Data Storage Systems Center

Discrete track and bit patterned recording media have been suggested as a means of improving areal density and recording performance while maintaining signal-to-noise ratio. In this study, we present the fabrication and spin stand analysis of discrete tracks and patterned bits fabricated in commercially available perpendicular recording media. Spin stand measurements of the discrete tracks show improved performance in adjacent track erasure (ATE) and increases in the amount of current needed to write the media, where both changes in recording performance are most likely caused by the same fundamental mechanisms relating to media exchange coupling. The wall angle of the patterned tracks is suggested as one possible factor influencing the writability of the tracks. Therefore, metal based reactive ion etching is currently being explored to improve wall angle. For bit patterned media, spin stand analysis of low and high frequency writing on patterned bits emphasizes the need for write head synchronization with the isolated islands. Magnetization reversal of patterned bits from a DC saturated state is examined as well.

Magnetic tunnel junctions have been studied extensively for their magnetoresistance and potential uses in magnetic logic and data storage devices, but little is known about how their performance will scale with size. Here we examined the electronic behavior of 12 nm diameter magnetic tunnel junctions fabricated by a novel nanomasking process [1]. The pillars were fabricated starting from a thin film consisting of 30 nm Ta/ 5[0.7 nm Co/1.8 nm Pt]/ 1.0 nm Al2O3 / 2[0.7 nm Co/1.8 nm Pt] / 1.8 nm Pt. Argon ion milling with nanoparticles as a shadow mask was used to pattern well-separated magnetic tunnel junctions. Scanning electron microscopy images indicated feature diameter of 12 nm, and atomic force microscopy showed a height of 10 nm suggesting that unmasked regions have been milled on average to the oxide barrier layer, and areas should have the remnants of the free layer exposed. Electrical contact was made to individual nanopillars using a doped diamond atomic force microscopy probe with a 100 nm radius of curvature at the tip. An algorithm was developed to account for sample drift to maintain electrical contact between the probe and nanopillar. Off pillar we observed a resistance of 6.4 ∙ 104 Ω, while on pillar we found a resistance of 6.1∙106 Ω. Based on the RA product for this film, 100 Ω∙μm2, a 12 nm diameter cylinder with perfect contact would have a resistance of 8.85 ∙105 Ω. The larger experimental value is consistent with a smaller contact area due to the curvature of the tip and pillar surfaces. The magnetoresistance characteristics of these magnetic tunnel junctions will be discussed.

Self-assembled nanoparticle arrays show promise as etch masks for bit-patterned media, since they exhibit long-range order with feature densities well beyond 1 Tb/in2. Previous work has demonstrated the fabrication of individual nanopillars by ion milling with nanoparticle etch masks[1].

Reactive Ion Etching (RIE) was used for pattern transfer of ordered nanoparticle arrays. The high chemical selectivity, along with the gas-phase nature of the reaction products, allows for deep etching. To etch between the particles, the interparticle surfactant must be removed. This tends to destabilize the array, leading to crack formation which gets worse as the etch progresses. This problem is solved by sputtering a conformal Si film on top of the array. The wafer was then affixed to a new, carrier wafer, flipped, and the original wafer was etched completely through using SF6, leaving the particles exposed from the other side. The surfactant could now be removed by H3PO4 without damaging the pattern.

There are two routes to using the pattern to make media. One is to sputter the entire storage stack on top of the nanoparticle array, and the other is simply to use the etched pattern as a master for nanoimprint lithography.

In order for discrete track recording (DTR) media to become a viable alternative, it is essential to produce accurate servo patterns in a cost effective manner. In this study, we present various ABCD burst amplitude and AB burst phase servo patterns, as well as address patterns fabricated on commercially available perpendicular magnetic recording media. Spin stand analysis of the DC erased patterned disks confirms the patterned servo produces clean position error signals with reasonable amplitude and good linearity. Various bit patterns and small scale servo patterns were also fabricated on the same perpendicular media. Recording analysis of the additional patterns remains to be performed, but the initial results on the patterned servo presented here look promising.

[1] M. T. Moneck, J.-G. Zhu, Y. Tang, K.-S. Moon, H. J. Lee, S. Zhang, X. Che, and N. Takahashi, J. Appl. Phys., 103, 07C511 (2008).

Nanoparticles create a mask for a variety of substrates at the nano scale. For this work, isolated nanoparticles are used to mask a magnetic tunnel junction multilayer. After Argon ion-milling, isolated magnetic pillars are exposed above the continuous oxide barrier. These pillars are studied with Magnetic Force Microscopy. The Magnetic Force Microscopy is also simulated with a simple model to better understand the experimental results.

This talk examines the use of scanned probe lithography as a method for fabricating patterned media. The challenges for mastering nanoimprint masks are examined, particularly with regard to the ultimate number of copies a master must produce. If generated by e-beam lithography, each master would take weeks to fabricate and would ultimately need to produce 1 million copies. This would probably be in a 1 or 2 stage replication process. Consequently, any damage to the master before completion of its use, will result in costly losses. Conversely, if scanned probe arrays are used for master pattern generation, reasonable assumptions suggest mastering could be done in minutes, rather than weeks. This greatly reduces the number of copies that a master must generate and the cost of damage to or defects in one master. Wilder, et al [1] have demonstrated scanned probe lithography at 50 nm line width. Examination of the physics of scanned probe field emission suggest that 10 nm lines are feasible.

[1] K. Wilder, C.F. Quate, B. Singh and D.F. Kyser, J. Vac. Sci. Tech. B vol. 16 (1998) p. 3864