Search

Spin transfer torque (STT) is one of the most promising switching approaches for magnetic tunnel junction (MTJ) nanopillars to build up innovative nonvolatile memory and logic circuits. It presents low critical current (e.g., < 100\ \muA at 65 nm), simple switching scheme, and fast-speed; however, it suffers from a number of reliability issues like stochastic switching effects, process voltage temperature (PVT) variations, and erroneous reading etc. The mainstream solution is to enlarge the write pulse duration to reduce error rate, which sacrifices the speed and low power advantages. In this paper, we present a new switching circuit for STT memory and logic, allowing “error-free” as the switching operation becomes deterministic benefiting from the self-enabled mechanism. The switching power efficiency can be also improved thanks to a shorter switching duration. By using an accuracy spice model of STT-MTJ and CMOS 65 nm design-kit, mixed simulations have been performed to demonstrate its high-reliable write/read operations and evaluate its potential area, power, and speed performance. [ABSTRACT FROM PUBLISHER]

Recently, ultra-low power circuits based on logic-in magnetic tunnel junction (MTJ) memory structure have been studied thanks to its non-volatility, infinite endurance, high access speed, and easy integration with CMOS process. However, this type of circuit suffers from low reliability both in memory cell and sensing amplifier circuits, which greatly limits its practical applications for logic computation. In this paper, we present a new design of magnetic full adder (MFA) to overcome this issue based on the thermally assisted switching (TAS) MTJ cell and pre-charge sensing amplifier (PCSA) circuit. By using CMOS 65 nm design kit and a precise TAS-MTJ model, mixed simulations have been performed to demonstrate its high reliability keeping low power and small die area. [ABSTRACT FROM PUBLISHER]

The current challenges of spintronic devices are the large power and long latency for magnetic switching. Current-induced domain wall (DW) motion is a new switching mechanism promising low-power, high-density, and high-speed circuits. It was first studied to build “race track” memory, which is considered as one of the most emerging technologies for future stand-alone memory. Based on the 3-D hybrid integration above CMOS circuits and using magnetic tunnel junction (MTJ) as the write/read heads, DW motion can be advantageously extended to logic and embedded memory applications. In this paper, we present the first DW shift register-based lookup-table circuit to build reconfigurable logic, which may nearly halve the die area compared with conventional SRAM-LUT by sharing a number of subcircuits and suggest some new functions such as multicontext configuration and run-time reconfiguration for further performance improvement. By using a DW electrical model and CMOS 65-nm design kit, its performances such as low power and high computing/reconfiguration speed have been simulated or calculated. [ABSTRACT FROM AUTHOR]

The VHSIC hardware description language (VHDL) simulation of domain wall logic allows to validate magnetic circuits functionality early in the design stage. In this paper, we present two VHDL simulation techniques for domain wall (DW) logic. The first one, closer to the physics, uses models of components described at the behavioral level. It allows to study the basic functions and reveals the specificities of DW logic. The second one, closer to the logic, models components with logic gates and delays and it allows very fast simulation of complex circuits. [ABSTRACT FROM AUTHOR]

The article provides information on the efforts of physicists to discover a device that will use only flow of spins. It highlights the research of Tao Yang and colleagues on how to switch magnetization of a particle from electric current to a pure spin current through the demonstration of non-local spin valve. It notes that there are many problems that need to be mastered before a free current tool will be realized.