By Rick Merritt EBN (09/12/2003 1:00 PM EST)

After 20 years of research and debate, the disk drive industry agrees that it must make a historic shift from longitudinal to perpendicular recording. However, fortunes could be won or lost in this massive but increasingly consolidated industry depending on exactly when each drive company makes the shift and at what cost.

Komag Inc. detailed progress in media and Seagate Technology LLC and Western Digital Corp. discussed their separate work on new heads for perpendicular recording at the annual Diskcon conference in San Jose this week. A consensus appears to be emerging that the shift will come for server and notebook drives in the 160Gbyte/platter generation in the second half of 2005, followed by a move for desktop drives in the 200Gbyte/platter generation.

The plans are unfolding at a time when growth in areal density -- the amount of data that can be packed on a square inch of media -- is throttling back from a torrid growth rate of 100% a year to less than 50%. Business managers generally welcome the slowdown, which they see helping keep supplies of drive Gbytes more in line with demand and diversifying what was once a singular and intensely expensive focus on capacity.

Drive leader Seagate Technology, Scotts Valley, Calif., once spent up to 96% of its R&D budget pursuing areal density, according to one executive. Now the company wants to spend more money tailoring drives to the unique needs of the growing notebook, server, and consumer segments and developing more small-form-factor drives.

Nevertheless, areal density remains a top concern for drive makers because its capacity gains can be used to cut costs by reducing the number of platters and heads per drive. Areal density improvements can also fuel an expected shift from 3.5in. drives that dominate the market to smaller form factors demanded by mobile, server, and consumer gear.

To keep areal density growing, drive manufacturers believe they must shift from storing magnetic charges longitudinally on a disk's surface to vertical storage. Benefits stem from the fact that it's difficult to push horizontal charges close together while vertical charges will naturally couple, opening the door to smaller recording tracks or more bits stored per track area.

"Areal density growth has been slowing because longitudinal recording is running out of gas," said Francis Lui, director of advanced recording at Read-Rite Corp., Fremont, Calif., which recently was acquired by Western Digital, Irvine, Calif.

Using a new shielded-pole magnetic head, a team consisting of Read-Rite, Fuji Electric, Komag, and Samsung developed an 80Gbyte/platter drive prototype that stored 63.8Gbits/sq. in., an 18% boost in areal density over a longitudinal drive. The drive had a linear density of 725Kbits/sq. in. and a bit error rate of 10-8.

"Today, perpendicular recording is 20% better than longitudinal, but it needs to be twice as good," Lui said.

In part that's because perpendicular drives could be an estimated 20% more expensive at introduction based on extra costs for the heads and media. Media alone will account for up to a dollar a platter in additional costs just for new materials, said David Wachenschwanz, senior technical staff member for Komag, San Jose.

Cost premiums for perpendicular heads -- the most expensive parts in a hard drive -- could be even greater.

"The very important next step" for drive companies is to build prototypes in volume to identify manufacturing issues, yields, and failures per million for the perpendicular drives, Lui said.

However, changing manufacturing lines is a risky issue for a 200-million-unit industry focused on low-cost production. So vendors are looking for ways to leverage existing technology or stave off the shift to perpendicular for a generation, potentially reaping a competitive advantage.

Western Digital's approach appears to entail an early shift to perpendicular recording but the continued use of existing current-in-plane giant magneto-resistive (CIP GMR) head sensors to control costs and hedge risks. The company would later move to new current-perpendicular-to-plane (CPP) sensors that interact more efficiently with the vertically charged media but are still based on thin-film metals.

Seagate is taking a different tack. The company detailed at Diskcon its tunneling magneto-resistive (TMR) heads, an aluminum sensor that responds to a tunneling effect rather than the direct magnetic charge of the media. TMR heads, like CIP GMR heads, orient current in the same direction as the vertically charged media, but could also be used to reap benefits from today's longitudinal media, said Sining Mao, director of advanced devices for Seagate's recording head division.

Mao said the TMR heads require new structures, materials, and physics, and have workable but still unsolved reliability and manufacturing issues. For example, the heads require a new deposition system costing about $6 million.

Nevertheless, TMR heads could be big enablers for 2.5- and 1in. drives, Mao said, noting that as many as seven independent or captive head makers have published or announced work on TMR heads. One wild card is the use of patterned media.

Today, drive heads determine the placement of data tracks, but researchers say that, ultimately, media will be pre-written with track and even bit placement information, opening the door to simpler heads.

"We believe that pattern media will be needed for tracking media to 1Tbit/sq.-in. densities," said Robert Evans, a director of technical support for drive component maker Hutchinson Technology Inc., Hutchinson, Minn.

While Lui and others said patterned media probably won't come into play until after the switch to perpendicular recording, others disagree.

"There are several companies, including Komag, developing versions of patterned media that could extend longitudinal recording by two generations," said Chris Bajorek, executive vice president of advanced technology at Komag.