Nanomagnetic and spin-based memories are distinguished for their high data endurance in comparison with their charge-based peers. However, they have drawbacks, such as high write energy and poor scalability due to high write current. In this paper, we apply the straintronics principle that seeks the combination of piezoelectricity and inverse magnetostriction (Villari effect), to design a proof-of-principle 2 Kb nonvolatile magnetic memory in 65 nm CMOS technology. Our simulation results show read-access and write-cycle energies as low as 49 and 143 fJ/b, respectively. At a nominal supply level of 1 V, reading can be performed as fast as 562 MHz. Write error rates <10 −7 and 10 −15 can be obtained at 10 and 5 MHz, respectively. In addition to nonvolatility, ultralow energy per operation, and high performance, our STRs memory has a high storage density with a cell size as small as 0.2 µm 2 .
Strain-mediated magnetization switching in a magnetic tunneling junction (MTJ) by exploiting a combination of piezoelectricity and magnetostriction has been proposed as an energy efficient alternative to spin transfer torque (STT) and field induced magnetization switching methods in MTJ-based magnetic random access memories (MRAM). Theoretical studies have shown the inherent advantages of strain-assisted switching, and the dynamic response of the magnetization has been modeled using the Landau-Lifshitz-Gilbert (LLG) equation. However, an attempt to use LLG for simulating dynamics of individual elements in large-scale simulations of multi-megabyte straintronics MRAM leads to extremely time-consuming calculations. Hence, a compact analytical solution, predicting the flipping delay of the magnetization vector in the nanomagnet under stress, combined with a liberal approximation of the LLG dynamics in the straintronics MTJ, can lead to a simplified model of the device suited for fast large-scale simulations of multi-megabyte straintronics MRAMs. In this work, a tensor-based approach is developed to study the dynamic behavior of the stressed nanomagnet. First, using the developed method, the effect of stress on the switching behavior of the magnetization is investigated to realize the margins between the underdamped and overdamped regimes. The latter helps the designer realize the oscillatory behavior of the magnetization when settling along the minor axis, and the dependency of oscillations on the stress level and the damping factor. Next, a theoretical model to predict the flipping delay of the magnetization vector is developed and tested against LLG-based numerical simulations to confirm the accuracy of findings. Lastly, the obtained delay is incorporated into the approximate solutions of the LLG dynamics, in order to create a compact model to liberally and quickly simulate the magnetization dynamics of the MTJ under stress. Using the developed delay equation, the efficiency of the straintronics switching over the STT method is highlighted by analytically investigating the energy-delay trade-off of both methodologies.
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.