Search

The exceptional capabilities of the human brain provide inspiration for artificially intelligent hardware that mimics both the function and the structure of neurobiology. In particular, the recent development of nanodevices with biomimetic characteristics promises to enable the development of neuromorphic architectures with exceptional computational efficiency. In this work, we propose biomimetic neurons comprised of domain wall-magnetic tunnel junctions that can be integrated into the first trainable CMOS-free recurrent neural network with biomimetic components. This paper demonstrates the computational effectiveness of this system for benchmark tasks and its superior computational efficiency relative to alternative approaches for recurrent neural networks. [ABSTRACT FROM AUTHOR]

The development of an efficient neuromorphic computing system requires the use of nanodevices that intrinsically emulate the biological behavior of neurons and synapses. While numerous artificial synapses have been shown to store weights in a manner analogous to biological synapses, the challenge of developing an artificial neuron is impeded by the necessity to include leaking, integrating, firing, and lateral inhibition features. In particular, previous proposals for artificial neurons have required the use of external circuits to perform lateral inhibition, thereby decreasing the efficiency of the resulting neuromorphic computing system. This work therefore proposes a leaky integrate-and-fire neuron that intrinsically provides lateral inhibition, without requiring any additional circuitry. The proposed neuron is based on the previously proposed domain-wall magnetic tunnel junction devices, which have been proposed as artificial synapses and experimentally demonstrated for non-volatile logic. Single-neuron micromagnetic simulations are provided that demonstrate the ability of this neuron to implement the required leaking, integrating, and firing. These simulations are then extended to pairs of adjacent neurons to demonstrate, for the first time, lateral inhibition between neighboring artificial neurons. Finally, this intrinsic lateral inhibition is applied to a ten-neuron crossbar structure and trained to identify handwritten digits and shown via direct large-scale micromagnetic simulation for 100 digits to correctly identify the proper signal for 94% of the digits. [ABSTRACT FROM AUTHOR]

The NiMnInSi Heusler alloy family is analyzed, and a self-similarity based method is used to analyze the first-order transition of Ni-Mn Heusler alloys. This method is appropriate to determine magnetic characteristics of the Magnetocaloric Effect providing that there is sufficient separation in their phase transition temperatures. A temperature scaling methodology is used to model the cluster compositions in the mixed-state regions where two stable magnetic states co-exist. The various descriptions and classifications of these transitions, however, are not critical to this analysis. [ABSTRACT FROM AUTHOR]

Inspired by the parallelism and efficiency of the brain, several candidates for artificial synapse devices have been developed for neuromorphic computing, yet a nonlinear and asymmetric synaptic response curve precludes their use for backpropagation, the foundation of modern supervised learning. Spintronic devices—which benefit from high endurance, low power consumption, low latency, and CMOS compatibility—are a promising technology for memory, and domain-wall magnetic tunnel junction (DW-MTJ) devices have been shown to implement synaptic functions such as long-term potentiation and spike-timing dependent plasticity. In this work, we propose a notched DW-MTJ synapse as a candidate for supervised learning. Using micromagnetic simulations at room temperature, we show that notched synapses ensure the non-volatility of the synaptic weight and allow for highly linear, symmetric, and reproducible weight updates using either spin transfer torque (STT) or spin–orbit torque (SOT) mechanisms of DW propagation. We use lookup tables constructed from micromagnetics simulations to model the training of neural networks built with DW-MTJ synapses on both the MNIST and Fashion-MNIST image classification tasks. Accounting for thermal noise and realistic process variations, the DW-MTJ devices achieve classification accuracy close to ideal floating-point updates using both STT and SOT devices at room temperature and at 400 K. Our work establishes the basis for a magnetic artificial synapse that can eventually lead to hardware neural networks with fully spintronic matrix operations implementing machine learning. [ABSTRACT FROM AUTHOR]

There are pressing problems with traditional computing, especially for accomplishing data-intensive and real-time tasks, that motivate the development of in-memory computing devices to both store information and perform computation. Magnetic tunnel junction memory elements can be used for computation by manipulating a domain wall, a transition region between magnetic domains, but the experimental study of such devices has been limited by high current densities and low tunnel magnetoresistance. Here, we study prototypes of three-terminal domain wall-magnetic tunnel junction in-memory computing devices that can address data processing bottlenecks and resolve these challenges by using perpendicular magnetic anisotropy, spin–orbit torque switching, and an optimized lithography process to produce average device tunnel magnetoresistance TMR = 171% and average resistance-area product RA = 29 Ω μ m 2 , close to the RA of the unpatterned film. Device initialization variation in switching voltage is shown to be curtailed to 7%–10% by controlling the domain wall initial position, which we show corresponds to 90%–96% accuracy in a domain wall-magnetic tunnel junction full adder simulation. Repeatability of writing and resetting the device is shown. A circuit shows an inverter operation between two devices, showing that a voltage window is large enough, compared to the variation noise, to repeatably operate a domain wall-magnetic tunnel junction circuit. These results make strides in using magnetic tunnel junctions and domain walls for in-memory and neuromorphic computing applications. [ABSTRACT FROM AUTHOR]

Analog hardware accelerators, which perform computation within a dense memory array, have the potential to overcome the major bottlenecks faced by digital hardware for data-heavy workloads such as deep learning. Exploiting the intrinsic computational advantages of memory arrays, however, has proven to be challenging principally due to the overhead imposed by the peripheral circuitry and due to the non-ideal properties of memory devices that play the role of the synapse. We review the existing implementations of these accelerators for deep supervised learning, organizing our discussion around the different levels of the accelerator design hierarchy, with an emphasis on circuits and architecture. We explore and consolidate the various approaches that have been proposed to address the critical challenges faced by analog accelerators, for both neural network inference and training, and highlight the key design trade-offs underlying these techniques. [ABSTRACT FROM AUTHOR]

In this work the magnetocaloric effect (MCE) and hysteresis losses of Ni45Co5Mn(37-x)In(13+x) alloy (x= 0, 0.4) are studied. Also we propose a new approach in calculating the average entropy change to eliminate the overestimation of the magnetic entropy change calculation using Maxwell's relations which is commonly seen in the indirect MCE calculation in the literature. The attributes of magnetostructural coupling have an important role in the magnetic behavior of these magnetofunctional alloys. This work exhibits that a strong magnetostructural coupling can be achieved over a wide temperature range in this alloy family which results in a higher MCE and relative cooling power (RCP) compared to other Ni-Mn-In based Heusler alloys. It's shown that by increasing the applied field the peak of the entropy change curves shifts to lower temperatures. Additionally, a small variation in the stoichiometry composition broadens the magnetic entropy change curve and shifts the critical temperatures by about 36 K to around room temperature. Furthermore, due to the enhancement in ferromagnetic exchange interactions in austenite, the magnetization change accompanying the martensitic transformation is greatly improved, resulting in a large magnetic entropy change of 31 Jkg-1K-1 under 5T field. The broadened and higher peak of the entropy change curves produced a significantly high relative cooling power in Ni45Co5Mn36.6In13.4 (488 JKg-1) and Ni45Co5Mn37In13 (317 JKg-1) under 5T field. These results are especially promising in finding low-cost, high-performance magnetocaloric material for magnetic refrigeration systems. [ABSTRACT FROM AUTHOR]

The vector properties of a newly developed Preisach-type magnetostriction model are discussed. The model uses a modified version of the Della Torre-Pinzaglia-Cardelli model to compute the irreversible and the reversible components of the magnetization. The magnetostriction can then be simulated by assuming that its magnitude is proportional to the magnetization and its direction is dependent on the magnetization history. The modeling approach is outlined for two types of isotropic media: native and polycrystalline. The preliminary results show excellent agreement with the rotational magnetization measurements for a sample of high-strength steel. [ABSTRACT FROM AUTHOR]

A new model to explore negative remanent magnetization is presented. Recently, a model for magnetic materials that have wasp-waist hysteresis loops was applied to a homogeneous thin film. This model postulated two interacting antiferromagnetically coupled media: one medium with low moment and relatively high coercivity, and one medium with a relatively high moment and low coercivity. This model is applicable to the situation where the "width" of the loop near zero is "negative." The major loop of such a material crosses itself, leading to negative remanent magnetization. The present paper introduces a new concept of a need for a third dimension to add to the usual magnetization versus applied field graph, demonstrates that there is no crossing, and discusses the nature of the losses in these materials. [ABSTRACT FROM AUTHOR]

Heusler alloys feature both conventional and inverse magnetocaloric effects near room temperature as they undergo two different transitions. In this paper, new data are presented and analyzed and a new mechanism to explain the complex hysteretic behavior of a Ni48Co2Mn35In13Ga2 Heusler alloy is developed. This mechanism explains isothermal loops near room temperature. The various descriptions and classifications of these transitions, however, is not critical to this analysis. [ABSTRACT FROM AUTHOR]

Nanoparticles provide a system where the nonequilibrium dynamics of magnons can be engineered, varied, and studied over many orders of magnitude. This paper describes the conditions for quasiequilibrium phases with either more or less magnons than predicted by Bloch’s law, including, if certain conditions hold, a Bose–Einstein condensate phase whose existence should strongly depend on system size. [ABSTRACT FROM AUTHOR]

Evidence has recently been given for the Bose-Einstein condensation (BEC) of magnons in nanostructures. This behavior occurs for nanoparticles some tens of nanometers in size (having some millions of atoms in a nanoparticle). Here we explore the effect of particle size on the BEC critical temperature and the blocking temperature. These considerations lead to a proposed phase diagram for magnetic behavior in nanoparticles, with distinct BEC, superparamagnetic, and ferromagnetic phases. [ABSTRACT FROM AUTHOR]

A systematic study of the magnetocaloric effect of a Ni51Mn33.4In15.6 Heusler alloy converted to nanoparticles via high energy ball-milling technique in the temperature range of 270 to 310 K has been performed. The properties of the particles were characterized by x-ray diffraction, electron microscopy, and magnetometer techniques. Isothermal magnetic field variation of magnetization exhibits field hysteresis in bulk Ni51Mn33.4In15.6 alloy across the martensitic transition which significantly lessened in the nanoparticles. The magnetocaloric effects of the bulk and nanoparticle samples were measured both with direct method, through our state of the art direct test bed apparatus with controllability over the applied fields and temperatures, as well as an indirect method through Maxwell and thermodynamic equations. In direct measurements, nanoparticle sample’s critical temperature decreased by 6 K, but its magnetocaloric effect enhanced by 17% over the bulk counterpart. Additionally, when comparing the direct and indirect magnetocaloric curves, the direct method showed 14% less adiabatic temperature change in the bulk and 5% less adiabatic temperature change in the nanostructured sample. [ABSTRACT FROM AUTHOR]

It is well known that the energy barrier for magnetization reversal, EB, varies quadratically with the magnetic field for the Stoner-Wohlfarth model. However, the enthalpy H (switching energy) required to reverse the magnetization is the sum of EB and the work Wf done by the Néel fluctuation field. The sum of these two terms gives an enthalpy of reversal. If the fluctuation field is uniaxial and parallel to the holding field, then the enthalpy barrier for switching is linear in the holding field. The linearity is consistent with certain experimental results. If the fluctuation field is anisotropic, then for the same size field, the probability of the particle’s switching will depend upon the direction of that field. The paper discusses the holding field variation of the energy barrier for different fluctuation field directions. [ABSTRACT FROM AUTHOR]

In order to determine the accuracy of an identification method of the fluctuation field, artificial data were generated. A major hysteresis loop, and its corresponding anhysteretic magnetization curve, was simulated and used to create thermal aftereffect curves as a function of a fluctuation field. Since the curves were idealized with known parameters, they could be used to test the method. In the Preisach-Arrhenius model, it was shown that the fluctuation field could be computed as the slope of the translation, in log time, of the aftereffect curves as a function of the holding fields. Gaussian noise was introduced into the curves to test the robustness of the method in the presence of experimental error. The limitations of the identification method are also discussed. [ABSTRACT FROM AUTHOR]

Wasp-waist and pot-belly hysteresis loops have been observed in many materials. When only the major loop is reported, the results are insufficient to establish which processes are involved. We present two models for wasp-waist materials that produce virtually indistinguishable major loops, but show that first-order reversal curves can be used to separate the effects. In the simplest model, we take a soft magnetic material and a hard material and exchange couple them. When the exchange is positive, the loop is conventional. However, for negative (antiferromagnetic) exchange, the wasp-waist loop is obtained. Negative coupling of two materials with different switching field distributions leads to pot-bellied loops. [ABSTRACT FROM AUTHOR]

Measurements of the fluctuation field of a perpendicular Co/Pt multilayer medium was carried out from 4 K to room temperature, using a procedure that for each holding field matches, in log-time, the aftereffect curve with the major loop. This is done by translation and scaling using the entire curve. The fluctuation field at each temperature is obtained from the slope of the required translation as a function of the holding field. The fluctuation field is directly proportional to the absolute temperature below the Bose condensation temperature, TBE, has a sharp peak at the TBE, then decreases until it approaches a constant value asymptotically from below for higher temperatures. This paper presents both the technique used to obtain this data and a theory of this magnetic aftereffect. [ABSTRACT FROM AUTHOR]

Bimodal magnetic behavior has been observed in a (0.3-nm Co/1.2-nm Pt)[SUB15] multilayer film. In that) film, the Kerr rotation and ellipticity exhibit magneto-optic hysteresis loops with two distinct field-dependent steps, so that there are four stable magnetization states. The magnetic aftereffects of the two distinct types of decay at the two sets of holding fields have been analyzed. Thus, it is possible to measure and calculate the two types of decay individually. We show that the shift in log time of the aftereffect curves for both the sets are linear. The fluctuation fields for the two states are calculated using these methods. We also confirm experimentally the conjecture of the Preisach-Arrhenius model that the shape of the aftereffect curve is the same as the major hysteresis curve. This helps in finding the chemical potential. [ABSTRACT FROM AUTHOR]

Nanocomposites have several advantages as a refrigerant for 100-300 K applications compared to the other common methods of assembling a magnetic refrigeration bed, such as a layered thermal bed, or mixing of different magnetic materials. This article discusses the thermodynamics and heat transfer analysis of an ideal and real active magnetic regenerative refrigeration cycle. An algorithm for the choice of optimum parameters is derived. [ABSTRACT FROM AUTHOR]

The temperature variation of the magnetic relaxation process obtained from dc magnetization measurements are investigated in fine Fe[sub 3]O[sub 4] powder in the temperature range of 5 K

Magneto-optical Kerr effect measurements performed on (0.3 nm Co/x Pt)[sub 15] multilayers with x=0.3–2.0 nm showed all samples to exhibit strong perpendicular magnetic anisotropy. Recent work has shown that for compositions near (0.3 nm Co/1.2 nm Pt)[sub 15], both polar Kerr rotation and ellipticity exhibit bimodal reversals in which the major hysteresis loops have two distinct field-dependent steps due to the contributions of two magnetic phases. To better understand this bimodal magnetization phenomenon, Kerr rotation and ellipticity were determined as a function of radial position from the center to the edge of this disk sample. These measurements exhibit a continuous variation from two-step (bimodal) behavior in the center to single-step (unimodal) behavior at the edge. Since these films were deposited via MBE with a heated substrate, it is believed that this phenomenon is related to thermally induced variations in interface effects such as alloying, abruptness, and morphology. © 2000 American Institute of Physics. [ABSTRACT FROM AUTHOR]

Unique properties of the Zr[sub 3](Rh[sub (1-x)]Pd[sub x])[sub 4] alloy system are reported here. Contrary to most metallically bonded systems which are not quick to change magnetic behavior with alloy concentration, this system does. Pure Zr[sub 3]Rh[sub 4] and pure Zr[sub 3]Pd[sub 4] are each nonmagnetic, while in the limited range 0.02

The rate of magnetization change (magnetic after effect) which occurs after the magnetic field applied to a magnetic material is switched suddenly to a new value, is generally assumed to increase when the temperature is increased. Deviation from this temperature behavior, which has been observed in several cases, has been ascribed to changes in pertinent magnetic properties, such as magnetocrystalline anisotropy, activation volume, coercive field, etc., with temperature. We report here on the magnetic after effect temperature dependence of a magnetic recording material with magnetic properties that change very little with temperature over a wide temperature range but which, nevertheless, deviates considerably from a monotonic increase in decay rate with increasing temperature. We postulate that this behavior is explained by replacing the Maxwell–Boltzmann statistics by the Bose–Einstein statistics appropriate for the magnon energy distribution. © 2000 American Institute of Physics. [ABSTRACT FROM AUTHOR]

Focuses on a magnetic aftereffect study of the reversal behavior in a bimodal magneto-optical medium with perpendicular magnetic anisotropy. Experimental procedure; Results and discussion; Conclusion.

Presents information on a study which investigated the magnetization reversal and magneto-optical images of cobalt/platinum multilayer films using Kerr microscopy. Bimodal features; Magnetic images of the multilayer film.

A systematic study of the size effect on the magnetic and structural properties of Y2Fe17< nanoparticles has been performed.We present newdata to explain the enhanced magnetic properties of nanostructured yttrium-iron alloy synthesized through alkalide reduction chemical synthesis. The properties of the particles were characterized by x-ray diffraction, electron microscopy, and magnetometer techniques. As the size of the nanoparticles is reduced, there is an increase in magnetization per unit of applied magnetic field, a decrease in the coercivity and a substantial reduction in hysteresis. [ABSTRACT FROM AUTHOR]

Deals with a study which measured the temperature dependence of the alternating current susceptibility for a copper-nickel alloy near the monel composition as a function of temperature for melt-spun material, both in the as-spun condition and after isothermal anneals. Differences between the curves for the as-spun and annealed materials; Curie temperature estimates.

Reports on the electroplating of alternate layers of cobalt (Co) and copper (Cu) of individual thicknesses from an electrolyte having a low concentration of Cu and a high concentration of Co atoms. Electrodeposition technique used to prepare the Co-Cu composition modulated alloys; Investigation of the effects of surface roughness and layer interactions.

Deals with a study which presented the Wigner-Seitz construction and its use in the analysis of the ternary compounds from an extensive family of alloys having complicated magnetic structures. Information on a magnetic ordering that has been inferred from magnetization; Factor related to the occurrence of a substantial magnetic moment; Proof for the existence of iron atoms.

Investigates the basic properties of the proposed Preisach accommodation properties of recording media on a commercial flexible disk material. Accommodation in Preisach models; Experimental analysis of accommodation in recording media.

Presents a study of the magnetization process of granular cobalt-silver films. Use of the magneto-optical indicator film technique and magnetic force microscopy in the study; Quality characterization of high-temperature superconductors; Analysis of magnetization domains.

Presents a study that investigated artifacts in magnetic resonance imaging (MRI) from metals. Importance of magnetic properties of clips or other surgical materials placed in the human body; Information on the artifacts produced by MRI by material magnetic susceptibility; Description of the nuclear MRI facility.

Examines the effect of nitrogen on the magnetic properties of hard magnets using the Wigner-Seitz construct. Background on the Wigner-Seitz cell; List of magnet materials which were found to have enhanced magnetism upon nitriding; Analysis of the results for αiron.

Presents a study which investigated the long-time magnetic relaxation measurements on a quench melt grown YBCO superconductor. Experimental methods; Results and discussion.

Focuses on a study which presented Monte Carlo simulations of the magnetocaloric effect in superferromagnetic clusters having uniaxial magnetic anisotropy. Background on magnetocaloric effect; Calculations and results; Conclusion.

Deals with a study which focused on the role of magnetic anisotropy in the magnetocaloric effect. Factor considered in the choice of optimizing the magnetocaloric properties of a material; Entropy change associated with applying the magnetic field; Features of the Monte Carlo calculations.

Focuses on the establishment of the existence of a magnetic aftereffect in nickel (Ni)/copper (Cu) multilayered alloys using a vibrating sample magnetometer at room temperature. Thickness of Cu-Ni multilayers; Information on the maximum values of the relaxation rates.

Presents information on a study which compared the results of measurements on single crystals of YBa[sub2]Cu[sub3]O[sub7-x] using magnetically modulated microwave absorption and SQUID magnetometry. Magnetization of the single crystals; Experimental details; Results and discussion; Conclusions.

Presents a study that determined the Barkhausen noise emanating from the surface of thin foils of iron and nickel. Analysis of the autocorrelation and power density spectrum of the noise; Evaluation of the jump correlations of the noise; Examination of the jump amplitude spectrum of the noise.

Focuses on a study which investigated the magnetism of rare-earth and 3d transition metals neodymium[sub2]iron[sub14]boron and neodymium[sub2]iron[sub17]. Correlation between the local site magnetism and whether those sites lie on nets of disclinations; Criterion for choosing candidate alloys with strong 3d moments; Relationship between crystal field anisotropies and the orientation of the major ligand lines.

Studies the magnetic properties of nanograin iron powder. Details on the experiment; Hysteresis loops of the powder samples; Information on the time effect phenomenon in multilayer systems.

Focuses on a study on giant magnetoresistance peaks in CoNiCu/Cu multilayers grown by electrodeposition. Introduction to giant magnetoresistance; Experimental details; Analysis of the high angle x-ray diffraction patterns.