October 22, 2009
Jian-Gang (Jimmy) Zhu, ABB Professor of ECE and director of the Data Storage Systems Center, is developing a technology seen as a dark horse in the race toward tomorrow's ultra-dense hard disk drives, according to an October 19 article in EE Times. Zhu is developing a prototype of his microwave-assisted magnetic recording (MAMR) technique he believes will pack three terabits of data in a square inch of a spinning disk.
The technique represents a third option in an ongoing debate over the next big shift in hard disk technology expected to emerge in the next year or two. (Read more...,)
Seagate Technology has been championing an approach called heat-assisted magnetic recording (HAMR) that uses a tiny laser light on each drive head to heat a portion of the disk just before a write operation. Rival Hitachi Global Storage Technologies is working on a way to pattern tracks and even bit locations on media.
All sides say they are actively exploring both technologies. Ultimately both technologies will be needed to deliver disks that pack 10 terabits or more per square inch.
Using perpendicular recording, drive makers are now shipping units that pack 530 Gbits per square inch. One analyst said the industry may have to choose within the next year one of the contending approaches to get to 1 Tbit/square inch and beyond.
Either patterned media or heat-assisted recording could deliver densities of 1-10 Tbits per square inch and have similar costs based on today's rough estimates. Zhu of Carnegie Mellon believes his approach could cost less than either approach and provide similar gains.
"What is needed to convince the industry is an experimental demonstration [and my lab] is in the process of fabricating such device and testing its performance," Zhu said in an email exchange.
Using MAMR, a drive head emits a microwave field that excites the electrons in the media building up energy that eases the process of writing data. The process uses a localized high frequency ac magnetic field generated by a magnetic thin film stack integrated with existing recording heads.
The stack consists of "a few more magnetic layers within the present recording heads and the additional process will only increase the complexity of the recording heads by 10 percent or less," said Zhu. Thus MAMR "represents the least disruptive approach with substantial gain of storage capacity," he added.
The patent office recently awarded the university a patent on a spin torque oscillator it uses to implement its film stack. "I am sure that every HDD company is starting some sort of effort to investigate this technique," Zhu said.
Indeed Hitachi GST and others have published papers on versions of MAMR.
Drive makers divided R&D managers at Hitachi and Seagate said they are working both on HAMR and patterned media and see MAMR as a more distant possibility. A spokesman for Seagate said the company has intellectual property relating to MAMR, "but we feel that HAMR or patterned media are more viable approaches."
Currently Seagate feels there is "maybe a little advantage for HAMR," said Mark Re, senior vice president of recording media operations at Seagate. That's because drive makers like Seagate already make their own heads. The HAMR approach only entails the 30 to 40 cent cost of buying a laser diode for each of the one to ten heads in a drive.
However, Seagate has not figured out how to integrate the laser and the head. "That's not a show stopper, because clearly DVD players integrate laser diodes into very cheap players," Re said.
By contrast, patterned media requires several new process steps and new equipment---such as nano-imprinting lithography---that must come from third parties. Talking on those capital costs and estimating the yields for such new equipment and steps is a more daunting challenge, said Re. "That could be one of the things that determine which way we go," he said.
Currie Munce, vice president of research for Hitachi GST, says patterned media has a slight advantage as a next step.
"We are already working with many vendors on their prototyping tools and evaluating their plans for production-level tools for patterned media," said Munce in an email exchange. "We feel there are still a few inventions or breakthroughs required to make [HAMR] viable," he added.
In the end, drive makers are expected to rally behind just one of the two approaches as the next step.
"For either HAMR or patterned media there's an infrastructure that needs to be build up in the supply base," said Re. "I think once the momentum goes in one direction, the industry probably will converge on it," he added.
In the short term, all sides agree drive makers will squeeze advantages from today's perpendicular recoding technology to make it last as long as possible.
"We think we can do [perpendicular recording] for another few product generations," said Re. ""Probably 2012 is when we will need one of these transitions to patterned media or HAMR," he said.
Advances in heads, media, materials and communications channels are giving drive makers confidence they can push perpendicular recording to densities beyond a terabit/square inch. However, no disks have been demonstrated at that level yet.
Source: EE Times
Jmmy Zhu, ABB Professor of ECE; Director, Data Storage Systems Center