Computer simulation of the operation of a magnetic bubble domain propagation circuit modeled after a current-access device

The operation of a bubble-domain straight-line propagation circuit has been simulated successfully. This simulation has been achieved by our approximating the motion of an s = 0 frozen-azimuth bubble placed under a drive fieldH_{Z}(X, Y, T)= -H_{p} cdot cos [2pi(X/R_{X} - n(T)/4)] cdot exp [-(Y/R_{Y})^{2}]. The simulation has been generated from a previously developed numerical scheme to simulate the motion of a bubble, whose domain shape and magnetization structure along its domain wall were variable. The drive field has been modeled after a dual conductor-sheet, current-access propagation structure, which has a bit period RXand a transverse width on the order of2R_{Y}. The entire field contour has been advanced stepwise in the positiveXdirection by an increase of the integern(T), which represents the drive-phase number. The bubble motion has been observed during the first six drive phases to produce operating margin diagrams for drive frequencies of 250 KHz, 796 KHz and 1 MHz. The method of calculation and the results of the simulation are given.     

On spontaneous nucleation in field-accessed bubble devices

The dependence ot the in-plane drive field at which bubble domains spontaneously nucleate in field-accessed bubble devices has been investigated as a function ofH_{k} - 4piM_{s}and of spacer thickness between the bubble film and permalloy propagation elements. The experiments were carried out on amorphous GdCoMo bubble films with T-bar and Y-bar structures. For a given structure and spacer thickness the nucleation field increases linearly withH_{k} - 4piM. Larger spacer thicknesses also lead to increased nucleation fields. A model based on the Stoner-Wohlfarth astroid is compared to these data and found to be useful in explaining the qualitative trends, but to be in poor quantitative agreement. It is concluded that since the drive field required in a device is proportional to4piM_{s}, Q - 1 = (H_{k} - 4piM_{s})/4piM_{s}must be greater than some minimum value for a given device structure and spacer thickness to permit reliable device operation.     

Gap-tolerant half-disk bubble device margins

A simple model is presented which allows accurate prediction of bias margins of gap-tolerant half-disk propagation tracks for bubble domains. After this is verified by comparison with experimental margin data, an "isomargin" plot is derived to show how the margin varies as a function ofWandG, whereWis the minimum linewidth andGis the inter-bar gap. The bias margin is shown to decrease along a fairly straight line which goes to zero whenW + Gequals the runout diameter, i.e., whenW+G approx 1.5 W_{s}, where Wsis the bubble stripwidth or average bubble diameter. This agrees with experiment, and means that the minimum resolvable feature for half-disk type patterns must be less than0.75W_{s}, and probably will not be much larger than0.5W_{s}to0.6W_{s}. It is concluded that, if made with perfect Permalloy, T-bars and half-disks should propagate isolated bubbles equally well. The advantages of half-disks over T-bars are 1) the fatal bar-crossing problem of T-bars with multiple bubbles is avoided, 2) the minimum propagation field is lower than for T-bars, and 3) half-disks seem more tolerant of "bad" (e.g., high-coercivity) Permalloy. Also tabulated are the effects on margins of variations in the device parameters of a representative design, as might be encountered in a fabrication process with finite tolerances. A brief discussion of stop-start margins is given in conclusion.     

Design and characterization of ion-implanted tracks for 16 Mbit/cm2storage density bubble memory devices

The ion-implanted propagation tracks with contiguous disk patterns (CD tracks) have been confirmed to be better for high density propagation tracks (≥ 16 Mbit/cm²) than those with snake patterns (snake tracks), because of less interactions between bubbles on the other side in the same track. The CD tracks with 1.8 µm × 2.0 µm cell size for 0.5 µm bubbles have been evaluated. The large operating bias field margin of 12.4 percent is obtained at the quasi-static operation with rotating field HRof 60 Oe. The minimum rotating field is 40 Oe. Interdigital folded minor loops are proposed and operated. The proposed minor loops are composed of straight propagation tracks connected alternately to relax the in-side turns. The overall operating margin of 8.4 percent (46 Oe) is obtained at HR= 60 Oe. The feasibility of 16 Mbit/cm²storage density bubble memory devices is confirmed.     

New stable state of bubbles containing Bloch lines

A new kind of bubble having two stable states for a bias field HBhas been found in thin garnet films. The bubble becomes smaller with increasing HBand disappears abruptly at some critical fieldH_{C1}. However, it does not collapse atH_{C1}. When HBis lowered, it comes into sight suddenly at another critical fieldH_{C2}. This means that for HBbetweenH_{C1}andH_{C2}the bubble has two stable states, one for a large bubble and the other for an unobservably small bubble. This has been well explained in terms of the stability of bubbles containing a definite number of Bloch lines.     

Hard bubble suppression and controlled state generation of one micron bubbles in ion-implanted garnet films

Hard bubble suppression with proton and neon implantation was achieved in one micron Eu.8Tm2.2Ga.5Fe4.5O12films. The dose and energy required in these one micron bubble films are similar to that required in five micron bubble films, despite the fact that the one micron films have larger uniaxial anisotropy to be overcome with the induced magnetostriction. In addition, controlled state generation (for bubble lattlce device applications) has been demonstrated in the proton implanted samples using two different techniques. The first technique, using a combination of an in-plane field and a critical velocity, is identical to what Hsu has reported with five micron bubbles, except that the in-plane field required in this case is 20 to 40% larger. The second technique, using DC and AC in-plane fields only without any field gradient, has also been shown. In the case of conversion from a bubble that propagates at an angle from the field gradient to one that propagates along the field gradient, a DC in-plane field of 300 Oe is needed. While in the reverse case, an AC in-plane field of 100 Oe or so is necessary.     

Magnetic bubble mass memory

An experimental magnetic bubble mass memory module complete with all control function and detection electronics has been built and operated. The module contains twenty-eight 16 448-bit mass memory chips and operates at a nominal rotating field frequency of 100 kHz. The module has an average access time of 2.7 ms, a read/write cycle time of 5.14 ms, and a data rate of 700 kbit/sec. A read error rate of <1.6 × 10¹²and error-free propagation in excess of 8.4 × 10¹⁵bubble cycles have been demonstrated.     

Bubble lattice formation in thick MnBi platelet by one-point nucleation

The dynamic domain formations in a single crystal MnBi platelet were examined under various external field strengths. Once the reverse domain is nucleated at one point of the saturated plate by pricking or external pulse field, the domain expands with various modes depending on the net field strength acting on the moving wall and forms various patterns. In a thick platelet, a bubble lattice is produced with no bias field, but it is suppressed by the external field; a radial stripe domain pattern or a spider web like pattern is formed depending on the field polarity. The phase velocity of the wall propagation during lattice formation was estimated as about 50m/sec by the pulse field method. This result brought us the important information about the truly complicated nature of the bubble lattice formation, i.e., the sequential process of the dynamic conversion, breakdown of velocity, static conversion, high speed propagation, and sharp deflection.     

Magnetic bubble device scaling and density limits

Scaling of magnetic bubble devices to smaller bubble sizes and higher density is considered. Drive field requirements, materials requirements, fabrication requirements, current requirements, and detector signal-to-noise ratio are all calculated as a function of bubble size and related to practical limits imposed by bubble materials, fabrication techniques, and electromigration limits. It is concluded that "conventional" bubble devices using Permalloy bars can be made practical with 1-μm bubble domains (storage densitysim6 times 10^{6}bits/cm²). Although it may be possible to extend these Permalloy bar devices to even smaller bubbles, it seems more likely that other bubble devices such as contiguous disk devices or bubble lattice devices will in fact be used for densities greater than 6 × 10⁶bits/cm².     

Stripe-domain dynamics in bubble-domain circuits

It is widely known that bubble domains can exist when the bias field is between the elliptic instability (runout) field Heand the bubble collapse field Hbc. Values for Heand Hbcwere calculated by Thiele. It is not widely recognized that long stripe domains can also exist in the bottom 20% of this range. Stripes are stable up to the "stripe contraction" field, Hsc, which is about 0.02 M/μ0above Hefor thickness over intrinsic length,h/l, values from 4 to 10. (Hbcis about 0.10 M/μ0above He.) Values for Hscwere calculated by Kooy and Enz of Philips Research Labs. in 1960, although the importance of their result to bubble-domain devices was not apparent at the time. The velocity of the domain tip during stripout and contraction such as occurs during domain detection or transfer is important for calculating maximum circuit speeds. This is given byV = cG (H-H_{sc}), whereHis the field (bias field plus local field from currents or Permalloy elements),Gis the wall mobility, andcis a constant approximately equal to 0.5.     

Effects of finite anisotropy parameter Q in the determination of magnetic bubble material parameters

Improved formulas are obtained to provide more accurate determination of magnetic bubble material parameters such as material length(l)and saturation magnetization(M_{s})based on the measurements of domain stripe width and bubble collapse field. The improvement takes into consideration the effects of finite uniaxial anisotropy parameter(Q)in the formulas. These improved formulas are based on the numerical results obtained from micromagnetic domain calculations of periodic stripes and of an isolated axisymmetric bubble.     

New fabrication technique for magnetic bubble circuits with submicron dimensions

A new technique which permits the fabrication of submicrometer bubble propagation circuits has been described. Straight line patterns and contiguous zigzag patterns are combined with an appropriate registration to form bubble propagation patterns. The straight line pattern width corresponds to the gap width in the Permalloy bubble propagation circuits. By controlling the exposure time in fabricating straight line photoresist patterns, submicrometer pattern gaps are easily obtained using photomasks with 1 μm minimum features. The 4 μm period and 0.5 μm gap width permalloy circuits fabricated using this technique provide promising propagation characteristics for 1 μm bubbles: 60 Oe bias field margin at 60 Oe drive field and 25 Oe minimum propagation drive field.     

Bubble domain propagation dynamics in field-access permalloy driving circuits

To improve the understanding of the bubble domain propagation a stroboscopic investigation of bubble displacement with a permalloy T-bar overlay was undertaken at various values of drive field rotation rate. In addition to stepwise change of bubble wall velocity, variation distortion of the bubble shape in the process of propagation and an asynchronous lag of the bubble relative to the drive field vector were also observed, the lag increasing with the rotation rate. To explain these results a simplified method for determining the bubble position in a magnetostatic well was proposed whereby elements of the permalloy overlay was divided into elementary rectangular components. The computed data are in good agreement with the experimental findings and this permits explanation of a number of phenomena observed in bubble circuits.     

Optimum design considerations for dual conductor bubble devices

A design for dual conductor, current-access bubble devices with 8-μm periods has been optimized with a numerical calculation method for bubble motion in a propagating magnetic field, generated around hole patterns in conductor layers. Magnetic bias field distributions are calculated for an oval hole chain in the conductor layers. Bubble motion equations are obtained with analytical field distribution functions approximating the calculated field distributions. Minimum drive current density Jminfor normal bubble propagation is determined by a solution to the equations. The hole shape has been optimized by the minimization of the drive power Pmin, the product of Jminand conductor resistance, which is calculated from current distributions around the hole pattern. Optimum layer thickness have also been obtained for 8-μm period bubble devices. Both registration tolerance between the two conductor layers and bubble skew effects have been studied semiquantitatively on the basis of the equations of motion. The numerical calculation method developed here is found to be a highly effective means to optimize pattern design for smaller period devices.     

A magnetic bubble passive time-slot interchanging gate

The design and operation of a magnetic bubble logic gate, able to perform the basic retardation operations for a magnetic bubble PCM time-slot interchanger,are reported. With this design no external current pulses are needed to perform those functions. With a 32 µ circuit periodicity and using (SmY)3(GaFe)5O12, bias field marginsashigh as 11.5% for the passive logic function are reported in a 25 Oe rotating field. Because no precautions are taken againsthard bubbles the unsuspicious range of frequency is limited to 20 kHz. Nearly no difference is seen in operating margins between low andhigh frequencies. Design rules are given that canlead to other bubble-to-bubble logic circuits with high operating margins.     

Bubble dynamics in T-bar-type circuits

A theoretical model for bubble propagation under T-bar type overlays due to an in-plane rotating field is developed. By making certain assumptions about the nature of the overlay magnetisation, it is shown that bubble driving forces can be readily calculated for a circular bubble of varying diameter moving freely in 2-dimensions. Expressions for the frictional forces acting on a bubble are derived and the resulting equations are solved to yield the bubble centre trajectory, the radius, velocity and phase variations. This analysis is applied to examine bubble propagation along a straight portion and around a corner of a T-bar circuit. Cases of successful propagation of an isolated bubble, as well as of failure to propagate, are given. Bias margins for a straight T-bar track at 100 kHz are obtained theoretically and compared with experimental results for a chip having the same parameters.     

Control function margin degradation in a bubble memory chip

Chip control function and propagation circuit margin degradation due to long-term memory operation, was observed, using the bias field switching technique. 16 kbit major-minor loop organized bubble memory chips with 28 μm bit period, which had an average access time of 2.7 ms for a 100-kHz rotating field, were used. It was seen that degradations in the lower side of the bias field range were independent of chip functional elements. However, at the upper side of the bias field range, degradations in the performance can be classified by dividing the elements into two categories. These were propagation circuits (Permalloy patterns only) such as H-bars, chevrons, etc., and control functions (Permalloy and conductor patterns), such as generators, replicators, etc. Also, it was found that the degradation in the performance of propagation circuits is small compared with that of the control functions. These differences were considered to be caused by a failure in the Permalloy steps over conductors and/or by the magnetic interaction of the bubble and the conductor current.     

New ion-implantation method for 4-µm period bubble device

It is well-known that bubble propagation margins for ion-implanted bubble devices depend strongly on ion-implantation conditions. A new ion-implantation method is reported that can significantly improve bubble propagation margins for minor loops with 4 × 4 μm bit cell size. The implantation was done through a Mo thin film layer so that the lattice strain and the anisotropy field change would be more uniform through the depth of the implanted layer. With this method, minor loops can be formed by hydrogen single implantation. Consequently simplification of the implantation process is achieved.     

Contiguous-disk bubble domain devices

Contiguous-disk bubble devices are an approach to higher bit density through the use of coarse overlay patterns in manipulating small bubbles to relax device lithography requirements. As a first step towards such an objective, a fully processed chip using ion-implanted devices has been tested, showing the feasibility of all required memory functions with 5-μm bubbles and 25-μm period overlay patterns. A critique of permalloy versus implanted contiguous-disk devices is made, pointing out their basic difference in magnetization reversal processes and explaining the superiority of the latter over the former in achieving a good edge affinity of bubbles. The requirements for a good implanted device are reviewed, including the selection of garnet material parameters (K1, λ111), of implantation parameters (ion energy and dosage) and of device pattern geometry (thickness and shape of implanted layer). An understanding of these requirements has made it possible to demonstrate 1-μm bubble propagation in several contiguous-disk type circuits with 4.5-μm periods, yielding an areal density of over 3 × 10⁷bit/in²made by conventional photolithography.     

High-frequency propagation and failure of asymmetric half-disk field access magnetic bubble device elements

High-frequency propagation characteristics and failure modes in 14-μm period, 1.8-μm gap, asymmetric half-disk field-access device were studied using a high-speed optical sampling technique. Propagation elements as well as normal and hand gun corners and chevron structures were included. The operating bias margin at 1MHz, for a structure that had 1.2 MHz as highest possible frequency, was about half of the margin for frequencies of 200 kHz and below. The phase lag between the bubble leading wall and the instantaneous rotating field direction was nearly 90° as the bubble moved through the center of the element where the lag was the greatest. The peak velocity of the leading wall of 55 m/s and the trailing wall of 46 m/s is attributed to bubble interaction with the Permalloy structure creating a ∼125 Oe in-plane field that greatly increases the free bubble "saturation" velocity.