No abstract

Discrete tracks as narrow as 0.5 pm have been fortncd by ctching the surface of thin film Co-alloy and y -Fe,O, disks. The recording performance of isolated tracks has been determined using recording heads -10 pm wide, and measurements of rcadback signal and media noise for many discrete track widths have been reported previously.'v2 These results dcmonstrate that magnetic recording is possible on discrete tracks as narrow as 0.5 pm and show that the readback signal and media noise scale with track width as expected by Conventional recording theory. We have extended this work to tracks of width comparable to, and wider than, the head width, permitting a more complete description of thc recording performancc of discrete tracks.One component of media noise results from irregular magnetic domains written at the edges of a track by the fringing field of the recording head," and this contribution to the total noise can be significant when the head moves ~f f t r a c k .~ One potential advantage of discrete tracks is that the media in the region between the tracks can be removed, reducing edge noise and other edge effects. We have demonstrated that edge noise is much lower for a patterned edge than for a written edge: and Fig. 1 shows components of media noise normalized by the isolated pulse amplitude, So, for a disk of Co,,Cr,,. The data for the written edge were acquired by writing a wide track and then probing the center, edge, and off-track regions with a narrower head! The edge noise for a patterned edge is sufficiently small that it cannot be measured reliably by this technique, so the edge noise in this case was determined by linear extrapolation of media noise as a function of (track width)'/, to zero track width.It is important to determine if this significant reduction in edge noise results in improved recording performancc. We have measurcd error rates as a function of distance offtrack for a patterned disk of a Co-based ternary alloy. An inductive head 12 pm wide was used with a peak detection channel a n d (1,7) code with a minimum transition spacing of I pm. The error rates for a discrcte track 11.1 pm wide and conventional media on the same disk are shown in Fig. 2. The media on either side of the conventional track were dc-erased before the track was written, and the discrete track is separated by 4.5 pm from other tracks. The data for these two cases are quite similar, although there is a small improvement in offtrack capability by N 0.5 pm for offtrack distances of about 5 pm which may reflect the fact that the discrete track is narrower than the read head. These error rate data show that edge noise in ternary Co-alloys is not necessarily a crucial factor in recording performance for heads 12 pm wide, although the effects of adjacent track interfcrcncc have not yct been determined. We intend to extend these measurements to narrower heads where edge effects may be more significant. These results also demonstrate that edge roughness or other defects which may be induced by lithographic processing do no...