Modern methods help disk drive manufacturers combat contamination
By Roger W. Welker
A wide variety of strategies is available to prevent contamination of the disk drive. Key to successful application of these strategies is detecting and quantifying the contaminants in their various forms. Four general classes of contamination that are of interest include: organic, ionic, particle and magnetic contamination (a special class of particle contamination).
Resources for controlling contamination generally are limited. In order to maximize effectiveness of these resources, a management tool is needed to prioritize contamination fixes. A simple management tool for prioritizing effort is presented.
Effects of organic contamination
Organic contamination inside a disk drive can cause several different types of failures.
Among these is stiction, caused by a thin film of condensed phase organic material wetting the interface between the head and disk when the heads are parked in the landing zone. If sufficient organic material is present, the slider sticks so securely to the disk that the motor may fail to rotate the disk pack. The use of very low power consumption spindle motors has increased the sensitivity of drives to this failure mode.
An interesting and particularly aggressive example of a material which can cause stiction is the lubricant used to draw voice coil wires. This material is a mixture of straight chain hydrocarbons. Several of these have significant vapor pressure at coil operating temperatures, so they appear in the vapor phase in the drive during operation when the voice coils are hot. At drive operating temperatures, many of these materials are liquid so they condense to form a very fine aerosol mist. The aerosol mist accumulates on the sliders. When the drive is turned off, the heads land on the disks. If sufficient hydrocarbon residue is on the slider, it can wick between the head and disk. When the drive cools, it solidifies, making a remarkably strong bond between the heads and disks.
Another failure mechanism is gas-to-solid conversion on the surface of the slider. If this conversion increases the apparent surface area of the air bearing, the head will fly higher, resulting in loss of amplitude and increased soft error rate. The use of 75 percent and 50 percent sliders, combined with very low fly height, has increased disk drive sensitivity to this failure mode.
A third mechanism is alteration of the sticking coefficient of surfaces on which the organic molecule has been adsorbed. Organic molecules which increase the stickiness of the surface of the head or disk accelerate accumulation of other contaminants, often resulting in head crashes.
A fourth effect is tribopolymerization. This can occur from organic materials in both the solid and liquid phase. Tribopolymerization of material in the lead-in portion of the air bearing surface often results in erratic flying and eventual head crashes. Tribopolymerization of material on the air bearing surface can also result in stiction.
Some classes of organic molecules interfere with the wetting characteristics and reflow properties of mobile disk lube. Such chemicals can result in localized lube depletion, accelerating wear of the disk and thus leading to premature head disk failure.
The above mechanisms are the result of transport of the organic materials from their source to the head or disk. In general, two primary transport mechanisms allow organic contaminants to get from a source to the head or disk. These are bulk aerosolization (formation of droplets) by motion of moving parts within the drive and vapor phase transport. Often, vapors outgassed from hot components inside the drive immediately condense to form hyperfine aerosols.
Effects of ionic contamination
Ionic contamination promotes corrosion of components within the drive. Anions such as chloride, nitrate and sulfate (three of the many anions controlled) can react in the presence of moisture to form their corresponding volatile hydrochloric, nitric and sulfuric acids. These high vapor pressure compounds then can travel to sensitive components within the drive, such as the heads and disks, where they promote corrosion.
Two categories of ionic contamination tests are in use. One type measures the amount of ionic contamination on the surface or extractable from within the part. The second type of test characterizes the reaction of the corrosion-sensitive components in the drive to the part.
Effects of particle contamination
Of all the forms of contamination, particle phase contamination is most often responsible for disk drive failures. Some examples of how particle contamination causes drive failures include the following:
Particle contamination on the head or disk causes a head crash. This is perhaps the most intuitive effect of particle contamination.
Hard particles stuck to the disk can result in excessive wear to the pole tip of the read/write element.
Hard particles stuck to the head can scratch the disk.
Particles embedded in or lying underneath the thin film of the disk can cause thermal asperities on magnetoresistive heads — a failure mode not observed with inductive, thin film read elements.
Particle contamination between the disk and motor hub, disk and spacer ring, or disk and clamp can warp the disk surface.
Loose particles inside the drive can cause particle count failures.
Particle deposits on the air-bearing surface can create local hot spots, causing thermal degradation of volatile material which, in the absence of a particle, would be thermally stable.
Effects of magnetic contamination
Magnetic particles are an especially dangerous form of contamination for magnetic disk drives, due to their ability to erase data. Loose magnetic particles in the disk drive can find their way to the recording head where, if they are attached to surfaces close to the disk, can distort the write field or erase data. Because these particles are generally not sticky, they can appear to erase data and then disappear, resulting in a failure mode described as transient erasure. These are especially difficult failures to analyze since the particles which caused an erasure often are no longer present.
Considering the high degree of sensitivity of disk drives to magnetic contamination, it is surprising that high energy magnetic materials are used to make components which are internal to the drive. Specifically, voice coil motors for positioning the heads are large, high-coercivity permanent magnets. Latches in all modern drive designs use small, high-coercivity permanent magnets and per manent magnets are used in spindle motors which are internal to the disk drive. In addition, many of the tools used in the assembly of disk drives can have magnets built into them.
One way to manage contamination control activities is source apportionment, whiich is used to decide where additional efforts should be directed to provide the largest potential benefit.
Contamination may be thought of as originating from broad categories of sources. A pie chart is a convenient way to visualize the source apportionment; the size of the wedge is proportional to the contamination contributed by that source. The input data for the source apportionment can be the results of failure analysis, yield loss analysis, or even a random sample of HDAs for chemical analysis of the contaminants within them.
In many cases, the contamination causing the failure cannot be found or identified. This can be the result of loss of the contaminant during shipping, handling and sample preparation. It can also be the result of misidentification of contamination as a suspect cause.
Also, in many cases the contamination cannot be attributed to a single probable source. Where such ambiguity exists, a fraction of the failure is apportioned equally to the possible sources. Consider the case where disk scratches caused a field failure and the scratch and head are found to be contaminated with 300 series stainless steel. Several parts inside the HDA are made from 300 series stainless, which is also used in the fabrication of the tooling used to assemble the disk drive. Since the particles found on the head or disk cannot be uniquely associated with one source, half of this failure is binned in the piece parts category and half is binned in the tooling category.
Contamination from piece parts and subassemblies
The strategy to control contamination from piece parts and subassemblies consists of two phases: materials qualification and on-going process control.
During materials qualification several types of tests are conducted. Contact and near contact stain tests are run to verify that proposed materials are compatible with the candidate heads and disks. In addition, ion chromatography, gas chromatography/mass spectrophotometry, Fourier transform infrared spectroscopy and SEM/EDX are used to determine if the materials are free of contaminants known to be hazardous to drive reliability. In parallel, LPC and percent weight loss measurements are made.
The results of these qualification tests are then combined with process control data collected early in the production life of the product. The process control data is then used to determine which on-going tests are appropriate for each part. For example, if integrated circuit data shows that a part is 6 sigma capable of meeting the anion cleanliness requirements, then little value will be gained by routinely testing the part for anion content. Periodic audit measurements should be sufficient to assure that the process remains in control. If, on the other hand, the process characterization data shows that the part is only 1.5 sigma capable of meeting the LPC requirement, then some test for measuring particles will be needed on a lot-by-lot basis, preferably as a source inspection.
Personnel and practices
Contamination from people working in the cleanroom is among the most difficult to control because so much of the problem is related to behavior. For example, the best designed cleanroom garment and the best laundry service will not prevent the introduction of skin, hair, street clothing fibers, soil particles, etc. into the cleanroom. Periodic audits and frequent retraining sessions are needed to maintain the high level of discipline required.
Tooling in this context includes fully and partially automated assembly fixtures, conveyors systems, power and hand tools. Tooling contamination issues must be dealt with in a number of ways.
First, materials for use in tooling must be subject to the same rigorous qualification procedures as are materials for piece parts, although different control limits may apply. For example, adhesives and other polymers in a disk drive are generally limited to no more than 0.1 percent weight loss. Conversely, adhesives used in locations on tools that are not in contact with or close proximity to the drives may be allowed as much as 1 percent, as long as no known drive-killing chemicals are included in the volatile effluent.
Second, the design of the tool, including selection and layout of components, is constrained by the requirements of the cleanroom in which the tool is used. The tools are reviewed for conformance to contamination control requirements during design, at the prototype level and during installation of the replicated, production tool set.
Third, tools are subject to periodic audits while in production. This is done using simple visual in spection to verify that critical and busy locations are free of contamination.
State-of-the-art disk drives are more susceptible to contamination than ever before. The use of MR read heads has introduced failure modes never before recognized. Can contamination control keep up with the challenges facing the manufacture of disk drives?
It is possible for disk drive manufacturers who use the most modern approaches to contamination control to stand up to this challenge, and do so in a cost effective manner.
Roger W. Welker is the former senior director of contamination and electrostatic discharge control for Micropolis Corp. (Chatsworth, CA).