Site: Fuji Electric Co., Ltd.
4-18-1, Tsukana Matsumoto
Nagano 390, Japan
Date Visited: 13 March 1998
WTEC: W. Doyle (report author)
Fuji Electric, a relatively unknown company in the United States, is a worldwide supplier of equipment and components for electronic and power applications. It was the parent of Fujitsu, which was spun-off in the late 1930s. The net revenue in 1997 was $5.7 billion, generated by 12,900 employees. About 9% is accounted for by the Disk Division located in Matsumoto near Nagano. Disks are produced at Matsumoto, at two factories in Yamanashi near Shirane and at a new site in Malaysia with a total of 19 production lines. The R&D laboratory for storage, located in Matsumoto, has about 100 researchers and participates in the SRC program. A broad product line supplies disks to all major markets.
Fuji Electric is a major supplier of hard disks, with a strong research effort focused on media for 20 Gb/in2 (SRC) and 40 Gb/in2 (ASET) demonstrations by 2001. Strong interactive discussions showed that Fuji management clearly recognizes the importance of superparamagnetic effects and is directing a major effort to optimize performance and thermal effects at high density. Fuji managers are also considering potential alternatives such as perpendicular recording, SIL, near-field and MO which they feel will be required beyond 40 Gb/in2. No discussion of probe based storage occurred. Fuji Electric is not a producer of optical media.
W. Doyle gave a 30 minute presentation on high speed switching in thin film media based on his invited paper at the 1998 Joint Conference. Dr. Fumoto gave an extensive review of future storage technologies based on projects in SRC and ASET. Both presentations were highly interactive.
The widely heralded technology demonstrations by IBM and Fujitsu have been followed in the last few years by products in less than 3 years. Demonstrations of 20-40 Gb/in2 in 2000 will require new media with higher coercivity, lower noise limit, and no greater sensitivity to thermal decay than today's low Mrt media, which already show severe problems at both short and long times.
In work done in collaboration with J. Judy, University of Minnesota, the recorded signal decay decreased rapidly below a grain size of 10 nm but appeared independent of recording density up to 150 KFCI. CoCrPtTa was found to be superior to CoCrTa, consistent with the higher anisotropy K in the Pt films. Based on simulation results, a target specification is K 2 x 106 ergs/cc, Ms = 460 emu/cc, MR = 320 emu/cc, with a grain diameter of 9 nm and a magnetic spacing <15 nm, which will give a predicted relaxation time of 108 secs (~3 years). To obtain sufficient SNR, the grain interaction must be reduced which can be correlated to higher HC/Hk ratios. The grain size must be reduced and a narrower distribution of grain sizes achieved. Fuji managers believe that an important advance was the reduction of gaseous impurities by two orders of magnitude, which resulted in an increase in HC from 2,200 Oe to 3,000 Oe. Impressive lattice images in (110) CoCrTa on (200) Cr showed tri-crystal symmetry. No specific explanation for the increased coercivity was offered. Results on these materials were published in 1995 and 1996. Some work on granular media showed very high viscosity, which has discouraged further work in this area.
It was suggested that a 20 Gb/in2 demonstration at slow disk rotation speeds would be announced by a Japanese company in the near future.
Present disks with mechanical/chemical texture and an Ra = 1 nm with a flying height of 16 nm have survived 400 K passes. Future disks will have an Ra - 0.3 nm with a flying height of 10 nm.
Simulations predict that the limit set by thermal effects for longitudinal recording will be ~20 Gb/in2 but that the limit for perpendicular may be >40 Gb/in2. However, no indication of a significant perpendicular media program was given and the hosts expressed interest in a collaboration through NSIC. The main problems associated with perpendicular recording, including the absence of an erase band, increased spacing loss, domain noise in a soft underlayer, and head-induced data erase, were clearly recognized and may explain Fuji's present research posture.
An evaluation of possible future technologies including SIL technology (Terastor), optically assisted Winchester (Quinta), near-field MO and MAMMOS (magnetic amplifying magneto-optical system) being considered as future alternatives was presented. The most likely candidate was judged to be MAMMOS, allowing 60 Gb/in2 on removable media, although the access time would be a problem. No discussion of probe storage occurred.
Fuji Electric is a highly competitive supplier of rigid disks with a clear awareness of the challenges faced by the media at densities >10 Gb/in2. As alternate technologies evolve, it should be expected that Fuji will continue to play a significant role.