Naoki Koyama

Ion-implanted and permalloy hybrid magnetic bubble memory devices

Hybrid bubble memory devices have been proposed and operated with the memory density of 4 Mbit/cm2. In the hybrid bubble memory devices, minor loops are composed of ion-implanted tracks with 4-µm period, and major lines and functional parts including block-replicate and swap gates are composed of Permalloy tracks with a longer period of 12 µm. Passive junctions between ion-implanted and Permalloy tracks have been developed, introducing the tapered ion-implantation technique. Improving the characteristics of the functional parts composed of Permalloy tracks, the hybrid bubble memory devices with block-replicate and swap gates have been operated, and the feasibility of the devices has been confirmed. In addition, the possibility of higher memory density has been shown.

Characteristics of junctions between ion-implanted and permalloy tracks in hybrid bubble devices

The characteristics of junctions between ion-implanted and Permalloy tracks are described for hybrid bubble memory devices. It has been found that the junctions which have a sharply formed edge of the ion-implanted layer have a problem in reproducibility of bias field margins. It is caused by the potential barrier due to the stress change and free poles which appear at the edge as well as the attractive poles in the ion-implanted tracks which tend to keep the bubbles in the ion-implanted tracks. To reduce their influence, we have introduced the tapered edge. A good bias field margin has been obtained for the tapered junctions with the modified pickax Permalloy pattern from the ion-implanted to Permalloy tracks. In the junctions from the Permalloy to ion-implanted tracks, the bias field margin strongly depends on the orientation of the edge of the ion-implanted layer with respect to the easy magnetization directions of the garnet. The edge should not be parallel to the easy magnetization direction, and the Permalloy pattern should be made larger near the edge. The improved junctions have a good bias field margin with reproducibility.

New single-mask approach to bubble device fabrication

Permalloy devices with two separate levels of fine permalloy and conductor patterns have been fabricated by the use of a single-mask process. This process makes possible the production of devices having an essentially conventional design in addition to a completely planar structure. The features of this new process include 1) preparation of a reticle on which patterns for both permalloy and conductor layers are superimposed, 2) use of SiO 2 , Mo/Au/Mo, SiO 2 , permalloy, and TiO 2 thin films deposited sequentially on the bubble materials, and 3) simultaneous delineation of both photoresist patterns with two different thicknesses and of the desired patterns by CF 4 plasma etching and ion-milling.

Study on high field transfer curves of GMR heads with damaged pinned layer by ESD

In order to study the effect of ESD on antiferromagnetic layer of GMR heads, transfer curves of GMR heads using high-field QST (quasi static tester) were investigated. Transfer curves of GMR heads vary in waveform by applying of ESD damage, but there are many kinds of waveforms of transfer curves of after ESD damaged heads. Analysis of these damaged heads in magnetic state using high-field QST suggested that ESD damage would degrade the interlayer exchange coupling between pinned layer and anti-ferro magnetic layer.

Simulation of magnetization distribution in magnetoresistive film under a longitudinal bias field

The effects of longitudinal bias field, used for domain control on the magnetization distribution in a magnetoresistive (MR) film, have been investigated by computer simulation. The longitudinal bias field was generated by an exchange‐coupled antiferromagnetic or permanent magnetic film formed on the MR film outside the sensing region. It was assumed that the magnetization in the part of the MR film on which the bias‐generating films were formed was fixed along the easy axis. The spatial sensitivity of the MR film along the track width was evaluated by calculating the dependence of the resistance change on the position of a narrow track recording medium. It was found that the resistance change in the MR film with the anti‐ferromagnetic film was roughly twice as large as the change in the film with the permanent magnetic film. The asymmetric sensitivity profile with respect to reflection about the track width mid‐plane was also obtained. The asymmetry in the track sensitivity profile was found to be caused...

Fabrication process for ion-implanted and Permalloy hybrid magnetic bubble memory devices

A fabrication process for ion-implanted and Permalloy hybrid bubble memory devices has been developed. The mask patterns of polyimide (PIQ) for ion-implanted tracks are fabricated using a tri-level resist process where the top imaging resists are delineated by deep UV contact printing. The minimum feature of the PIQ pattern is 0.75 mu m and can be controlled precisely. The sidewalls of PIQ patterns are smooth and vertical. At the junctions between ion-implanted and Permalloy tracks, tapered resist patterns are used as masks for implantation. Novolak-based resist is used and is reflowed thermally to obtain a tapered profile. Fabrication conditions, such as type of resist and baking temperature, were investigated to achieve good reproducibility. Tapered resist patterns are also used as masks for iron milling which reduces the thickness of garnet film in the area of the Permalloy tracks. This is necessary to adjust the operating bias field of Permalloy tracks to that of the ion-implanted tracks. A prototype of a 4-Mb memory device was fabricated, confirming the feasibility of the present process. >

Sensitivity distribution asymmetries in magnetoresistive heads with domain control films (abstract)

Suppression of the sidetrack reading behavior is important in achieving⋅high track‐density recording with a magnetoresistive head. In the present work, we investigate the sensitivity distribution profiles of MR heads through experimentation using a microtrack technique1 and computer simulation. Shielded, soft, adjacent‐layer‐biased MR heads with 3∼6−μm‐wide tracks are used. The ends of the MR elements are coupled to antiferromagnetic or hard ferromagnetic films for domain control. In the simulation based on the Landau–Lifshitz–Gilbert equation,2 the magnetization configurations in the MR elements of the domain stabilizing films are calculated to analyze the sensitivity distribution. The measured sensitivity profiles in all cases have asymmetries that are reversed when the sensing current direction or the magnetization direction of the MR element is reversed. These results agree with the simulation. The asymmetries of sensitivity profiles are explained by the magnetic poles at the domain control film edges...

Quasistatic Bubble Propagation in 1.5 μm Period Ion-Implanted Tracks for 64 Mbit Magnetic Bubble Memory Devices

Ion-implanted tracks with a 1.5 μm period for use in 64 Mbit magnetic bubble memory devices have been fabricated and operated. The bubble material used was (BiSmLu)3 (FeAl)5O12 garnet film, supporting 0.45 μm bubbles. The resist mask patterns for implantation were formed using a 1/10 reduction projection aligner, using rectangular patterns on a reticle. The bubble propagation tracks were fabricated by double deuterium implantation and annealed at 400°C for 30 minutes, with an SiO2 layer deposited on the garnet surface to stabilize the implanted layer. An operating bias field margin in excess of 7% was stably obtained in quasistatic bubble propagation. This shows that the ion-implanted devices have great potential for use in 64 Mbit magnetic bubble memory devices.

Study on Optimum Condition for Fabrication of Tapered Ion-Implanted Layer in Hybrid Type Bubble Memory Devices

Methods for precise formation of tapered ion implanted layers for use as propagation tracks in 4 Mbit magnetic bubble memories were discussed. After forming a mask pattern using multi-layer resist methods, a resist pattern was formed on the propagation line connector, and heat treated. This heat treatment changed the resist cross-section angle ¿ and the pattern lengthening l; their dependences on the heat treatment temperature were determined. Optimum precisions of ¿ = 48 ± 2.3° and l = 0.43 ± 0.02 ¿m were obtained at 170°C without ultraviolet exposure.