Micro-drives: Hard-disk revival driven by thirst for mobile storage
09/01/2005
Rapid growth in consumer electronics, particularly for mobile applications, is driving a surge in demand for hard-disk storage, particularly micro-drives. Meeting areal density challenges will require a shift from longitudinal to perpendicular recording. Also, new process tools capable of handling multiple small form-factor disks will be needed to increase throughput.
The recent explosion of consumer electronics devices has generated new demand for a large amount of storage for media files. These new applications are fueling a revival in the hard-disk drive (HDD) business. “Hard disk drives are everywhere now, and they are enabling a digital lifestyle in ways the world has never before seen,” states Bill Watkins, Seagate CEO. “Our industry is at the beginning of a growth cycle that could be as big as the PC revolution of the ’80s. Consumer electronics is the primary growth driver because people are voracious when it comes to getting more gigabytes for their MP3 players, their TVs, gaming devices, and PCs. And cell phones, autos, digital cameras, and storage in your home appliances will soon be eating up even more gigabytes and terabytes of hard disk drives.” Figure 1 shows the projected growth of disk drive volumes, primarily in non-PC and laptop applications.
 Figure 1. Hard-disk drive shipments by application. |
Furthermore, the disk drive industry is projected to exceed $30 billion and 550 million units annually by 2008, according to Gartner Research, up from $21.4 billion in 2004. Consumer electronics accounted for 9% of all hard drive shipments in 2003, but this is predicted to grow to 40% in 2008. Figure 2 shows the projected growth of disk drive volumes by form factor. Consumers have demonstrated an insatiable appetite for digitally storing music, photos, video and other personal documents. These applications demand ultrahigh storage capacity, small size, and low cost. The ability to produce even higher-capacity disk drives without raising cost will continue to enable new applications.
 Figure 2. Hard-disk drive shipments by form factor. (Permission to reprint from Coughlin Associates, Copyright 2005) |
null
Going mobile
One of the fastest-growing areas for disk drives is in mobile consumer electronics. Current applications include MP3 players, digital cameras, PDAs, handheld digital video players, GPS units, and multimedia cell phones. The increasing storage requirements for these devices are driving the demand for micro (1 in. and smaller) disk drives. These drives are to be distinguished from the somewhat larger 1.8-in. types used in iPod and similar portable audio systems, which can provide as many as 60GByte of storage capacity. The 1.8-in. units are too large for mobile telephones and similar applications.
Seagate recently announced its 8Gbit 1-in. hard drive that can hold up to 4000 high-quality music files. The 8Gbit drive also comes in a standard CompactFlash type II card with capacity for more than 2400 high-resolution (6 megapixel) pictures in a single card. PalmOne’s LifeDrive is the first handheld to have a 4Gbit integrated hard drive. It lets users work on office files, check e-mail, surf the Web, listen to music, organize photos, and view videos.
Samsung unveiled the first cell phone containing a hard drive last year with a 1.5Gbit 1-in. drive. This provides enough capacity for up to 350 MP3 songs. Toshiba demonstrated a working prototype of its 4Gbit 0.85-in. drive at the last Computex show. The product has the distinction of being the smallest commercially available HDD and is currently available with 2Gbit capacity. The 4Gbit version will be available later this year and is slated to be used in Nokia Corp.’s upcoming N91 cell phone.
These miniature drives successfully compete against solid-state memories by providing higher capacities at lower cost. This creates an urgent need to continually increase storage capacity while maintaining low cost. The small recording surfaces of these disks require rapid advances in areal densities to maintain this advantage.
Micro disk-drive production requires new design and manufacturing processes for the miniature drive components and assembly. This article will discuss two key challenges on the magnetic recording medium (the rotating hard disk that records the data): the changes in the sputter films required to continue increasing storage capacities and the need to create cost-effective methods to handle the tiny disks.
Capacity increases
The small recording surfaces of the micro disk-drive media require rapid advances in areal densities to maintain an advantage over flash memory.
For almost five decades, the hard drive industry has continuously grown storage capacity using longitudinal recording technology. The average areal density of a hard-disk platter increased at rates over 100% annually during the late 1990s to early 2000s. This rapid increase in data density was accomplished during a steady decrease in the price of the disk drive. This combination of low prices and the improvements in data density was critical in enabling new consumer applications.
For longitudinal recording, increases in areal density have been primarily achieved by reducing the size of the data bits. This is limited by the magnetic grain size and the anisotropy energy along the recording direction.
Longitudinal recording is now near the point where further reducing the grain size will affect the data integrity due to the super-paramagnetic effect. This fundamental technology roadblock has contributed to slowing hard-disk storage capacity growth. Perpendicular recording offers an immediately available means of extending areal densities.
Transitioning to perpendicular recording
Longitudinal magnetic recording aligns the magnetization of the data bits horizontally, parallel to the disk surface. For the highest-density bit pattern of alternating ones and zeros, the adjacent bits end up with the like poles next to each other to create a high demagnetization field. This creates an unstable configuration where thermal fluctuations can cause the bits to “flip” and corrupt the data. Figure 3 shows the difference between longitudinal magnetic recording and perpendicular recording.
 Figure 3. Longitudinal magnetic recording vs. perpendicular recording. |
In contrast, perpendicular recording aligns the magnetization vertically - or perpendicular to the disk surface. The adjacent bits end up with the opposite poles next to each other (attraction), resulting in a lower demagnetization field to allow higher packing densities. The magnetic easy axis is highly oriented as defined by the sputter process. The addition of a magnetically soft underlayer on the disk effectively functions as part of the write head to enhance the write field. This allows the use of films with higher anisotropies to further extend the super-paramagnetic limit.
Disk-drive manufacturers have been battling each other to demonstrate the highest densities using perpendicular recording over the past couple of years. Several major manufacturers have already announced disk-drive products based on perpendicular recording this year.
The transition to perpendicular recording will drive significant changes in the magnetic film deposition process and equipment. The sputtered films can be more than 10× thicker than the films used in longitudinal recording. The soft magnetic underlayers can also be complex structures with 3-8 layers by themselves. A seed layer is added on top of the underlayer; the magnetic layers are on top of this; and finally an overcoat is added. Most of the existing hard-disk sputter machines were developed for longitudinal recording and do not have enough chambers to create these multilayer films.
The early hard-disk media for longitudinal recording consisted of only three layers (chrome underlayer, magnetic layer, and carbon overcoat). This has evolved to 10- and 11-layer structures on current longitudinal production. The current perpendicular hard-disk media is starting with 14-15 layers. Figure 4 illustrates the differences. The sputtering process for perpendicular recording is likely to get even more complex as engineers strive to further increase areal densities.
 Figure 4. Longitudinal and perpendicular thin-film structures. |
The disk media costs will be higher for perpendicular recording due to the increased complexity, thicker films, and investment in process equipment. This has less impact on the micro-drive media since many more of the smaller disks can fit into a sputter panel for batch processing.
Disk-handling challenges
Most of the current hard-disk production equipment was originally designed for disks of 95mm (3.74 in.) dia., which continues to be the dominant form factor. Much of this equipment has been adapted to work with smaller disks (down to 2.5 in.) by simply scaling the disk handling devices and process heads. By running the smaller form factors on the same lines, the disk manufacturers avoided the major expense of building new production lines and maintained the flexibility to switch between disk sizes as needed. It also allowed them to leverage existing deposition processes by using the same process chambers. Even though this solution resulted in higher material usage and larger equipment footprints, it was still very cost-effective when compared to the large investment for new equipment.
With the introduction of sub-1 in. disks, we have reached the point where converting the existing equipment may no longer be the lowest-cost solution. In many cases, scaling the mechanisms is no longer feasible or reliable. There are several common disk sizes currently in production, ranging in size from 0.85 in. to 95mm. The 0.85-in. disk actually fits in the center hole of the 95mm disk. In some cases, the existing mechanism may not be precise or gentle enough to handle the miniature disks. The disks are light enough to float away when immersed in a cleaning tank. Even sensing the disks can be a challenge.
The sub-1 in. disks are primarily used for consumer applications where cost is critical. This plays well into batch processes; material cost and throughput are direct functions of the batch size. For instance, the cost of sputtering the expensive magnetic layers on the disk is fixed for a particular chamber size regardless of the quantity of disks being sputtered. Six times as many 1-in. disks can fit inside the footprint of a 95mm disk, theoretically increasing throughput (or film thickness) and reducing material cost by a factor of six. The trick is to design a carrier for the small disks and either load 6× as many disks in the same interval or develop new equipment to load these disks offline.
The same economies of scale do not apply for single disk processes. Here, the process time decreases due to the small surface area of the disk, and the cycle time becomes dominated by disk handling. These systems may have to be completely redesigned to balance throughput with the batch processes. In some cases, this may involve providing multiple process stations, which can be serviced by the same disk and cassette handler. For other situations, there may be an opportunity to convert a single disk process to a batch process.
The handling of sub-1 in. media requires new designs for ID grippers, mandrel, spindles, and vacuum cups as well as OD grippers and saddles. The greatly reduced areas available for holding the disk require precise alignment for hand-offs. In some cases, retrofitting existing equipment with alternative mechanisms (e.g., changing from ID to OD gripping) has been possible with minimum impact to the controls and software. In other situations, clever design of automation may allow the simultaneous loading of multiple small disks in place of a single large disk.
What’s next?
Sub-1 in. disk drives have become one of the fastest-growing segments of the data storage industry. Capacity increases and cost reductions are required to compete with solid-state memories and open up even more market opportunities. The transition to perpendicular recording will allow areal density to continue increasing, but will also require investment in new sputtering equipment.
Achieving the necessary cost reductions for manufacturing sub-1 in. hard disks will require innovative approaches to disk handling, not simply the scaling of existing methods. Those manufacturers that make the first move to invest in dedicated small form-factor equipment will be best positioned to capitalize on this segment’s tremendous potential.
Bob Fung is director of engineering at Owens Design Inc., 47427 Fremont Blvd., Fremont, CA 94538; ph 510/576-1800, e-mail [email protected].