Search

We demonstrate spin-transfer torque magnetoresistive random access memory (STT-MRAM) arrays achieving 2.8e-10 write error rate (WER) performance at 3 ns write duration at a magnetic tunnel junction (MTJ) diameter of 40 nm. The bit-to-bit distribution of the write voltage at a WER of 1e-6 is characterized by a relative standard deviation of 3.7% for W0 and 4.5% for W1, sufficient to meet the write voltage distribution requirement for last-level cache (LLC) applications at 1x nm nodes.

Magnetic tunnel junctions operating in the superparamagnetic regime are promising devices in the field of probabilistic computing, which is suitable for applications like high-dimensional optimization or sampling problems. Further, random number generation is of interest in the field of cryptography. For such applications, a device's uncorrelated fluctuation time-scale can determine the effective system speed. It has been theoretically proposed that a magnetic tunnel junction designed to have only easy-plane anisotropy provides fluctuation rates determined by its easy-plane anisotropy field and can perform on a nanosecond or faster time-scale as measured by its magnetoresistance's autocorrelation in time. Here, we provide experimental evidence of nanosecond scale fluctuations in a circular-shaped easy-plane magnetic tunnel junction, consistent with finite-temperature coupled macrospin simulation results and prior theoretical expectations. We further assess the degree of stochasticity of such a signal.

We report reliable 5 ns switching of spin-transfer torque magnetoresistive random-access memory devices of nominal size 43 nm and a resistance area product of 11 Ω·µm2. We measured 256 devices with a 100% write-error-rate (WER) yield at a WER floor of $10^{-6}$ and a steep WER slope as a function of voltage. A single device had a WER less than $10^{-10}$ for 5 ns write pulses. We show promising 3 ns switching performance, with a 94% WER yield at a $10^{-6}$ WER floor, for 64 devices of nominal size 50 nm.

We report for the first time reliable 2 ns switching of spin-transfer torque magneto-resistive random access memory (STT-MRAM) devices by demonstrating 100% write-error-rate (WER) yield at 1e-6 write-error floor of 254 devices with tight distributions and steep WER slope at a nominal size of 49 nm. A single device was demonstrated with 2 ns write pulses to have less than 1e-11 write-error rate, limited only by test time. We further demonstrate reliable 3 ns switching performance, with 99% WER yield at 1e-6 write-error floor of 256 devices with nominal size of 43 nm and a single device with less than 1e-11 write-error rate with a completely different free layer materials design. These two different free layer materials designs address one of the major remaining challenges for STT-MRAM to replace SRAM for last level cache (LLC) applications.

The dependence of the write-error rate (WER) on the applied write voltage, write pulse width, and device size was examined in individual devices of a spin-transfer torque (STT) magnetic random-access memory (MRAM) 4 kbit chip. We present 10 ns switching data at the ${10^{ - 6}}$ error level for 655 devices, ranging in diameter from 50 nm to 11 nm, to make a statistically significant demonstration that a specific magnetic tunnel junction stack with perpendicular magnetic anisotropy is capable of delivering good write performance in junction diameters range from 50 to 11 nm. Furthermore, write-error-rate data on one 11 nm device down to an error rate of $7{\times}10^{ - 10}$ was demonstrated at 10 ns with a write current of $7.5\;\upmu{\rm A}$ , corresponding to a record low switching energy below 100 fJ.

We report the impact of four key parameters on switching efficiency of STT-MRAM devices with perpendicular magnetic anisotropy: device size, device resistance-area product (RA), blanket film Gilbert damping constant (a), and process temperature. Performance degradation observed in 400°C-processed devices was eliminated by optimizing the perpendicular magnetic tunnel junction (p-MTJ) materials. Furthermore, 400°C-compatible double MTJs were developed for the first time and showed 1.5x improvement in switching efficiency compared to single MTJs with identical free layers.

Barrier interface condition is critical for spin-polarized tunneling and spin-transfer torque switching in CoFeB∣MgO∣CoFeB-based magnetic tunnel junctions. The differential tunnel conductance gV contains information on CoFeB’s magnetic properties at tunnel interfaces. Experimentally, we find gV to follow a “cross-normalization” relationship between the parallel and antiparallel alignments. This we show originates from the leading order spin-flip scatter terms related to CoFeB interface magnetic properties such as its exchange-stiffness. By connecting the observable gV slopes to electrode-specific spin-flip scatter rates, we obtain an efficient measurement for mass-screening of junctions for interface magnetic differences. This provides valuable information for device and fabrication process optimization.Barrier interface condition is critical for spin-polarized tunneling and spin-transfer torque switching in CoFeB∣MgO∣CoFeB-based magnetic tunnel junctions. The differential tunnel conductance gV contains information on CoFeB’s magnetic properties at tunnel interfaces. Experimentally, we find gV to follow a “cross-normalization” relationship between the parallel and antiparallel alignments. This we show originates from the leading order spin-flip scatter terms related to CoFeB interface magnetic properties such as its exchange-stiffness. By connecting the observable gV slopes to electrode-specific spin-flip scatter rates, we obtain an efficient measurement for mass-screening of junctions for interface magnetic differences. This provides valuable information for device and fabrication process optimization.

With the promise of nonvolatility, practically infinite write endurance, and short read and write times, magnetic tunnel junction magnetic random access memory could become a future mainstream memory technology.

Exchange biased magnetic tunnel junction (MTJ) structures are shown to have useful properties for forming magnetic memory storage elements in a novel cross-point architecture. MTJ elements have been developed which exhibit very large magnetoresistive (MR) values exceeding 40% at room temperature, with specific resistance values ranging down to as little as ∼60 Ω(μm)2, and with MR values enhanced by moderate thermal treatments. Large MR values are observed in magnetic elements with areas as small as 0.17 (μm)2. The magnetic field dependent current–voltage characteristics of an MTJ element integrated with a silicon diode are analyzed to extract the MR properties of the MTJ element itself.

We have measured and simulated the dynamics of magnetization reversal in 5 nm by 0.8 by 1.6 $\ensuremath{\mu}\mathrm{m}$ ${\mathrm{Ni}}_{60}{\mathrm{Fe}}_{40}$ thin films. The films measured form the upper electrode of a spin-polarized tunnel junction so that the magnetization direction of the film can be probed by measuring the tunneling resistance of the junction. When a magnetic field pulse is applied, the time to switch the film magnetization changes from greater than 10 ns to less than 500 ps as the pulse amplitude is increased from the coercive field to 10 mT and beyond. We have simulated these transitions using micromagnetic modeling of the exact experimental conditions. The simulations agree well with the experimental measurements.

The use of lithography and electroplating to fabricate variable-reluctance, nearly planar, integrated minimotors with 6-mm-diameter rotors on silicon wafers is described. The motors consist of six electroplated Permalloy® horseshoe-shaped cores that surround the rotor. Copper coils are formed around each core. The Permalloy and copper electroplating baths, electroplating seed layers, and through-mask plating techniques are similar to those used to fabricate inductive thin-film heads. High-aspect-ratio optical lithography or X-ray lithography was used to form the various resist layers. The rotors were fabricated separately, released from the substrate, and then slipped onto the shaft, which was plated as part of the stator fabrication process. The fabrication processes for stator and rotor are described in this paper, and initial minimotor operation data are presented.

Using a wide-field Kerr microscope, we have studied the magnetic domain structures of epitaxial and polycrystalline La0.67Sr0.33MnO3 thin films as well as a film having thermally induced 〈110〉 microcracks. The epitaxial film on a (001) SrTiO3 substrate has different magnetic domain behaviors for in-plane fields applied along the 〈100〉 and 〈110〉 directions. Magnetic domain orientation and contrast suggest a biaxial magnetic anisotropy with 〈110〉 easy axes. Defects such as microcracks and grain boundaries have a strong perturbing effect on the local magnetization and can lead to an enhanced and controllable spin-dependent scattering.

We examine the sweep-rate dependence of magnetic switching field, Hs, in submicron magnetic tunnel junctions where shape anisotropy is dominant. Experimental data support the use of a single-domain thermal activation model for description of activated magnetic reversal in junctions 0.2 by 0.5 μm or less in size. A scaling law is obtained for the thermal activation energy which varies as the cube of junction size.

We present measurements of the magnitude of Neel “orange-peel” coupling due to interface roughness in a series of magnetic tunneling junction devices. Results from magnetometry and transport measurements are shown to be in good agreement with the theoretical model of Neel. In addition, we have used transmission electron microscopy to directly probe the sample interface roughness and obtain results consistent with the values obtained by magnetometry and transport methods.

We have studied the switching properties of micron-scale magnetic tunnel junctions in two-dimensional magnetic fields. We present data on interlayer magnetic coupling for multiple samples. We interpret these data as the sum of a magnetostatic and a Neel coupling contribution. The data are presented as functions of layer structure. In addition, we have extracted information about interface roughness. We have also studied the area of switching critical curves as a function of device geometry.

The magnetic switching behavior of micron-size magnetic tunnel junctions has been studied in two-dimensional magnetic fields. By measuring junction resistance, we obtain information about the magnetization state of the free ferromagnetic layer. Magnetic properties of this layer are explored using the Stoner–Wohlfarth rotational model as a starting point. We use geometric parameters of the critical curves to obtain information about interlayer coupling and domain structure effects in the free layer.

We have studied bias and magnetic field dependence of voltage noise in metallic magnetic tunnel junctions with areal dimensions on the order of 1 μm. We generally observe noise with Gaussian amplitude distribution and pure 1/f power spectra at low frequencies. The 1/f noise scales with bias voltage as V2. Two kinds of deviations from this low frequency behavior have been observed. One is at fixed magnetic field when the junction bias reaches above a critical value, the other occurs at a fixed bias when the external magnetic field brings the sample to certain magnetic configurations. In both cases the noise spectra become dominated by Lorentzian noise and in both cases we have observed two level fluctuators in the time domain. We attribute the bias dependent noise to charge traps in the tunnel barrier. The field dependent noise is associated with the switching of the magnetization direction of portions of the top electrode, which we believe to be reversible.

We show that shape anisotropy can be used to control the response characteristics of magnetic tunnel junctions. By varying the junction shape, the resistance versus field curve was made to vary from a nonhysteretic linear curve with a high-field sensitivity (0.3%/Oe) to a hysteretic response curve with high squareness.

Although considerable work has been done investigating the properties of arrays of magnetic elements, there have been few investigations on the reverse geometry, i.e., an array of nonmagnetic regions defined within a magnetic thin film. The 10 Hz BH loops, 10–500 MHz permeability spectra, and domain patterns of homogeneous, single layer 100 nm radio frequency (rf) sputtered Ni81Fe19 thin films with arrays of 23-, 50-, and 100-μm-diam holes defined by laser ablation were measured. The holes were defined in a grid along the hard and easy axes of the sample. Letting (x,y) represent, respectively, the hole spacing parallel to the easy and hard axes, the point to point spacing of the ablated circular regions was varied from (5, 2 mm) to (0.1, 0.1 mm).

Infroducfion: Small-size, highefficiency motors, fabricated by low cost batch fabrication techniques, are expected to have application in storage, automotive, medical and other fields. We describe a new approach to fabrication of an integrated experimental variable reluctance magnetic mini-motor, based on lithography and electroplating techniques, similar to those used in inductive head fabrication. Strumre Fabrication: The mini-motor consists of six, nearly planar, horseshoe-shaped electromagnets which surround a rotor. The stator diameter is 17 mm, rotor radius 2.976 mm, core width 0.7 mm, coil width 60 pm, and there are 120 turns per core., The motor thickness is -5OOpm, with a substrate approximately 1.0 mm. The fabrication sequence consists of a series of lithography, electroplating, seed layer removal, dielectric fill, and planarization steps. The process for the stator begins by sputterdepositing a Cu seed layer on an oxidized Si wafer, lithographically forming the plate-thru mask for the bottom part of the windings, and electroplating Cu through the mask. After removing the seed layer, a dielectric is applied across the entire structure, which is then planarized, and a pattemed dielectric is formed on top of the Cu to insulate the windings from the magnetic core. The Permalloy core is plated, and the top of the windings are formed by similar process steps. This simplifies the fabrication process. The vias that complete each tun of the coil are formed of PennaIloy during the core plating step. The rotor is fabricated by thru-mask plating of Permalloy on a separate substrate, detaching it, using a sacrificial layer, and then slipping it onto the pin. 'The pin is formed during the same lithography and plating steps thar created the core and the winding. Seed Layers: The seed layers for Cu and Permalloy plating respectively are of sputtered Ti/Cu (I 000-3000A) and of 1000-2OOOA sputtered Permalloy (80120) NiFe).

Wide-field magneto-optic Kerr observation techniques are described for the imaging of magnetic recording heads and other small magnetic devices. The imaging problems encountered for these samples are quite different from those for continuous films and lead to a different set of experimental approaches. Observation techniques that preserve the magnetic state of the sample allow for the implementation of new image acquisition algorithms based on multiple additions and subtractions that progressively build up a weak Kerr contrast even under unfavorable experimental conditions. Examples of results are presented. >

We observe a strong correlation between magnetization relaxation and electrical resistivity in thin Permalloy $({\mathrm{Ni}}_{81}{\mathrm{Fe}}_{19},$ ``Py'') films. Electron scattering rates in the films were affected by varying film thickness and deposition conditions. This shows that the magnetization relaxation mechanism is analogous to ``bulk'' relaxation, where phonon scattering in bulk is replaced by surface and defect scattering in thin films. Another interesting finding is the increased magnetization damping with Pt layers adjacent to the Py films. This is attributed to the strong spin-orbit coupling in Pt, resulting in spin-flip scattering of electrons that enter from the Py.

The authors report on a newly discovered domain configuration, observed with a magnetooptical microscope, in narrow stripes of NiCoFe, CoFeCu and NiFe. The stripes-1 to 25 mu m wide and 0.5 to 2.0 mu m thick-are used as a model for a narrow-track pole of a recording head. The new domain configuration, with its magnetization oriented mostly in a plane perpendicular to the long dimension of the stripe, suggests the possibility of obtaining a large high-frequency susceptibility even in recording head poles as a narrow as 1 mu m. >

We have measured the rate of thermally assisted magnetization reversal of submicron-sized magnetic thin films. For fields H just less than the zero-temperature switching field H(C), the probability of reversal, P(exp)(s)(t), increases for short times t, achieves a maximum value, and then decreases exponentially. Micromagnetic simulations exhibit the same behavior and show that the reversal proceeds through the annihilation of two domain walls that move from opposite sides of the sample. The behavior of P(exp)(s)(t) can be understood through a simple "energy-ladder" model of thermal activation.

The role of grain boundaries in the magnetoresistance (MR) properties of the manganites has been investigated by comparing the properties of epitaxial and polycrystalline films of ${\mathrm{La}}_{0.67}{D}_{0.33}\mathrm{Mn}{\mathrm{O}}_{3\ensuremath{-}\ensuremath{\delta}} (D=\mathrm{C}\mathrm{a},\phantom{\rule{0ex}{0ex}}\mathrm{S}\mathrm{r},\phantom{\rule{0ex}{0ex}}\mathrm{o}\mathrm{r}\phantom{\rule{0ex}{0ex}}\mathrm{v}\mathrm{a}\mathrm{c}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{s})$. While the MR in the epitaxial films is strongly peaked near the ferromagnetic transition temperature and is very small at low temperatures, the polycrystalline films show large MR over a wide temperature range down to 5 K. The results are explained in terms of switching of magnetic domains in the grains and disorder-induced canting of Mn spins in the grain-boundary region.

The cause of the readback waveform distortion in polycrystalline ferrite metal-in-gap (MIG) heads was investigated by using a magnetic force microscope (MFM) and a polarized optical microscope capable of microellipsometry for grain contrast and Kerr microscopy for domain contrast. It was revealed that, for high-asymmetry MIG heads, relatively large leakage of magnetic flux associated with complex multiple domains exists at remanence in the leading-side ferrite surface adjacent to the gap. The strengths of the magnetic flux and the magnetic domain states change irreproducibly when coil current is repeatedly applied. Such a remanent flux associated with the complex remanent state of magnetization distribution in the ferrite can cause an inhomogeneous flux coupling between the head and the recorded track, resulting in readback waveform asymmetry. >

The magnetic domain structure at the media-facing surfaces of the single-crystal ferrite MIG heads is investigated using the magnetooptical Kerr effect. Domain behavior depends strongly on the stress state and surface treatment of the head. The pole tips display coercive zig-zag domain walls, which are nucleated at the gap by write current pulses. For the samples studied, nucleation and motion of these walls are different between heads which exhibit small secondary pulses in the read-back signal and heads which do not. >

A novel magnetooptic imaging technique was used to investigate the dynamics of magnetization response in thin-film head yokes. Completed head devices were excited with a sinusoidal current applied to the integrated coils. The amplitude (20 to 40 mA pp) and frequency (1 to 50 MHz) of excitation were chosen to simulate the write process. Pulsed laser illumination permitted stroboscopic observation of domain-wall and flux-flow dynamics with a time resolution limited only by the 5-ns pulse width (full width at half maximum). Results suggest that high-frequency write performance is degraded by two mechanisms not considered previously for thin-film heads: first, 180 degrees walls appear to impede the flux-flow across the plane of the wall and, second, the inhomogeneous rotational magnetization response observed is known to be much slower than the typically assumed coherent rotation. The effect of NiFe composition on dynamics was also investigated by comparing responses of two heads, one having positive and the other negative magnetostriction. Flux in the head with positive magnetostriction flows in a constricted path along yoke edges, rather than at its center, which is consistent with the significantly lower efficiency measured electrically for this head. >

A thermally assisted magnetoresistive random access memory device (TAS-MRAM) with reduced power for reading and writing; the memory device comprising a tunnel barrier 14 sandwiched between a ferromagnetic sense layer 16 and a ferromagnetic storage layer 12. An antiferromagnetic pinning layer 30 is disposed adjacent to the ferromagnetic storage layer 12. The pinning layer 30 pins a magnetic moment of the storage layer until heating is applied. Either or both of the storage and sense ferromagnetic layers includes a non-magnetic material to reduce the magnetization of the respective layers. The reduction in the storage layer magnetization and sense layer magnetization reduces the magnetostatic interaction between the storage layer and sense layer, resulting in less read/write power. The ferromagnetic materials in the sense and storage layers may include at least one of Co, Fe, Ni, and any alloy including Co, Fe, Ni, whilst the non-magnetic material includes at least one of Ta, Ti, Hf, Cr, Nb, Mo, Zr and any alloy containing Ta, Ti, Hf, Cr, Nb, Mo, Zr. The antiferromagnetic pinning layer may have a diameter less than 250nm based on the reduction in magnetization of at least one of the storage or sense layer. The ferromagnetic storage layer may be formed by sputtering ,chemical vapour (vapor) deposition CVD or physical vapour deposition PVD , and may involve co-sputtering the ferromagnetic and non magnetic material, or forming multi-layers of ferromagnetic and non magnetic material. The ferromagnetic sense layer may also be formed by co-sputtering of ferromagnetic and non magnetic material or forming multilayers of the two materials. An alternative embodiment (figures 7A/B) comprises a tunnel barrier layer 14 sandwiched between a ferromagnetic storage layer 16 and a synthetic antiferromagnetic storage layer 12, which includes a first ferromagnetic storage layer 11 adjacent to the tunnel barrier layer and a non magnetic coupling layer 15 sandwiched between the first ferromagnetic storage layer 11 and a second ferromagnetic storage layer 13. The alternative structure further allows for a relative increase in the thickness of the first ferromagnetic layer 11.

Ferrite MiG heads intended for narrow track (≲10 μm) digital recording were recently investigated in the critical pole‐tip region at the air‐bearing‐surface using micro‐ellipsometry, Kerr microscopy, and electron back‐scatter diffraction from individual grains,1 and using magnetic force microscopy to detect air‐gap remanent fields.2 Comparison of these direct observations with readback‐after‐write waveforms from written test tracks, and consideration of granularity influences on bulk permeability and domain size, indicate that waveform instability and asymmetry from polycrystalline ferrite (PCF) heads would be diminished by suitable size and orientation of the grains.1 The use of single‐crystal ferrite3 (SCF) for advanced laser enhanced etch definition3 of narrow pole MiGs can avoid this type of distortion. However, secondary signals4 often appear as weak pulses separated in time from the main gap pulse. We have associated this effect with a zig‐zag shaped wall seen nucleated and propagated from the pole ...

Laminated Permalloy, with edge‐curling walls replacing closure domains, has been proposed to increase permeability and reduce wall noise in recording. However, in structures meeting the criteria for Slonczewski’s ‘‘easy‐axis’’ state, normal walls often coexist with edge‐curling walls. We have used our laser magneto‐optic microscope to study inductive‐head‐yoke shaped elements of two and four Permalloy layers separated by nonmagnetic, metallic spacers. In the four‐magnetic‐layer sample a state with a single wall, terminating at the edge‐curling regions and lying along the easy‐axis direction, is often observed on the top and bottom layers. Some elements may be driven into an easy‐axis state with no observed domain walls. The two‐magnetic‐layer sample also exhibited simultaneous one‐wall structures on the top and bottom layers. The other stable configuration was a no‐wall state on the top layer and a two‐wall (three‐domain) state on the lower layer. These ‘‘coupled’’ states were exceptionally stable in both...

A magnetooptic microscope was used to observe the time-averaged magnetization distribution in an inductive thin-film head excited by continuous sine waves. Domain activity in yokes driven with sinusoidal currents (1 to 20 MHz) was observed using the Kerr effect at video frame rates (0 to 30 Hz). Thus, the average location and shape of domains in the top yoke of the head could be recorded. It is shown that the domain pattern generally undergoes significant changes in a slow, repeatable evolution. Some changes lead to abrupt conversions of domain states. Although specific behavior varies from head to head, these conversions follow measurable curves having a common trend in the amplitude versus frequency space. Previous work on analysis of head response has not considered this type of dynamic response, although it appears to be common to many magnetic system. In addition to possible response at the excitation frequency, the wall network can also undergo large changes with a time scale much longer than the excitation period. Three possible mechanisms driving the domain conversions are outlined. >

Understanding the magnetic switching behavior in micron and submicron scale specimens is important for a number of applications. In this study, magnetic-tunnel junctions of various sizes and shapes were fabricated and their switching behavior was studied in detail. Using exchange bias to offset the magnetic response of one electrode, the response of the other (free) electrode was determined from measurements of junction resistance. Switching threshold curves were measured by sweeping magnetic fields in both easy and hard direction. Single domain like switching was observed in some of our smallest submicron junctions. The observed behavior was compared with predictions from the Stoner–Wohlfarth rotational model and from numerical calculations.