN - 9781479930173
T3 - Proceedings - Design Automation Conference
BT - DAC 2014 - 51st Design Automation Conference, Conference Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 2 June 2014 through 5 June 2014
ER -