Site: National Institute for Advanced Interdisciplinary Research
Tsukuba, Ibaraki 305, Japan
Date Visited 10 March 1998
NAIR was founded in 1993 as part of MITI's Agency of Industrial Science and Technology (AIST) to pursue interdisciplinary research in fundamental and frontier areas of industrial science. The government-funded research is collaborative, and includes industrial, academic, and government participants. Management is based on four principles: extensive openness, flexibility and mobility of staffing, international collaboration, and objective evaluation of research progress. Though NAIR does retain permanent research staff, term employees drawn from government, industrial, and academic organizations, as well as foreign guest researchers, make up the majority of research staff.
The four major research areas are divided into the following groups: Atom Technology, Cluster Science, Bionic Design, and Optical Memory. The Optical Memory Group had an FY 98 budget of ¥160 million and 13 researchers (4 NAIR staff, 2 assigned from other AIST institutes, 3 assigned from universities, 3 industrial researchers, and one foreign post-doc). The group began its research program in April 1997 after completion of a one-year feasibility study.
The Optical Memory Group's objective is to develop a technological basis for ultra-high density data storage using near-field optics. The research program is five-fold: (1) explore near-field recording principles; (2) explore suitable recording materials; (3) develop fabrication techniques for various probe types; (4) reduce feature size while improving dimensional accuracy and enhancing power throughput; and (5) evaluate dynamic signal detection. To address these issues, the group is divided into two projects, with one focusing on optical system design and another working on recording media design.
A prepared presentation by Dr. Atoda described several issues facing media selection. First, media must record and erase without significant deformation to reduce the possibility of head crashes. Second, the low optical power provided by near-field probes suggests that thermally assisted recording is desirable. Finally, clear recording thresholds are needed to obtain good edge definition. These issues make current organic write-once media less attractive than phase change or magneto-optic materials.
Early work has looked at the behavior of phase change (PC) materials. A test bed using a classical microscope head has been built for making 200 nm diameter marks on PC materials to test near-field readout. One demonstration showed that near-field readout modulation doesn't significantly degrade for dielectric cover layers as thick as 20 to 30 nm.
The researchers have observed a 5% reduction in volume due to writing (amorphous to crystalline) in SiN/GST/SiN sandwich structures. Study of the sandwich indicates that GST crystallization begins at the dielectric interface and is initially confined to within 15 nm of the interface. This suggests the possibility of two layers of recording if the GST film is thicker than 30 nm.
Another area of research involves array write/read. Researchers have produced and characterized 200 nm diameter holes in Si wafers; these appear useful as masks that could be placed over a laser beam to produce a two-dimensional array of probes. Ball lenses inserted over the holes have improved the optical throughput by up to two orders of magnitude. One result of this work might be a flying head array with integrated laser and detector. NAIR researchers estimate that near-field probe data throughput is limited (by optical power considerations) to 10 to 100 kb/s; a read/write array could increase this to 10 Mb/s, so array technology is important for near-field systems.
When WTEC's hosts were asked to name the single most difficult aspect of this project, tracking was the quick reply. Reduction of track pitch raises trade-offs between tracking accuracy and the dynamic range needed to accommodate track noise, wobble, etc. The registration of read/write arrays to tracks becomes more difficult with decreasing track width, and may require challenging array fabrication tolerances or array agility to accommodate compatibility among systems. For optical storage in general, the development of short wavelength lasers was considered a key area.
Though this is a long-term research project, there was discussion about the market forces that affect commercialization, and concern was expressed about the profitability of the storage industry. Because prices fall as participants introduce competing products, it's important to have a product that is resistant to these pressures. One approach is to introduce products early in the market cycle and benefit from this higher profit period. Near-field recording is very challenging, however, and the hosts estimated that five years would be needed for prototype development, and as much as 20 years for a commercial product.
Developing a near-field optical storage system is an ambitious undertaking that presents many challenges, and the NAIR team, though in early stages of research, clearly recognizes this and harbors no illusions. It is an interdisciplinary problem, however, and may benefit from the flexible staffing approach used at NAIR; as problems are solved and new issues become important, the group can draw from many sources to enlist researchers with expertise in the current problems.