The endurance of MgO-based magnetic tunnel junctions has been studied using a time-dependent dielectric breakdown method. Series of successive electrical pulses of constant duration, varying voltage and varying time interval between successive pulses (t) were applied until electrical breakdown of the tunnel barrier occurred. We show here that two electrical breakdown regimes exist depending on the time interval t between pulses compared to a characteristic escape time 0 of the order of 100ns. For t< 0 , the breakdown is caused by a high average charge trapped in the barrier. For t> 0 , the breakdown is ascribed to large temporal variations of the amount of trapped charges causing alternating stress in the barrier oxide. Between these two regimes, the tunnel junction reaches a very high endurance. PACS numbers:Spin-transfer torque magnetoresistive random access memory (STT-MRAM) is a promising memory technology because of its non-volatility, high speed operation, very large endurance, high density and compatibility with standard CMOS process 1 . As the magnetic tunnel junction (MTJ) size shrinks, the MTJ resistance must remain comparable to the resistance of the selection transistor in a one transistor-one MTJ (1T-1MTJ) design. Hence, a thinner tunnel barrier that does not compromise on reliability is required. Oxide barrier breakdown is one of the key integration and reliability issues for advanced semiconductor memory technology. Despite numerous studies on the tunnel magnetoresistance of MgO-based magnetic tunnel junctions (MTJ), the breakdown mechanism of ultrathin MgO-MTJ has not been thoroughly investigated 1-8 .However, a comprehensive understanding of MTJ reliability issues is essential for the success of STT-MRAM or of other devices based on hybrid CMOS/MTJ technology.In this work, we studied the lifetime of junctions using a time dependent dielectric breakdown (TDDB) 3,4 technique. Studies of time-dependent dielectric breakdown (TDDB) in magnetic tunnel junctions (MTJ) are usually carried out by applying a DC voltage while recording the time to breakdown 3,4 . However, the normal operation conditions of an MTJ in an MRAM device require applying a large number of read/write voltage pulses a few nanoseconds long.This work reports on the breakdown behavior of MgO-based tunnel barriers submitted to successive voltage pulses. We studied their endurance as a function of the time interval between pulses and pulses amplitude and polarity. An earlier study has shown that MgO dielectric breakdown measurements carried under DC voltage and under cumulative pulsed voltage yield equivalent results for pulsewidths longer than 100 ns 7 . In the present study, we therefore chose to work with shorter constant pulse-width of 30ns. A very peculiar behavior was observed characterized by a very large enhancement of the MTJ endurance around an intermediate characteristic time interval 0 between successive pulses of the order of 100ns.We ascribe this behavior to charging/discharging effects within the MgO barrier du...
An investigation of barrier endurance till electrical breakdown in MgO-based magnetic tunnel junctions (MTJs) is presented. Samples were tested under pulsed electrical stress. By studying the effect of delay between successive pulses, an optimum endurance of MTJs is observed for an intermediate value of delay between pulses corresponding to an optimum trade-off between the average density of charge trapped in the barrier and the amplitude of its time-modulation at each voltage pulse. A charge trapping–detrapping model was developed which shows good coherence with experimental results. The influence of the delay between pulses on the trapped charge density in the tunnel barrier and on its time-modulation is discussed. The average density of trapped charges and its time-modulation are, respectively, responsible for a static and dynamic stress within the tunnel barrier, both leading to breakdown. The probability of breakdown of the MTJ for different applied pulse conditions has been evaluated. The expected endurance of the MTJs was deduced depending on the characteristics of the electrical stress in terms of delay, and unipolarity versus bipolarity. The study emphasizes the role of electron trapping/detrapping mechanisms on the tunnel barrier reliability. It also shows that extremely long endurance could be obtained in MTJs by reducing the density of electron trapping sites in the tunnel barrier.
International audienceThe write endurance and the 1/f noise of electrical origin were characterized in CoFeB/MgO/CoFeB magnetic tunnel junctions (MTJ) for spin transfer torque or thermally assisted magnetic random access memories. A statistical study carried out on a set of 60 nominally identical patterned junctions of 200 nm diameter revealed a correlation trend between the electrical 1/f noise power in the unexercised MTJs and the number of write cycles that these MTJs can withstand before electrical breakdown. The junctions showing the largest 1/f noise power before the write endurance test (successive 30 ns pulses of 1.73 V) have the lowest endurance. In contrast, MTJs initially exhibiting lower 1/f noise tend to have a better electrical reliability, i.e., much longer write endurance. This correlation is explained by the presence of electron trapping sites in the MgO barrier and the role of electron trapping/detrapping phenomena in both MTJ reliability and its 1/f electrical noise power. These results suggest that 1/f noise could be used as a predictive characterization of the MTJ endurance
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2025 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.