TY - JOUR
T1 - Reliability prediction with MTOL
AU - Bernstein, Joseph B.
AU - Bensoussan, Alain
AU - Bender, Emmanuel
N1 - Publisher Copyright:
© 2016 Elsevier Ltd
PY - 2017/1/1
Y1 - 2017/1/1
N2 - Here, we develop a comprehensive reliability prediction of FPGA devices from data motivated by physics of failure. The Multiple Temperature Operational Life (MTOL) testing method was used to calculate the failure in time (FIT) of 3 different mechanisms on both 45 nm and 28 nm technologies. We confirmed that there is significant hot carrier injection (HCI) at sub-zero temperatures in 45 nm technology. Surprisingly, we found that 28 nm exhibits no HCI degradation even with up to 1.6 V on the core. As a result, we show that there is no effect of frequency on the reliability. This means that at 28 nm and possibly smaller technologies, the devices can be de-rated or up-rated based only on the NBTI model and therefore reliability is dependent only on operating Voltage and Temperature with a single activation energy. Notably, the activation energies and voltage acceleration factors for both technologies are remarkably similar. This demonstration shows that, unlike other conventional qualification procedures, the MTOL testing procedure gives a broad description of the reliability from sub-zero to high temperatures. This procedure provides FIT prediction which can be applied to newer technologies, specifically 20 nm and 16 nm and beyond.
AB - Here, we develop a comprehensive reliability prediction of FPGA devices from data motivated by physics of failure. The Multiple Temperature Operational Life (MTOL) testing method was used to calculate the failure in time (FIT) of 3 different mechanisms on both 45 nm and 28 nm technologies. We confirmed that there is significant hot carrier injection (HCI) at sub-zero temperatures in 45 nm technology. Surprisingly, we found that 28 nm exhibits no HCI degradation even with up to 1.6 V on the core. As a result, we show that there is no effect of frequency on the reliability. This means that at 28 nm and possibly smaller technologies, the devices can be de-rated or up-rated based only on the NBTI model and therefore reliability is dependent only on operating Voltage and Temperature with a single activation energy. Notably, the activation energies and voltage acceleration factors for both technologies are remarkably similar. This demonstration shows that, unlike other conventional qualification procedures, the MTOL testing procedure gives a broad description of the reliability from sub-zero to high temperatures. This procedure provides FIT prediction which can be applied to newer technologies, specifically 20 nm and 16 nm and beyond.
KW - Accelerated testing
KW - Microelectronics reliability
KW - Ring oscillators
UR - http://www.scopus.com/inward/record.url?scp=85006981296&partnerID=8YFLogxK
U2 - 10.1016/j.microrel.2016.09.005
DO - 10.1016/j.microrel.2016.09.005
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:85006981296
SN - 0026-2714
VL - 68
SP - 91
EP - 97
JO - Microelectronics Reliability
JF - Microelectronics Reliability
ER -