Sunday, March 11, 2012

Hard Disk Drives: Magnetic head processing technology for small form-factor hard drives


Platform-based process equipment and integrated process and metrology tools will be pivotal in advancing reliable, economical manufacturing of small form-factor hard disk drives (HDD). This article looks at key fabrication challenges and current equipment directions for next-generation magnetic heads.
By 2010, the primary end market for HDDs will shift from computers to a diverse family of consumer electronic (CE) devices that require large amounts of low-cost memory. Not only are HDDs found in MP3 players, they are beginning to proliferate in such products as digital cameras, portable gaming devices, personal digital assistants, digital video recorders and players, and especially in full-featured mobile phones. Gartner Research has raised its forecast for total mobile phones shipped in 2005 to 750 million. With the oncoming arrival of 3G mobile technology and a forecast for per capita penetration in dense regions like Europe that exceeds 100% in 2007, the potential market for full-featured phones utilizing HDD storage is enormous. In other sectors, market-research firm In-Stat/MDR anticipates that CE devices will absorb one-third of HDD production by 2008, while Hitachi Global Storage has projected that by 2010, households will commonly utilize 10-20 HDDs in a variety of CE devices. On average, the studies seem to point to a CE HDD utilization 5× that for computers by the end of the decade.
Keenly aware of this groundswell, HDD manufacturers face formidable technological and manufacturing challenges as they transition from the storage density plateau of 100Gb/in.2 to a target level of 250Gb/in.2 and beyond. Accompanying the gain in areal density will be a shift toward low-cost, high-quality, small form-factor (1-in. and smaller) drives suitable for handheld devices such as mobile phones.
From a market perspective, HDD manufacturers are continuously challenged by flash memory to provide a more cost-effective storage solution. Storage capacity and other performance characteristics for 1-in. and smaller form-factor drives presently give HDDs a 3-4× advantage over flash memory. But with lower-cost, higher-capacity flash memory on the horizon, that advantage could be lost unless significant gains are realized in HDD areal storage density.

Figure 1. a) Perpendicular recording with magnetization perpendicular to the plane of the medium. b) Magnetoresistive sensing of the magnetization state of a storage bit in the HDD medium with perpendicular recording.

Click here to enlarge image
Finally, as the data storage marketplace expands to new CE applications, there is an expectation of improved product quality. HDD manufacturers, seeking to expand their reach into new storage offerings, are actively seeking manufacturing methods that dramatically reduce defect parts-per-million to levels acceptable to a new breed of consumer.

Figure 2. Trends in thin-film magnetic head demand. (Source: TrendFocus 2005 RHO Study)

Click here to enlarge image
Demand for increased areal density will require the industry to employ new technologies like tunneling magnetoresistance (TMR) sensors [1] and perpendicular magnetic recording (PMR) (Fig. 1). The expected demand trend for thin-film magnetic recording heads is illustrated in Fig. 2. And because mechanical factors like slider shape and advanced air-bearing surfaces (ABS) (Fig. 3) are so critical to control of slider reliability and fly height (the nanometers of distance separating an HDD head from the rotating magnetic media), the fabrication of next-generation femto sliders will challenge fundamental limits in manufacturing technology.

Figure 3. An etched Al2O3-TiC composite slider showing the ion-milled air-bearing surface (ABS) that controls airflow and pressure while the disk is spinning. The TFMH is mounted to the trailing edge of the slider (shown in blue and red).

Click here to enlarge image
null

What’s ahead in HDD manufacturing

Thin-film magnetic head (TFMH) fabrication involves state-of-the-art manufacturing processes that are challenged by three near-concurrent changes in HDD sensor technology: the shift to femto form-factor sliders requiring advanced ABS for fly height control, the introduction of tunneling giant magnetoresistance sensors/current perpendicular-to-plane (TGMR/CPP), and perpendicular recording [2]. The shift to nanodomain manufacturing with increasing cost and quality pressures will drive new manufacturing solutions to address the added number and complexity of process steps. The answer, in part, will lie in the increased use of integrated and common platform toolsets.
A typical example of an integrated toolset offering advantages in footprint, throughput, and product quality can be found in the formation, isolation, and biasing of the TMR sensor (Fig. 4). The industry standard for these processes involves ion-beam etch (IBE) for sensor definition, ion-beam deposition (IBD) or atomic-layer deposition (ALD) for isolation, and IBD for the metal underlayer and hard magnet. Because each of these process steps is critical to ensuring sensor performance, many HDD manufacturers have sought to combine these individual modules into a single toolset to eliminate any variability due to vacuum breaks between process steps. As these processes mature, as well as alternative processing techniques such as substituting ALD for IBD on the isolation layer, this approach offers cost benefits to the user because these available modules can be redeployed to other processes.

Figure 4. TMR sensor with process requirements indicated.

Click here to enlarge image
As TFMH sensor scaling continues, conventional liftoff photolithographic patterning is no longer adequate. With an integrated process, read sensor definition can be accomplished with a liftoff technique based on chemical mechanical polishing (CMP) (Fig. 5). Another advanced process, ALD, provides a diamond-like carbon (DLC) film above the sensor stack to act as a CMP stop layer. This limits wafer bowing, allows reduced shield-to-shield spacing, maintains symmetry of the sensor structure that is critical to reliable performance, and permits damage-free removal of the DLC prior to shield deposition. It also reduces inboard/outboard deposition asymmetry during wafer-level rowbar formation.

Figure 5. CMP-based photolithographic definition of a femto slider read sensor.

Click here to enlarge image
Another example of integrated processing is found in the perpendicular pole formation and encapsulation. A PMR write pole is an extremely small structure with a cross-section ≤0.02µm2 that requires encapsulant protection immediately following etch. As shown in Fig. 6, this encapsulation step can be achieved more effectively through high-rate ALD than through conventional physical vapor deposition (PVD), since the latter tends to have poor fill characteristics because of the write pole’s re-entrant sidewall angles. Pinhole-free ALD promotes better slider processing for subsequent CMP, slicing, grinding, and lapping operations.

Figure 6. ALD used for write-pole encapsulation promotes process reliability.

Click here to enlarge image
Advanced process control and reliability demands are expected for advanced TFMH fabrication and have moved capital equipment suppliers toward the design of integrated processing systems that leverage a number of process technologies on a common hardware and software architecture.

Metrology

Advanced metrology now plays a critical role in control of production processes for HDD femto slider fabrication. New techniques and instrumentation that employ advanced algorithms to measure partial etch depths with photoresist in place eliminate unneeded iterative photoresist strip and measurement procedures. Another tool in the emerging optical measurement kit is an innovative, long working-distance objective that, with a complementary illumination source, allows 2-20× viewing of surfaces right through cover glass and compensating glass layers. Computer film modeling and simulation have advanced to the point of allowing good correlation of in situ measurements with actual film thickness so process control can be improved.
External to the process toolsets, atomic-force microscopy (AFM)-based instrumentation is being employed to measure critical parameters of PMR poles, including re-entrant sidewall angles; pole width and height; and throat length. A combination of optical and probe-based techniques can provide AFM resolution and the speed and large-area metrology capability of optical interferometry.

Slider manufacturing challenges

One of the primary challenge areas lies in the manufacture of advanced femto form-factor (0.85×0.7×0.23mm, 0.6mg) thin-film magnetic read and write heads (sliders). Slider dimensions well into the nanodomain (<100nm) are anticipated, reaffirming that these small devices are already among the most complex and hard-to-make nanostructures in production today.
The shape of the slider and its ABS are critical to proper HDD operation, and mechanical processes that form these are an essential part of slider shape control. Advanced slider formation typically includes a row slice and grinding operation to establish a parallel slider shape and a series of lapping sequences. Rough lapping quickly removes 20-25µm of material at high throughput for enhanced productivity; fine lapping removes 3-5µm more of material and establishes an acceptable surface finish; and a final lap (sometimes called a “kiss lap” or “touch lap”) provides final stress control. New TFMH technology, such as PMR, requires new levels of control for micro angle-adjustment of wedge edges to set reader-writer offset and even finer control of stripe height. The result is a slider body with parallelism and orthogonality of its surfaces precisely defined, and reader-writer performance established through closed-loop process control.
Small form-factor drives have also introduced new demands on slider dicing operations to reduce latent particle contamination from dicing residue - a leading cause of HDD failure. One approach to this problem is two-pass slicing/polishing using a dual spindle tool. This has reduced average slider-edge surface roughness (Ra) to <2nm and pressure ridge height to <2nm, while facilitating particulate reduction.

Summary

As manufacturers move toward building HDDs with densities beyond 250Gb/in.2, problems brought on by the combination of nano dimensions, low signal levels, and quantum-level noise will require solutions based on advanced materials science, such as novel magnetic materials and grain size engineering, as well as a new level of manufacturing control coupled with advanced metrology. The nanotechnology threshold has only recently been crossed in high-technology manufacturing, yet common challenges have already surfaced, and HDD solutions and methods may come from unexpected sources.
With a new generation of consumer electronics demanding vast increases in storage capacity at ever decreasing cost, the HDD industry is being pressed to surmount nanolevel manufacturing issues while maintaining a cost advantage over competing technologies. Reducing HDD form factors will be the engineering challenge, but over the next decade the key to HDD price point survival will be steady growth of areal storage density while improving HDD quality generation after generation. TFMH process equipment and metrology that provide cost-effective, flexible manufacturing solutions to support development and ramp to volume will be critical to advancing the HDD industry.

References

  1. Recent work has shown TMR ΔR/R values with advanced film stacks of up to 300%. (S.S.P. Parkin, et al., Nature Materials, Vol. 3, Dec. 2004.)
  2. K.E. Williams, “Hard Disk-drive Technology Revolutionizes Processing,” Solid State Technology Data Storage Supplement, p. S21, Sept. 2004.
Adrian Devasahayam received his PhD in electrical engineering from Carnegie Mellon U. and is director of process development at Veeco Instruments Inc., 1 Terminal Dr., Plainview, NY 11803.

No comments: