B. Van de Wiele

MuMax: A new high-performance micromagnetic simulation tool

Abstract We present M u M ax , a general-purpose micromagnetic simulation tool running on graphical processing units (GPUs). M u M ax is designed for high-performance computations and specifically targets large simulations. In that case speedups of over a factor 100 × can be obtained compared to the CPU-based OOMMF program developed at NIST. M u M ax aims to be general and broadly applicable. It solves the classical Landau–Lifshitz equation taking into account the magnetostatic, exchange and anisotropy interactions, thermal effects and spin-transfer torque. Periodic boundary conditions can optionally be imposed. A spatial discretization using finite differences in two or three dimensions can be employed. M u M ax is publicly available as open-source software. It can thus be freely used and extended by community. Due to its high computational performance, M u M ax should open up the possibility of running extensive simulations that would be nearly inaccessible with typical CPU-based simulators.

Generation of propagating backward volume spin waves by phase-sensitive mode conversion in two-dimensional microstructures

We present the generation of propagating backward volume (BV) spin waves in a T shaped Ni81Fe19 microstructure. These waves are created from counterpropagating Damon Eshbach spin waves, which are excited using microstrip antennas. By employing Brillouin light scattering microscopy, we show how the phase relation between the counterpropagating waves determines the mode generated in the center of the structure, and prove its propagation inside the longitudinally magnetized part of the T shaped microstructure. This gives access to the effective generation of backward volume spin waves with full control over the generated transverse mode.

A micromagnetic study of the reversal mechanism in permalloy antidot arrays

A numerical analysis is focused on the influence of patterning and finite-size effects on the hysteresis properties and magnetization reversal of permalloy antidot films with square lattice and square holes. Simulations are performed by solving the Landau-Lifshitz equation. The aim is to explain the relationships between the shape of the hysteresis loop and the different stages of the reversal process. In particular, the switching mechanism is characterized by the nucleation of domain chains that destroy the periodic symmetry in the magnetization present when infinite periodicity is considered. This behavior is strongly influenced by the demagnetizing effects arising both at the film boundaries and at the hole edges.

Fast Semianalytical Time Integration Schemes for the Landau–Lifshitz Equation

Aiming at a micromagnetic model that describes the macroscopic material behavior starting from microstructural features, there is a need for efficient time stepping schemes for the integration of Landau-Lifshitz-equation in material samples with dimensions of order mum. This paper presents two related semianalytical time schemes. The time stepping algorithms are compared with other explicit time stepping schemes. Here, memory efficiency, time efficiency, convergence, and precision are checked. Further, special attention goes to 1) preservation of the magnetization magnitude in each FD cell; 2) a nonincrease of the free energy when applying a constant field; and 3) conservation of the systems free energy in the case of zero damping

Application of the fast multipole method for the evaluation of magnetostatic fields in micromagnetic computations

Micromagnetic simulations are elaborated to describe the magnetic dynamics in ferromagnetic bodies. In these simulations, most of the time is spent on the evaluation of the magnetostatic field in the magnetic material. This paper presents a new numerical finite difference scheme for the evaluation of the magnetostatic field based on the fast multipole method (FMM). The interactions between finite difference cells are described in terms of far and near field interactions. The far field computations are conducted using the spherical harmonic expansion of the magnetostatic field while the near field computations are accelerated using fast Fourier transforms (FFT). The performance of the presented FMM scheme is studied by comparing the scheme with a pure FFT scheme. The FMM scheme is more memory efficient and more flexible then the FFT scheme. It is accurate and still has a good time efficiency.

Current-driven domain wall mobility in polycrystalline Permalloy nanowires: A numerical study

A complete understanding of domain wall motion in magnetic nanowires is required to enable future nanowire based spintronics devices to work reliably. The production process dictates that the samples are polycrystalline. In this contribution, we present a method to investigate the effects of material grains on domain wall motion using the GPU-based micromagnetic software package MuMax3. We use this method to study current-driven vortex domain wall motion in polycrystalline Permalloy nanowires and find that the influence of material grains is fourfold: an extrinsic pinning at low current densities, an increasing effective damping with disorder strength, shifts in the Walker breakdown current density, and the possibility of the vortex core to switch polarity at grain boundaries.

Logic and memory concepts for all-magnetic computing based on transverse domain walls

We introduce a non-volatile digital logic and memory concept in which the binary data is stored in the transverse magnetic domain walls present in in-plane magnetized nanowires with sufficiently small cross sectional dimensions. We assign the digital bit to the two possible orientations of the transverse domain wall. Numerical proofs-of-concept are presented for a NOT-, AND- and OR-gate, a FAN-out as well as a reading and writing device. Contrary to the chirality based vortex domain wall logic gates introduced in Omari and Hayward (2014 Phys. Rev. Appl. 2 044001), the presented concepts remain applicable when miniaturized and are driven by electrical currents, making the technology compatible with the in-plane racetrack memory concept. The individual devices can be easily combined to logic networks working with clock speeds that scale linearly with decreasing design dimensions. This opens opportunities to an all-magnetic computing technology where the digital data is stored and processed under the same magnetic representation.

A numerical approach to incorporate intrinsic material defects in micromagnetic simulations

Spintronics devices like racetrack memory rely on the controlled movement of domain walls in magnetic nanowires. The effects of distributed disorder on this movement have not yet been studied extensively. Defects give rise to a pinning potential that can be characterized in terms of a depth and an interaction range. We investigate how the effects of defects can be realistically introduced in micromagnetic simulations by comparing the properties of the pinning potential to experimental results in the literature. We show that the full 3-dimensional simulations can be replaced by equivalent 2-dimensional ones and propose two approaches to include defects.

Comparison of Finite-Difference and Finite-Element Schemes for Magnetization Processes in 3-D Particles

This paper compares two numerical schemes for the simulation of magnetization dynamics in 3-D particles. The first one is based on the finite-difference computation of the Landau-Lifshitz equation and on the magnetostatic field evaluation by fast Fourier transforms. The second one handles the spatial discretization of the effective field expression by a nodal finite-element method and solves the Poisson equation with a finite-element/boundary-element technique. The convergence of the methods is studied by varying the time and space discretization.

Thermal effects on transverse domain wall dynamics in magnetic nanowires

Magnetic domain walls are proposed as data carriers in future spintronic devices, whose reliability depends on a complete understanding of the domain wall motion. Applications based on an accurate positioning of domain walls are inevitably influenced by thermal fluctuations. In this letter, we present a micromagnetic study of the thermal effects on this motion. As spin-polarized currents are the most used driving mechanism for domain walls, we have included this in our analysis. Our results show that at finite temperatures, the domain wall velocity has a drift and diffusion component, which are in excellent agreement with the theoretical values obtained from a generalized 1D model. The drift and diffusion component are independent of each other in perfect nanowires, and the mean square displacement scales linearly with time and temperature.