Logo image
Correct Extrapolation Model for TDDB of STT-MRAM MgO Magnetic Tunnel Junctions
Conference proceeding

Correct Extrapolation Model for TDDB of STT-MRAM MgO Magnetic Tunnel Junctions

J.H. Lim, N. Raghavan, V.B. Naik, J.H. Kwon, K. Yamane, H. Yang, K.H. Lee and K.L. Pey
IEEE International Reliability Physics Symposium proceedings, pp.1-7
03/2019

Abstract

Current density Extrapolation Magnetic tunnel junction (MTJ) Magnetic tunneling Mathematical model MgO Power law model Pulsed voltage endurance Stress Temperature distribution Temperature measurement Time-dependent dielectric breakdown (TDDB)
Magnesium oxide (MgO) has become a dielectric of growing importance, given its use in spin-transfer torque magnetic random access memory (STT-MRAM) devices as a magnetic tunnel junction (MTJ). From recent research studies by several groups, it is apparent that the fundamental physics of time dependent dielectric breakdown (TDDB) in MgO is different from HfO 2 and SiO 2 due to several factors which include ultra-thin dielectric (<1 nm), ferromagnetic material interfaces on both sides, different energetics of the defects and the combination of "very high" current density and "very low" voltage of operation in STT-MRAM technology. These factors have an impact on the TDDB lifetime extrapolation model that can be used for M gO reliability qualification. We analyze the applicability of the commonly proposed (I) ξ-model, (II) 1/ξ model and (III) power law model for front-end dielectrics and assess the suitability of each of them using evidence from large sample size TDDB tests performed on MTJ arrays at various voltage and temperature stress conditions. The most suited extrapolation model along with an underlying physical explanation that supports it will be presented. Our analysis reveals that a combination of the power law and thermochemical ξ-model would best represent MgO dielectric breakdown (BD).

Metrics

1 Record Views

Details

Logo image